Increasing ISO to Reduce Grain in Your Images (Podcast 661)

Increasing ISO to Reduce Grain in Your Images (Podcast 661)

Today I’m going to provide some clear evidence of the benefits to increasing your ISO to help reduce grain. Sounds crazy, I know, but this works, and after this post I think you’ll believe me.

I have been meaning to do this for a while, as although I’ve talked about Exposing to the Right and how increasing the ISO can help to reduce noise, I do realize that this is counterintuitive. I also received an email from listener Matthew Wells, as follows:

In several podcasts, you mention how you will shoot higher ISO levels by exposing to the right to brighten the image to reduce the visible noise. Could you put together a short video for your Youtube channel to show your methods for doing this? I have tried to play with the idea a little without success so far.

Matthew Wells via email.

Thanks very much for your suggestion Matthew. I decided to do this as a straight post rather than a video so that I can provide some examples that you can open and look at directly, and I think this format will help more for now.

Why Take Control?

So, the first thing I want to do is explain why I take control, with an example you can try for yourself really easily. All you need is something white, and something black, or a very dark color. Literally, a large piece of paper or a bedsheet is fine for the white thing, and a dark coat will work for the dark thing, and this works best if the things you find are not too glossy or shiny.

In my studio, I have both white and black backgrounds on a pulley system so that I can change the background for my photographs quickly. I have a roll of black paper that we won’t use for this, and a roll of white paper, and a roll of felt cloth, that we will use. You can see what I mean in this photograph, with the black background extended part way so that you can see them both.

Black and White Backgrounds

White is Grey!

Here now, is a photograph of the back of my camera, in LiveView, and set in Aperture Priority mode with Auto-ISO turned on and no Exposure Compensation dialed in. There are a couple of things that I want you to look at. Firstly, while noting that I’m pointing the camera at a sheet of white paper, look where the data of the histogram is falling. All three channels are only one third from the left side. This means as you can see from the photo that the white background would be recorded as grey, not white.

White is Grey!

The other thing that I want you to look at is that the shutter speed has been set to 1/160 of a second and the ISO is 4000. I set the aperture to f/5.6 myself, and the camera has automatically selected the other settings.

Black is Grey Too!

Now let’s look at another photo, and the only difference between the two images is that I’ve drawn down the black felt background, as you can see in front of the camera. I didn’t touch the camera other than to half-press the shutter to display the exposure details. Look at what’s happened to the color of the background in LiveView. Nothing! The black background is also grey, and the histogram shows us that it’s very similar, if not almost identical to the grey in the photograph of the white background.

Black is Grey Too!

We can also see that the settings have changed. The shutter speed has been increased to 1/15 of a second from 1/160 and the ISO has gone from 4000 to 12800. Using the Exposure Calculator in my Photographer’s Friend app I can quickly see that the camera has changed the exposure from 17 2/3 EV to 16 EV, a difference of one and two-thirds of a stop. Just to quickly walk through this, the shutter speed change is 3 and a third of a stop, and the ISO change is one and two-thirds of a stop, so if we deduct the ISO change from the shutter speed change, we get one and two-thirds.

I could have left Auto-ISO off to make that easier to understand, but I also wanted to show how the camera would want to increase the ISO instead of taking the shutter speed much longer.

So, what does this tell us? Well, for one thing, it tells us that our cameras are still pretty stupid when it comes to metering certain scenes and setting its exposure. This is why I almost always shoot in Manual mode. You can use Exposure Compensation too, and which method you use is completely up to you, but I find using Manual easier, mostly because I’ve been doing it so long, but also because as the size and position of the various elements in your frame change, the exposure will also change, so you have to continuously adjust the Exposure Compensation, especially when you are Exposing to the Right, which we’ll discuss shortly.

This is extremely important when photographing things like the Red-Crowned Cranes that we shoot on my Japan Winter Wildlife Tours. These are a white bird against a white background when they are on the snow, but when they fly, the background is much darker, and as you’ve seen, the camera always tries to make a dark scene brighter and a bright scene darker.

So if you dial in up to plus two stops of Exposure Compensation, which is required to make the snow and white bird actually white, then recompose and start to include a darker background, the camera will increase its exposure to lighten the dark background, and the white bird essentially becomes completely over-exposed.

A Real-World Example

Here is a screenshot of Capture One Pro showing two photographs that I shot as I explained this to a participant on this year’s tour. To prove the point, I put the camera into Aperture Priority mode and pointed my camera down so that only white snow was in the frame. This is how I set my exposure in Manual Mode, because the cranes are also mostly white.

Exposure Example
Exposure Example

Keep in mind that to set my exposure in Manual Mode, all I do is increase my exposure while looking at the caret on the meter, until it’s between +f and+2 stops over zero. For brightly lit snow, it’s closer to +1 and for snow on an overcast day, it’s usually around +2 stops. Now that I’m shooting with the mirrorless EOS R camera, I have a live histogram and can actually just keep on increasing the exposure until the data on the histogram is almost touching the right shoulder, but without this feature, I used to just take a test photo, just like the one you see on the left here. As long as it’s white, but not over-exposed, I know that I’m then good to go, and because the bird is white, like the snow, and under the same light, I am then free to just shoot away until the light changes again.

In an automatic exposure mode though, such as Aperture Priority in our example, as you can see, as soon as I recompose to include an even slightly darker background, the camera tries to lighten up the scene, and my whites start to get over-exposed. So, if you want your whites to be white, not grey, you really must take control of the exposure by locking it down in Manual mode, or you have to change your Exposure Compensation every time you recompose.

I’ve added a few bits of markup to the screenshot to point out a few key things. First, notice on the left that I have +2 stops of Exposure Compensation dialed in, and I’m in Aperture Priority mode for this example. Also, note how the shutter speed changes from 1/500 of a second for the correctly exposed shot and drops to 1/200 of a second as the camera tries to compensate for the darker background. It would probably make more sense to use shutter priority for birds in flight, but as an example, it doesn’t really matter. The point is, the exposure runs off as we recompose.

Also note that the red that you see over the bird and the snow are not markup, that is the Exposure Warning in Capture One Pro. You can see that this is turned on also by the fact that the icon in the toolbar is yellow. This is grey when the Exposure Warnings are turned off.

Higher ISOs are Less Grainy with ETTR

So, let’s move on to the benefits of increasing the ISO rather than being afraid to, for fear of introducing grain. Most people have the idea that increasing the ISO causes grain in your images, and of course, that is true, if you leave the exposure up to the camera, because the camera is generally going to underexpose your images. Here is an example with two photos that I shot as a test, just of some ornaments and a little metal EOS R that I got as part of the release campaign here in Japan.

ISO Example
ISO Example

If you click on the image to open it up in the lightbox, even at the web size you’ll be able to see the visible grain in the right image, which I shot with the camera’s built-in meter at zero, exactly where the camera thought the exposure needed to be. It was indoors late afternoon, and because the ISO performance is so good on the EOS R, I had to increase it to 51200 for you to really see the grain.

For the left image though, to prove my point, I used Aperture Priority mode again, with +2 stops of Exposure Compensation dialed in, as I’ve circled in red on the left of the screenshot. You can see underlined in red, that both images were shot at 51200 ISO and the +2 stops images shutter speed was 1/125 of a second, exactly two stops slower than the right image, which was shot at 1/500 of a second.

Comparison

Here now, is a 100% crop of each of the images from the above screenshot, so that you can check the details. As you can see, despite them both being shot at ISO 51200, the grain visible in the +2 stops version, which is I should emphasize, exposed to the right, is virtually undetectable, compared to the image at the camera’s recommended metering, which is very grainy.

Again, if you click on the images to view them in the Lightbox you’ll get the best view of the detail, but the difference I’m sure you’ll agree is huge. Note too that although in comparison, the lighter image may appear too bright, it’s actually only slightly over-exposed, with the highlights just clipping slightly. If you wanted to darken it back down again, you could use the Highlights sliders or a tone curve, and you’d still get a cleaner image by exposing to the right like this, then darkening it back down to suit your needs.

Another Take On This

Another take on this, which is perhaps easier to understand from a shooting workflow perspective, is the fear element that generally prevents people from increasing the ISO. Imagine you are in a situation where the light is low, and you are already forced to shoot your image at 12800 ISO, at 1/125 of a second to avoid subject blur, and you need an aperture of f/8 for sufficient depth of field. The camera is metering at zero, with the information in the histogram way over on the center decreasing down the left side. The actual ISO will depend on your camera, but many people start to shy away from shooting higher than around 3200, some even as low as 1600. For me, as I get used to the EOS R’s ISO performance, I would probably not have gone higher than 12800 in the field, because I hadn’t yet done these tests. On my 5Ds R bodies, I tried to avoid going above 6400 ISO, based on tests.

ISO 12800 Camera Zero Metering @ 100%
ISO 12800 Camera Zero Metering @ 100%

As you can see from this image, with its histogram embedded for reference, there is a little bit of grain starting to creep in. Still incredible for a 12800 ISO image at zero metering, but you can see the grain.

+1 Stop Increase

So, what do you? Most people would shoot at 12800 or whatever your own personal soft-ceiling is, and be afraid to increase the ISO any more for fear of introducing any more grain, but wait! To get your histogram data over to the right, in this hypothetical example, our only option is to increase the ISO further. Here is another image in which I’ve increased the ISO by one stop to 25600.

ISO 25600 +1 Stop @ 100%
ISO 25600 +1 Stop @ 100%

This image is double the ISO at 25600, but the grain is actually less than the 12800 ISO image because the ISO increase has made the image brighter. You can see from the histogram, which I screen-captured from the full sized image, not the cropped version, that although it’s more to the right than the previous images, there is still a gap.

One More Stop!

For this next image, I increased the ISO yet another stop, to 51200. This is actually the lighter image of the pair that we just looked at in my first example. and as you can see from the histogram, we are now exposing to the right. There is a small gap, but that’s Capture One Pro giving us a little back. In the camera, the specular highlights were just starting to over-expose, so this is as far as I would like to take this image exposure-wise.

ISO 51200 +2 Stops @ 100%
ISO 51200 +2 Stops @ 100%

A Few Tweaks

You can see that increasing the exposure with the ISO has actually reduced the amount of grain visible in the photograph. The darker parts of the image are now starting to suffer from a little bit more grain, but if you really had to push your ISO this far, you could do a few tweaks such as adjusting the levels and curves to darken down the dark areas a little, which helps to mask the grain that does creep in there.

ISO 51200 +2 Stops Tweaked @ 100%
ISO 51200 +2 Stops Tweaked @ 100%

And I haven’t even touched the noise removal options. They have been set at the default settings all along. So, as you can see, although I know it’s counter-intuitive and probably goes against everything you’ve been taught, if increasing the ISO helps you to expose to the right, then the brighter image will almost always have less grain in it than the lower ISO image, if the lower ISO image is already very high.

I should add at this point, that for lower ISOs, where you really can’t see any grain anyway, there is something called ISO invariance, that we looked at in Episode 520, which basically means that you can shoot according to your camera’s meter at lower ISO, and then increase the brightness in post if necessary, and you still won’t see any grain, because there is nothing in the base image for you to amplify. What I’m talking about today is more for higher ISOs, which is generally the area that starts to get people nervous. It’s that fear that I’m hoping I can help you to break through with this post.

My ETTR Workflow

I’m sure that part of Matthew’s original question was also referring to how I actually adjust my exposure, so I’d like to add a little more detail on this before we close. As I mentioned earlier, I pretty much always shoot in Manual Mode, though occasionally I do shoot in Aperture Priority and turn on Auto-ISO, and use Exposure Compensation to Expose to the Right. My thought process is similar in either shooting mode.

Aperture

I start usually by selecting my Aperture, as this directly affects my depth of field, and that is often one of my most important decisions as I start to set my exposure. Once I have set my aperture to something appropriate, i.e. a small number like f/2.8 for a wide aperture and shallow depth of field, or a larger number like f/11 for a smaller aperture, and more depth of field.

Shutter Speed

I generally then set my shutter speed in Manual Mode. I will select a faster shutter speed, like 1/500 to freeze a moving subject or a faster shutter speed of 1/1600 of for fast moving subject or birds in flight. Or I might select a slower shutter for a landscape, or even use a Neutral Density filter to slow down the shutter speed even more for some of my landscape work. If I am in Aperture Priority mode I often control the shutter speed by setting a minimum speed in the camera settings, just to help the camera to avoid going too slow, but getting faster is generally not a problem in these cases.

ISO

Finally, I adjust my ISO. It’s actually my ISO most of the time that I use to actually adjust and fine tune my exposure. This is why I am really enjoying the new Control Ring on the front of the Canon RF Lenses, because I have this set to adjust my ISO, so I can now really easily adjust it while looking through the viewfinder and keeping my eye on the live histogram.

With the live histogram I literally just adjust my ISO until I see the data on the histogram just about start touching the right side. For landscapes, even with my older DSLR bodies, I used to turn on LiveView and do this on the LCD.

Highlight Warnings

It’s also important to turn on Highlight Warnings on your camera so that you can see when you do start to get over-exposed and pull it back a little. I will often use a very small amount of flashing or the “blinkies” as an indication that I’m right where I need to be exposure-wise, but once a larger area starts to blink, I pull the exposure back a little.

RGB Histogram

And it’s also important to use the RGB histogram rather than the black and white brightness histogram, because the brightness histogram is an average of all three channels, and doesn’t always show you if one color is becoming overexposed before the others, and it can cause a nasty blotchiness in your colors if you allow one color to blow-out.

Camera Meter in Manual Mode

Also, note that even when I am in Manual Mode, I still reference the camera’s meter reading. People often think you are flying blind in Manual Mode but that is not the case. When you half-press the shutter, the meter still kicks in and shows you where it thinks the light levels are on the meter, so for example when I am adjusting for the white snow, I can see the caret moving on the meter scale, and use that to see when my settings result in my exposure being two stops over.

If it’s not a white snow scene, I will just guess at where the caret should be on the meter scale, based on the balance of light and dark objects in the scene. If it’s literally 50/50, then the meter might be at zero, or perhaps just a little over to get my histogram data over to the right.

Note too that I’m always talking about the right-most data, not all of the histogram data. The rest of the data represents your mid-tones and shadows, and can, depending on your scene, extend all the way over to the left shoulder. Occasionally you might shoot a scene with very high contrast, and find that even when exposing your highlights to the right, your shadows can start to spike up the left side. That’s when you might consider merging multiple images in an HDR photo, but personally, I have not had to do that for many years, while using my ETTR techniques.

Thanks for the question Matthew, and I hope this was useful for all that stop by and take the time to read or listen to this episode.

Complete Namibia Tour & Workshop

Before we do finish, I’d like to mention that we’re at the time when we need to start finalizing the numbers for my 2020 Complete Namibia Tour & Workshop This is an epic trip that encompassed landscape, culture and wildlife and we next year will be my sixth visit, and the tour just keeps getting better and better, so I’d love for you to join me there. Please check out the details at https://mbp.ac/namibia2020 and drop me a line if you have any questions. If you visit this post after the 2020 tour closes for bookings, check our Tour & Workshops page for a list of currently available tours.

Complete Namibia Tour & Workshop 2020

Show Notes

Check out my ISO Invariance article here: https://mbp.ac/520

See details of my 2020 Complete Namibia Tour & Workshop here: https://mbp.ac/namibia2020

Music by Martin Bailey


Audio

Subscribe in iTunes to get Podcasts delivered automatically to your computer.

Download this Podcast as an MP3 with Chapters.

Visit this page for help on how to view the images in MP3 files.


Developing Your Style and Self-Acceptance (Podcast 635)

Developing Your Style and Self-Acceptance (Podcast 635)

In this week’s post, I’m going to relay my thoughts on spending lots of time on lengthy post-processing techniques to “improve” our images, or more accurately, what I do to avoid spending this time, including some self-acceptance advice, based on communication with a listener.

This episode was spurred by a recent email exchange with listener Evan Stewart from Saint Louis, Missouri, so I want to start by thanking Evan for the mail, your questions, and providing some food for thought.

In the exchange, Evan asked how much I use Photoshop and mentioned that he is at the point in his photography where he can’t help being curious about trying more advanced techniques such as focus stacking, exposure blending, and luminosity masking, and was asking if I used some of these techniques in my own photography. My reply may well be useful to others, so I decided to share and expand on that reply here on the blog.

Personal Preference

To set the stage and relay a kind of disclaimer about what I’ll follow on to, I will start by saying what I said to Evan in a follow-up to my main reply, which is that some people feel as though they must spend hours on a photo to make it good, and I feel for them. Aimed at anyone that spends a lot of time on images, I do understand that some people simply enjoy working on images.

I personally do not, which is why I shoot and process the way I do, as I’ll explain shortly, but I can spend hours on my computer doing other things. It’s all personal preference, so please understand that what I’m going to talk about today is not condemning anyone for spending time on images if that’s what you like doing.

If however, you find yourself spending a lot of time working on images to recuse them, and the time spend is putting pressure on you, it may well be time wasted, so I hope that what I’ll relay today will help you to save some of that time, so that you can spend it doing something else more enjoyable.

How Often Do I Use Photoshop?

In reply to Evan’s question about how much I use Photoshop to process my images, I said that I do use Photoshop to work on my photos very rarely. Probably less than 0.1% of my images are edited outside of Capture One Pro, but when they are, I use Photoshop. I used to edit some of my images in the Nik Software suite, named Silver Efex Pro and Color Efex Pro, but I haven’t used these applications at all since switching to Capture One Pro just over two years ago now. Some of the effects that I was getting can be done in Photoshop, but I did like the results I was getting with Nik until Capture One Pro gave me an alternative that I liked.

Time Spent on Image Processing

Partly because of how I shoot, I actually spend less than 30 seconds on the vast majority of my images. Occasionally I shoot something that requires more work, say for example if there are power lines running through a scene, and I decide to shoot it anyway, this kind of photo would require more work, but even then I generally don’t spend more than five minutes or so.

I understand that how little I do may result in my images not being quite as impactful as some of the work of the other photographer’s but this is my style, and I personally enjoy my results. As I’ve mentioned before, I’m under no delusions that I’m some kind of amazing photographer, but I do make work that I am happy with, and for the vast majority of photographer’s out there, I think this is a good place to be in our work.

Content, Not Complacent

Having said that, I also think it’s important to make a distinction between being content with our work and being complacent. Being content leave room for improvement, and I’m always on the look-out for a better photograph, and ways to improve my work, but I want to make those changes in the field, during the creation of the images, not after the event on my computer. 

Complacency, on the other hand, leaves no room for improvement, and this is a place that I never want to be. Even though I shoot some locations every year, there is always an opportunity to improve, and being self-critical is vitally important to that process, and for me, this self-criticality starts in the field, while I still have a chance to change my raw materials, literally my raw images.

Raw, Not RAW

That statement reminds me of a time when Jeffrey Friedl, who you may know for his amazing Lightroom Plugins, pulled me up for spelling raw in capitals, like an acronym. Jeffrey is probably one of the most knowledgeable people I know and realized I’d been getting this wrong when he told me that raw images literally are just raw, as in not cooked, rather than R.A.W. actually meaning something.

It doesn’t help that most camera manufacturers and many people in the industry use RAW, but I can say with my hand on my heart that since Jeffrey explained this to me, I have not only always used the lower case raw, but I think I’ve had a deeper appreciation for what raw images really are, and this also leads nicely to one of the main reasons that I work hard not to make copies of my images in a format that takes me away from my raw files.

Non-Destructive Editing

The moment you export and edit your images in a program like Photoshop, or the Nik plugins, you will eventually end up saving your image in a format other than your original raw image format, and once you do that, you lose the ability to easily update your photographs to the latest and greatest versions of the processing engines that come with your chosen image editing and management software. I use Capture One Pro, but this is the same with Lightroom and many other applications that are what we call “non-destructive”. 

When I first started using Capture One Pro just over two years ago, it was at version 9, and since then, there have been two major releases taking us to version 11, and both of these releases have brought improved image processing with an updated processing engine. When I first processed the image that we’ll look at in a moment, I had some cloning to do that was not possible in Lightroom, so my only option at the time was to take my image into Photoshop to do my editing. This was one of the rare times when I round-tripped out of Lightroom and ended up with a PSD file with my changes baked in.

I resented the time I’d spent in Photoshop though within just a few weeks. I’d taken my original photo into Silver Efex Pro and converted it to black and white, but as I lived with the image I realized that I’d made the trunks of the trees in the foreground a little too dark. I had to go back into Silver Efex Pro and reprocess it, and that meant that I had to also go back into Photoshop and clone out the cable car lines a second time. This felt like such a waste of time to me.

Then, a few months later, in the summer of 2016 I switched to Capture One Pro, and during my initial tests of the product, I used the same image to see if I could both create black and white images that were as good if not better than Silver Efex Pro, and I checked to see if I could do cloning in Capture One Pro, so that I could avoid baking my changes into a destructive file format rather than keeping them in their original raw format.

It turned out that Capture One Pro passed both of these tests, so I now had a copy of my original raw file with all of the cloning done in Capture One and so when versions 10 and then 11 came out, I just clicked a button to update the processing engine, and the image grew incrementally better with more subtle detail and better handling of the shadows. I was able to improve my photo twice, with no more effort than a couple of mouse clicks, simply because I had been able to keep my image in its original raw format, so this is a major time-saver and benefit.

My “No See, No Edit” Policy

Let’s take a look at the image I’m referring to as we start to discuss more of the methods that I use to help improve my photography and save me time in post-processing. Since I started shooting digital back in 2000 I made a conscious decision to kill two birds with one stone. I’m not a purist in the sense that I will not clone anything out of my images, but I don’t like to make the decision to do so lightly. I want to modify the content of my images based a deliberate, conscious decision.

Couple with a desire to create images that require as little as possible work on the computer, I decided to train myself to be more observant in the field, by sticking to my personal policy not removing anything that I was not aware of when I exposed the photograph. If I didn’t see it in the field, I do not allow myself to change it later in postprocessing.

What this means is that if I see the cable car lines running through a scene, I am OK with removing them later, but if I do not see the cable car lines or anything else distracting in the image until I get it onto my computer, I do not allow myself to remove the distraction. I’m left with two options, live with the annoying element, or throw the image out, and 99% of the time I go with the latter option.

To illustrate this, here’s a photograph from Mount Asahi in Hokkaido, with the lines of the cable car running through the scene. I knew about these when I composed the photograph but liked this scene so much that I decided to spend the time to remove them later. There is actually also a large pillar to support the cables, but I positioned my camera so that I hid that behind the third foreground tree from the left. Another decision that I made to help save me time on the computer.

To see the cable cars on the left side of the photo, grab the vertical line in the middle and drag it over to the left side of the image. You’ll see the cables behind the trees across most of the left side of the photograph. Another minor benefit of keeping my images in raw is that I could show you the processed black and white image today, and just turn off the cloning adjustments for the before/after images.

I have found though that my policy of not allowing myself to remove anything that I didn’t see when I initially made the photo has really helped me to be more deliberate in my compositions. There were times when I had to throw out images that I otherwise really liked, so I soon learned that I had to do better. It’s surprising how much a little self-kicking can do. I really recommend it to all photographers.

No Exposure Blending

Evan had asked if I had come to peace with not using any exposure blending techniques, but for me, it’s not so much a case of coming to peace with not using techniques like exposure blending, I really just don’t believe it’s necessary. Part of this in my case is because I don’t often shoot in very bright weather that might cause really high contrast images, but apart from a few times well over ten years ago, when our cameras weren’t as capable as they are now, I really just haven’t photographed any scenes that in my mind needed more than one frame to capture the necessary dynamic range.

Part of this is also my own sense of the aesthetic. You might remember me talking about this photo (below) earlier this year about how I often let the windows go white in my Namibia work from Kolmanskop. This is not because I’m too lazy to take multiple shots and blend them together, but because I really just prefer to see the images that way. It leaves more to the imagination and to me, feels a little more surreal than a photo where we can see both the inside of the room and the exterior perfectly exposed.

Kolmanskop Accountant's Bathroom
Kolmanskop Accountant’s Bathroom

Similarly, I don’t mind in this photograph that the walls are leaning outwards from the angle that I shot this. That is not to say that I never do any keystone adjustments mind. I do, but only when the effects that I a remove are a distraction for me. I think people sometimes feel that they have to remove any and all form of distortion, and again, that’s each individuals decision to make, but personally, I don’t feel that every photograph has to be absolutely perfect in this respect.

Don’t Deliberate, Be Deliberate!

In complete contradiction to that sense of not caring for the above image though, I should note that I do quite often take the time to ensure that my camera is at just the right height to prevent my vertical lines from leaning in at least part of the time. This falls under my heading of being deliberate. I find it ironic that the same word pronounced differently has both ends of the spectrum covered for me in this respect.

I won’t “deliberate” over some issues, such as leaning walls, or spending hours on images in post, but I absolutely will be “deliberate” and take the time necessary to craft my compositions, so that everything is just as I want it, as I did for this photo (below) where I moved the frame on the floor to a more pleasing location and spent extra time ensuring that my camera was at the right height to get all of the vertical lines perfect straight.

Kolmanskop School Corridor
Kolmanskop School Corridor

I guess it’s also ironic that I completely don’t care that we can’t see everything outside the windows to the right in this shot as well. I think my point is, once again, that we have to develop a sense of what we are happy with, and in many respects, this takes time and confidence in our work.

Self-Acceptance

As I think about this for today’s post, I realize that Evan’s email to me that prompted this post had a more appropriate subject than I had initially thought. I started this post with a different title, but as I write, I’ve adopted the subject of Evan’s mail, which was “Developing your Style and Self-Acceptance”.

In some areas of my photography, I’m a stickler for getting it right and ensuring that everything is perfect, but in other areas, I simply decide to not give a hoot, and that really does come from reaching a point where you can not care what other people think. Having the confidence to say that it is simply not important to be able to see what’s outside, and deciding that the walls are leaning outwards doesn’t matter, comes from having a strong sense of self-acceptance.

The Roll of Your Trusted Critique

If you listen to everything that everyone will tell you it can be quite paralyzing, as you really will never be able to please everyone. This is where the significance of your trusted critique comes into play, and equally important, is developing the ability to ignore the words of the untrusted critique.

I’ve mentioned a number of times in the past that my wife is who I consider my main “Trusted Critique”, although there are a few others around the world whose advice I will head. Whenever I have a decision to make about my photography or any other creative pursuit for that matter, I don’t consider my work to be complete until I’ve run it past my wife. Her advice means everything to me, but I should note that I don’t take on board everything she says.

There are times when I like something that she doesn’t, and at the end of the day, it’s my art, and I have the final word, but for example, if I’m working on a set of images, say a portfolio, or a selection of work to send a client, nine times out of ten, if she says that something has to go, it goes. A lot of the time, I use her opinion to check my own suspicions, and over the years, I’ve learned to preempt her advice, and simply remove images that I know she will dislike.

Cooling Off Period

I also recall a story from my good friend Graham Morgan, an award-winning photographer from Australia, who was on his way home from Antarctica and showed his wife a photo that he was about to delete but she told him to keep it. The following year he won the Australia Nature Photographer of the Year with that photo! This is a powerful reminder of the need to seek advice from a trusted critique.

It’s also an important reminder that making the final decision on photos that we’ve shot straight after shooting them isn’t always a good idea. Giving yourself at least a few days, ideally a week or more after a shoot to make a more subjective decision about your work is very important.

When I return from a trip, I generally want to start talking about it within just a few days, so I force myself to complete at least a preliminary cull as soon as possible after getting home, but then I always continue to refine the selection for another week or two until I make my final selection. Invariably I find that as the memory of the shoot fades slightly, I am able to be more ruthless in my selection, and the more we can remove from a selection, the more condensed and rich the final set will be.

Controlling Exposure for Optimal Dynamic Range

As I mentioned earlier, in part of my reply to Evan, I had also stated that if you are careful with your exposure, you can generally get a better quality image from a single frame than from blending. Now, I realize that there are techniques that can help to blend images together without really being able to tell it was done, but the majority of the time they require more work than I personally am prepared to put into each individual images, so once again, this is personal preference.

My advice to people has always been that if using exposure blending or any other HDR techniques feeds your creativity, then go for it. If however, you are doing it to overcome limitations on your camera, then first learn how to expose your photos to get maximum dynamic range in a single image, then decide whether or not you need HDR. Once you get used to looking at the histogram, there isn’t even any guess-work. You can see right on the back of your camera if the image fits in a single frame, simply by referencing your histogram.

Exposing To The Right (ETTR)

Now, I’ve talked about ETTR or Exposing To The Right a number of times in the past, so I’m not going to go into great detail on this, but I want to summarise why I still do it, and a few other considerations that you might keep in mind if you are still formulating your own strategy.

First of all, I want to mention that thanks to the wisdom received from Graeme Nattress, who’s forgotten more about image processing than I’ll ever know, I’m no longer talking about the cause of noise in images based on the old Luminous Landscape article that used to be my primary reference. Graeme helped me to understand that the old explanation is not valid, but also that the benefits of Exposing to the Right are very real all the same.

I have though been Exposing to the Right instinctively for many years though, and experience has proved to me that my images are better quality for it, so I’m still using this technique for almost all of my work.

Basically, what I do exposure all of my images so that the lightest part of the histogram data falls just short if not slightly touching the right shoulder of the histogram. On the rare occasion when I find myself photographing a scene with the sun in the frame, I will allow it to become over-exposed slightly, to give myself some leeway in the shadows, but even then, one frame generally gives me enough dynamic range.

As an example, here’s a photo (below) of a lighthouse with the late afternoon sun shining through the windows of the lantern room, shot on this year’s Hokkaido Landscape Photography Tour. When looking through the viewfinder, as you might imagine, everything except the sun looked almost completely black. I could see from the histogram though, that if I set my exposure so that the sun is just starting to over-expose, the shadows were not spiking up the left shoulder.

Again, this is a comparison image, so you can slide the vertical bar from left to right to reveal more of the final processed image on the left, to see how much detail I brought out in the bottom third of the photograph. All I did was ran a gradient mask over the bottom third and opened up the shadows slider to around half way. I also pulled out the shadows across the entire image with a very slight tone curve adjustment and placed a second gradient mask across the top of the sky to darken it down a little. This isn’t the greatest image I’ve shot by a long shot, but I think it helps to illustrate my point.

Reading the Histogram

Basically, expose so that the information on the right of the histogram is just about, or in this case, just touching the right shoulder, and check to see if you have a spike on the left side. If you do see a spike on the left side, it means that the shadows are going completely black, which, similar to highlights going completely white means that it might be difficult or even impossible, to recover any detail in these areas. As long as the shadows on the left side are not spiking though, there is detail in the shadows that can be used.

Noshappu Lighthouse Photo Histogram
Noshappu Lighthouse Photo Histogram

Do also be aware that the camera’s histogram is generally based on the JPEG preview, so it tells a somewhat harsher story than your image processing software probably will. Both Lightroom and Capture One Pro give you around a stop of light back, compared to how the images look on the camera. 

I also use a piece of software called RawDigger to check my images, especially ones like this, to see what the image really looks like, and I can also create a very details histogram, such as this one (right) that I got from the original raw file of the photo we just looked at.

In the RawDigger histogram, we can see a small spike on the right side of the histogram, which represents the bright sky around the sun that I overexposed slightly. On the left you can see that the shadow data tapers off nicely, showing that there is a little bit of information lost, but enough to be able to bring out detail in the shadows.

I don’t have a photo of the histogram on the camera, but here two are two histograms from Capture One Pro with the original raw file on the left and the processed raw file on the right. You can see how the shadows were steep and close to the left shoulder, but not spiking. Also, note the small spike in the highlights for the sun, but see how I was able to bring that under control in Capture One Pro. Going over a little bit is rarely a problem, but we do need to avoid going over by a large amount.

Histograms

So, as you can see, using the histogram as a tool can really help to make educated decisions about your exposure. Based just on the image, or even just what you see through the viewfinder, you might think that a scene like the Lighthouse sunset would need multiple exposures merged together, or even HDR processing, but in my opinion it really isn’t necessary, unless, as I mentioned, you actually enjoy the process and that process feeds your creativity. 

For me, there is nothing more off-putting than getting home from a shoot and having to spend hours working on images, and as fast as the process may have become, as far as I’m aware, there are still no applications that will allow you to keep a processed and tone-mapped HDR image in raw format, so you cannot benefit from future processing engine updates.

Don’t Fight Your ISO

That example was for a high contrast scene, but as I said, I Expose to the Right for all of my work, and I have received feedback from people that know, that the quality of my images is higher than most, and we attribute this to the way I shoot, taking control of my exposure. 

Basically, the darker your images get, the more noise you’ll see in the shadows. I have talked about ISO Invariance in Episode 520, and I agree that if your base ISO could be ISO 100 you can increase the exposure in post for a number of stops, and really not see a lot of degradation in image quality. This means that especially when you are running and gunning and the light is changing so fast that manual exposure may cause you to miss shots, you can go to an automated mode and then brighten up the images in post if necessary.

This is only the case though if the scene is bright enough that ISO 100 would get you to within two or three stops of your required exposure. Once your scene is so dark that you ISO 100 does not get you to within two or three stops of your necessary exposure, the only option is to increase the ISO, and when you start to increase your ISO, you can’t benefit from ISO invariance. Your only option for darker scenes is to increase your ISO and if you are afraid to increase it high enough to essentially expose to the right, getting the histogram data as close to the right shoulder as possible, then your shadows are going to get noisy.

Use Highlight Warnings

Another thing that I want to add here, is that it’s also important to turn on and use the Highlight Warnings or “blinkies” in your camera so that you are made aware when small areas of the scene or subject start to become over-exposed.

Himba Smile
Himba Smile

For example, when photographing the dark-skinned Himba people inside their huts in Namibia, their eyes, teeth, and shells around their neck start to overexpose as you increase the ISO, but it’s virtually impossible to see this on the histogram. You have to check for these areas blinking in the preview image, or if you are using a mirrorless camera, some of them have these warnings in the electronic viewfinder too.

Some people make the mistake of trying to get the peak or hump of the histogram over to the right side, but when shooting images like this one of the young Himba Girl (left), that will result in her eyes, teeth, and regalia all becoming overexposed, and although just a little is fine and controllable, we don’t want to overexpose these areas too much.

To be clear though, the goal is to increase the exposure to the point where the brightest parts of the scene or subject are as close to the right as possible, but not overexposed, and sometimes, this requires us to increase the ISO so much that some people start to back off.

This image was shot at ISO 5000, and that scares some people, but because I was still essentially exposing to the right, until the bright areas started to overexpose, the shadows still have almost no grain to speak of, and this is with a 50 megapixel camera, which was supposed to be really bad in low light. The reality is that even with such high-resolution cameras, if you push the ISO to the point that you need it to be at to give you an ETTR image, then the shadow areas will still be clean enough to give you a useable image.

You might also recall that this is one of the ten images that I printed at 44 x 62 inches for display in an exhibition at Canon’s Headquarters here in Tokyo, and the feedback I’ve received is that people were very surprised that this image was shot at such a high ISO, and they were looking at a five foot tall print, so I think that’s proof enough that my technique has helped me to create images of a pretty high quality, often in somewhat challenging environments.

The "Namibia" Wall
The “Namibia” Wall

Manual Exposure

This may be obvious already, and I certainly mention this a lot, but just to be thorough, I should also mention that I shoot almost exclusively in Manual Exposure mode. I just find this easier, as it enables me to get my camera set up for a specific scene, exposing to the right, and then just shoot, without having to worry about exposure compensation, which I find completely annoying. 

I know that some people like to just let the camera decide, and if your scene is bright enough that ISO Invariance will help you to brighten up your images as necessary without losing any image quality, you really won’t be negatively affected by shooting in an automated mode. For me though, I’ve been shooting Manual for so long, that it’s just more intuitive, and second nature.

The Manuals Were Wrong

Another thing that I wanted to talk about, with regards to using ETTR techniques to get better quality images, is that sometimes people quote the camera manuals, that often say that a good histogram has one large hump, and that should be in the middle of the histogram. Forget this if it’s something you tend to bear in mind, for two reasons. Firstly, having the data in the middle of the histogram may result in unnecessary noise in your images, especially in the shadow or dark areas. 

The other reason is that the histogram represents a mapping of all the tones in the image. It is not uncommon for a histogram to have multiple spikes. This image, for example, has a spike on the right side, that represents all the white in the crane’s and the snow, and a spike of the left, for the dark background. And there is nothing in the middle, which is where the camera manuals will have you believe the data should be.

Cranes Taking Flight (with Histogram)
Cranes Taking Flight (with Histogram)

Of course, as a guideline, for an average scene, it’s not a complete failure as an example, but it’s important to understand what the histogram represents so that it can be used as a tool to work towards shooting images of higher quality, rather than allowing something you’ve been told to lead you to make mistakes.

OK, so the last thing I want to relay is that it is also not necessary to try and have your image data fill the histogram. Some scenes do not contain a wide enough range of tones to create a full histogram from the left to right side. As an extreme example, and to make one last point, here (below) is another snow scene from Hokkaido.

See if you can guess where the histogram spike will be for this image, then drag that vertical bar over to the left to reveal the embedded histogram and see how close you were. 

Conclusion

Let’s start to wrap this up now, with a few last words on why I put this post together. I kept Evan’s part about developing a style in the title, even though we didn’t go into detail on this, because I think that the way I shoot plays a large part in the look of my work. I don’t I have a distinctive style that comes from a type of processing, because as I’ve said, most of the time it isn’t the processing that defines my work. That in itself is, to me, one of the most important things that I wanted to relay today. 

I’ve been told that people can tell my work, because of the quality of the images, and I like to think that this is because of the care I take in setting my exposure, and how careful I am to compose my images in such a way that they contain as few distracting elements as possible. A lot of my work is somewhat if not straight on minimalist, but rarely cluttered, and I think these things all inform my style.

I’ll stop writing for now though, and get this out there. I’ve already been writing for a day and a half, but there is still so much more to say related to this topic. I’ll try to follow up next week with a discussion of some of the things that I consider as I decide on my compositions, as we didn’t really get into that today. If there is anything else that you’d like me to cover, just drop me a line or write a comment below. I always enjoy hearing from you.


Show Notes

My post on ISO Invariance is here: https://mbp.ac/520

Music by Martin Bailey


Audio

Subscribe in iTunes for Enhanced Podcasts delivered automatically to your computer.

Download this Podcast in MP3 format (Audio Only).

Download this Podcast in Enhanced Podcast M4A format. This requires Apple iTunes or Quicktime to view/listen.


ISO Invariance in the Canon EOS 5Ds R (Podcast 520)

ISO Invariance in the Canon EOS 5Ds R (Podcast 520)

Over the last six months or so, many people have asked for my opinion on ISO Invariance, and I didn’t really have one, because I hadn’t done any tests for myself, so I corrected that, and today will share my results.

Let’s start with a bit of information about ISO Invariance, as this is perhaps a new term to some people. I first started to hear this term when the Sony Alpha 7R II camera was released. This camera has a very wide dynamic range, meaning it can capture a wider range of tones from total black to full white than most other cameras.

We measure dynamic range in stops, which are the measurements used to describe increases or decreases in exposure. We might talk about making our shutter speed one stop slower for example, which could be something like changing it from 1/250 of a second, to 1/125 of a second. One stop faster would mean changing 1/250 of a second to 1/500 of a second.

In terms of aperture, one stop smaller or slower than f/8 would be f/11, and one stop wider or faster than f/8 would be f/5.6. We can also use ISO to change the exposure, for example, making the sensor one stop more sensitive, by changing ISO 100 to ISO 200. We’ll use the ISO for most of our tests that I’ll share with you shortly.

DxO Mark puts the Sony Alpha 7R II at 13.9 stops of dynamic range. The Nikon D810 which also uses a Sony sensor beats that with a huge dynamic range of 14.8 stops, and my Canon 5Ds R comes in at just 12.4 stops in their tests. Compared to just a few years ago, all of these cameras are capturing enough dynamic range to give us a lot of freedom in our photography, although obviously to varying degrees.

The wide dynamic range of the Sony and Nikon cameras though raised the question of ISO Invariance and whether using Expose to the Right techniques that I use is still necessary. Exposing to the Right or ETTR is basically where we adjust our exposure for any given scene so that the highlights are on the right side of the histogram.

Basically you will see more noise in the mid-tones of a photograph compared to the lighter areas, and the shadows are even more noisy than the mid-tones. This means that we can create cleaner images by exposing them with the information in the scene as close to the less noisy right side of the histogram as possible, even if we then reduce the brightness of the image in post processing later.

In a great article published on DPReview called “Sony Alpha 7R II: Real-World ISO Invariance Study” they share some test results which show that because the noise floor is so low in the Sony Alpha 7R II and Nikon D810 with its Sony sensor, it can actually be beneficial to keep your ISO down at 100 or 200, and photograph your scene much darker, and brighten it in post. This is basically the opposite to Exposing to the Right.

The idea is that all we are doing when we increase the ISO is making the pixels more sensitive to light or amplifying the signal, but with such high dynamic range on these cameras, the noise added by increasing the ISO in camera is comparable to the noise added by pushing the exposure in post processing, but with the added benefit in the latter, of stopping highlight areas from blowing out or blooming.

This got me curious, as I know that my 5Ds R does not have such a wide dynamic range as these Sony sensor cameras, so I did a test to see if Canon EOS 5Ds users could benefit from this technique, or if it was better to continue to use ETTR techniques. Here are my results…

First of all, I shot six frames of a guitar, in my studio, with just light coming in through the lace curtain in my window. The guitar is black and shiny, so you can see the reflection of the curtains, and it has some chrome on there too, which also reflected light. I adjusted my exposure at ISO 100, so that my resulting photograph was just starting to blow out the reflection of the window in the chrome and a little bit in the reflection of the curtain on the shiny black guitar. This is probably how I would expose this photograph if I was exposing to the right (below).

ISO 100, f/8 for 1 second

ISO 100, f/8 for 1 second

For this exposure, my shutter speed was down to 1 second at f/8, with ISO 100. In Lightroom, I can adjust the Exposure of my images up to +5 stops, buy increasing the Exposure slider all the way over to the right. So, I adjusted my 1 second exposure at ISO 100, to the same exposure at ISO 3200, which is 5 stops more sensitive than ISO 100. Just double the numbers five times to check for yourself. (100 -> 200 -> 400 -> 800 -> 1600 -> 3200)

Because I was going to be increasing the sensitivity of my sensor by five stops, I also needed to decrease my shutter speed by five stops to start my test. So that would be 1 second halved to, 0.5 seconds, then 1/4 sec, 1/8 sec, 1/15 sec to 1/30 of a second. I then shot a series of six images, starting at ISO 3200 at 1/30 of a second, and continued to shoot five more frames, decreasing the ISO by one stop for each image. In this screenshot from Lightroom (below) you can see the ISO 3200 image in the top left, going down to the ISO 100 image in the bottom right.

ISO 100 to 3200

ISO 100 to 3200

Next, I went into the Develop module in Lightroom, and increased the Exposure slider for the images from ISO 1600 to 100. I increased ISO 1600 by 1 stop, making it the same brightness as the ISO 3200 image. I increased the ISO 800 image by two stops, the ISO 400 image by three stops, the ISO 200 image by four stops, and the ISO 100 image by five stops, making them all look the same brightness, as you can see in this screenshot (below). I literally just clicked the number to the right of the Exposure slider in Lightroom, and typed in 1, 2, 3, 4 and 5 against each image.

 ISO Invariance Tests

ISO Invariance Tests

As you can see, in the thumbnail view, it’s hard to tell the difference between the ISO 3200 image in the top left, and the ISO 100 image in the bottom right, with its +5 stops of Exposure applied in Lightroom. Let’s take a look at the amount of noise that is added to these images though, as we increased the ISO. Click on the image on the top left, and then navigate back and forth through these images with your mouse or arrow keys on your keyboard. Note that the images might automatically advance, but to stop that, just place your mouse over the image.

For me, I can see no real difference in the amount of noise between ISO 3200 and ISO 1600 pushed by 1 stop, and ISO 800 pushed by 2 stops. From the ISO 400 image pushed by 3 stops I can start to see a little bit more noise, and ISO 200 pushed by 4 stops gets quite bad, and ISO 100 pushed by 5 stops is really noisy.

What does this tell us?

So, what does this tell us? Well, whereas the Sony sensor cameras are showing results proving that you can push images up to 5 or 6 stops without seeing very much noise added, based on this test alone, it would seem that with the Canon 5Ds R, you could push your images by up to 3 stops without degrading the quality of the image, assuming that there was some reason for you to be using a high ISO in the first place.

For example, imagine I wanted to photograph a bird in flight in low light, which would require that I increase my ISO to get a faster shutter speed, and there was a lot of dark areas in the scene, I could shoot at say 1/1000 of a second at ISO 100, then push my exposure in post to +3 or +4 at a push, and not see any more noise than I would have seen if I shot the same scene at ISO 800 or 1600.

Keep in mind at this point, that the amount of noise that you would see at ISO 800 or 1600 is minimal anyway, and I’m not saying that either of these methods is better than the other, but it’s one technique that we could keep in our digital toolbox in case it becomes useful at some point.

High and Low ISO Comparison

I’m going to state the obvious here before we move on, but if however, you don’t need to worry about shutter speed, and can do a longer exposure, it is always going to be better to shoot at a lower ISO and increase the length of your shutter speed, for cleaner images.

Here’s a diagonal splice of two images to illustrate (below). The top left triangle is ISO 100 for 1 second, and the bottom right triangle is ISO 1600 at 1/30 of a second. Don’t forget to open up your browser window and click on the image to view it at 100% to see the detail. The version that is embedded in the blog post has been reduced in size a little.

ISO 100 for 1 sec and ISO 1600 for 1/30 sec 100% Crop

ISO 100 for 1 sec and ISO 1600 for 1/30 sec 100% Crop

From this, you can see that there is a certain amount of grain in the image, even at ISO 1600, although ISO 3200 looks very similar to this. We can even see a little bit more grain in the highlights on the chrome, and there is really no blooming to be seen in the ISO 100 image, so personally, I’d always go for the ISO 100 shot when shutter speed is not an issue.

White on White

The next question I asked myself was, does this mean that my use of the Expose to the Right technique with the 5Ds R and possibly my earlier Canon cameras a total waste of time?

The first test I did was a predominantly dark scene, with a few areas of highlights. For something like my Japanese Red-Crowned Cranes or Whooper Swans, on a white snowy background, we’re talking mostly white on white, with some dark patches on the animals. I don’t have a Red-Crowned Crane handy, or a field full of snow for that matter, but I do have this almost white cat ornament with some darks spots, and a roll of seamless, so I used these things for my next test.

Here you can see a screenshot from Lightroom of a photo of the white cat on the white background, shot in Aperture Priority mode, with Exposure Compensation set to zero on the left, and Exposure Compensation set to + 1 1/3 of a stop on the right (below).

Invariance Test White Cat on White Background

Invariance Test White Cat on White Background

The cat has a little bit more reflectivity than the cranes usually have, but I stopped increasing the exposure at 1 1/3 of a stop, to prevent a patch of his belly from blowing out. Usually for a snow scene, if it was overcast/shade, like this, I would probably be a third or two-thirds more than this, but let’s continue with this example, as it is enough to draw some conclusions. It also represents a real world scenario, where I might leave the camera in Aperture Priority mode at zero Exposure Compensation, then increase the exposure later.

Click on the top left image below and again, with your browse window opened up wide, look at the noise in the dark patch above the cat’s eye. I’m sure you’ll agree that in every case, the image that was shot with 1 1/3 of a stop in Exposure Compensation applied in camera is cleaner than the one shot at zero Exposure Compensation, then pushed by 1 1/3 of a stop in Lightroom. Also, you’ll see that as the ISO is increased, pushing the image by 1 1/3 of a stop in post introduces a lot more grain than you’d see as opposed to increasing the exposure in camera.

For me, this test shows that for lighter subjects, you definitely get cleaner images by adjusting the exposure, and essentially exposing to the right in camera, than you do by leaving exposure to the camera and then lightening them later. This isn’t so noticeable at ISO 100, so again, ISO invariance is at play here. The difference is less noticeable and probably more acceptable at ISO 100 than it is at the higher ISOs.

Average Scene Test

Finally, I wanted to see what happened when I photographed a scene with a variety of tones and colors that pretty much average out to a zero compensation exposure, so I shot a series of photographs of an X-Rite Digital ColorChecker SG card, at ISO 100, 400, 1600 and 6400. I then adjusted the shutter speed, making three more frames at each ISO which were minus 1, 2 and 3 stops, as you can see here (below).

Average Scene Tests with X-Rite ColorChecker SG

Average Scene Tests with X-Rite ColorChecker SG

Once I had these images in Lightroom, I increased the Exposure of the under exposed images by +1, +2 and +3 with the Exposure slider, making them all the same brightness. Here is a 100% crop from each image, and again, you’ll need to click on these with your browser window wide, and navigate back and forth with your mouse or arrow keys to make a comparison.

As you can see, once again, at ISO 100, you can push the image by 2 stops, even 3 if necessary, and really see very little degradation in the quality of the image. At ISO 400, we’re probably talking 2 stops, and at ISO 1600, even pushing the image 1 stop introduces a lot of grain, and at ISO 6400 pushing the image at all pretty much ruins the image. This is in line with our earlier findings though. The invariance is only really valid in the high ISOs and quickly degrades as we increase the ISO.

Conclusion

OK, so what conclusions can we draw from this? Well, with the Canon EOS 5Ds R and probably the 5Ds as well as the 5D Mark III which all had very similar dynamic range, I think we can safely say that if you are shooting in a situation where ISO 100 gets you to within 3 stops of where your exposure needs to be for an average tone or dark scene, then shooting without any kind of exposure compensation, either in an automated mode or in Manual, and  then increasing the Exposure in post, you will not really see any degradation in the quality of your images.

If you are shooting a brighter scene, at least from these tests, ISO 100 does still show a slight increase in noise in the dark areas if you push in post as opposed to increasing the exposure in camera. For this reason, it’s probably still better to expose to the right, and get your whites white, when shooting winter scenes such as a white bird on white snow. Having said that, the amount of degradation is negligible at ISO 100, and if it’s a toss-up between fighting with exposure to avoid blowing out highlights, and under exposing just a little bit, this certainly does mean that we can give ourselves a little bit more wiggle-room, even for white on white winter scenes.

I can see me perhaps at least trying only going over by +1 stop, and maybe even trying zero compensation in next year’s winter tours, to at least see how the image looks with a true, real-world example. I expect that at ISO 100 this will be a valid way to shoot, as we only need to push the image up by 2 stops. It may also be fine up to ISO 400, if my tests above are anything to go by. Below ISO 400, and you really don’t want to be pushing your images in post.

I still need to get my head around applications of this in the field, but I’d say that on the Canon 5Ds R, if ISO 100 can get you to within 3 stops of the shutter speed that you need to freeze a subject, this is now a viable shooting technique. Allow the image in camera to fall dark, and increase the exposure in post. As I mentioned earlier though, the resulting images don’t look a lot different to those that you will get by increasing the ISO, so the benefits to Canon shooters at this point in time, are really minimal in terms of images quality, but if this technique helps you to avoid blowing out highlights or adjust for a certain shutter speed, it’s worth bearing in mind.

Let’s also bear in mind that during these tests, I have been looking at my images while zoomed in at 100% on 50 megapixel files. This is the level of detail that I like to work at, because I sometimes print large or need to crop my images, but if you are only shooting for web, or low resolution images, the differences that we can see will be less important to you. I’d recommend you think through this yourself or do some similar tests to make up your own mind how useful relying on ISO Invariance could be for your own shooting workflow.

If you are shooting with a Nikon D810 or the Sony Alpha 7R II, and this is the first you’ve heard of ISO Invariance, do check out that article on DPReview, as you may well be able to benefit from this more than us Canon users at this point in time. Either way, I hope this episode has been useful, and helped to shed some light on the topic of ISO Invariance. I’ll continue to update you through the blog and podcast as implement at least some parts of what I’ve found in my own shooting.

The Complete Namibia Tour & Workshop 2017

Before we finish, I’d like to mention that we have just started taking bookings for an incredible 17 day tour in Namibia from June 3 to 19, 2017. There are just a few spaces left, so they might already be gone by the time you see this post, but check out the tour page at https://mbp.ac/namibia and sign up if you’d like to join us. If it is already sold out, please contact us to be placed on the cancellation list.

Complete Namibia Tour 2017

Click for Details

 


Show Notes

DPReview Article on ISO Invariance on the Sony Alpha 7R II: http://www.dpreview.com/articles/7450523388/sony-alpha-7r-ii-real-world-iso-invariance-study

See details of my 2017 Namibia Tour here: https://mbp.ac/namibia

Music by Martin Bailey


Audio

Subscribe in iTunes for Enhanced Podcasts delivered automatically to your computer.

Download this Podcast in MP3 format (Audio Only).

Download this Podcast in Enhanced Podcast M4A format. This requires Apple iTunes or Quicktime to view/listen.


Why Expose to the Right? (Podcast 381)

Why Expose to the Right? (Podcast 381)

Today I’m going to go into detail on why I expose to the right, as I get asked about this a lot. Exposing to the right means adjusting your exposure so that the image data we see in the histogram falls as close to the right shoulder as possible. I’ve touched on this a number of times, but never really gone into detail about the reason, so here we go.

Actually, before we jump into this, I found out last week that images in this Enhanced Podcast not being displayed in iTunes 11 is not a bug, but the way Apple have decided it’s going to be. You have to hold down the Command key on a Mac or CTRL key on a Windows machine, then click the artwork in the little window at the top of iTunes. This will open the Enhanced Podcast viewer, so you can follow along with the images. I’ve also put a page together to help you understand this at mbp.ac/ituneshelp, so do take a look if you’ve been missing the images.

Tanchou Study #7

Tanchou Study #7

Firstly, let’s look at why this helps you to create cleaner images, starting with a little background. When I first started shooting with a Digital SLR camera, some 12 years ago now, it was common for people to slightly underexpose the images to give them a little punch. It seemed to make the colors more vibrant. The problem with that, as we’ll see, is that we were inadvertently introducing more noise to the images.

After a few years I started to travel to Hokkaido where I now run my Winter Wonderland tours, and before very long I started to shoot in Manual mode to ensure that I was getting well exposed snow. As you know, camera’s meters still try to expose everything as an 18% or neutral grey, which means if you shoot a snow scene, the camera will under-expose the photo by around two stops of exposure, making the snow look like a muddy grey.

You can use exposure compensation, adding two stops to counter this, but then if the predominantly white Red-Crowned Cranes that I shoot up there move from their white background to a dark background, you’d have to change your exposure compensation from plus two stop to minus two stops to compensate for the now dark background.

For example, one moment I could be photographing something like this photo (right) of a white crane on a white background, and the next moment I could spot another crane flying into the area, on a dark background.

These next two images are screenshots from Lighroom, to show you the histogram as well. You can see that this first image (below left) of a crane on a dark background is predominantly dark, but with a small spike to the right of the histogram representing the white bird. You can also see a gap between the bird and the right shoulder of the histogram, but this is Lightroom giving me an extra stop of exposure to play with. In the camera, this will have been just touching the right side of the histogram. (Click on the images to enlarge for a closer look.)

Soft Arched Wings in Lightroom

“Soft Arched Wings” in Lightroom

Soft Arched Wings 4 Stops Overexposed

“Soft Arched Wings” 4 Stops Overexposed

Had I simply raised the camera from the shot of the white crane on a white background though, still with plus two stops of exposure compensation dialed-in, the camera would have seen the dark background and brightened it up to a mid-grey, probably to the tune of around 4 stops, because you need to add 2 stops for white, or reduce by around 2 stops for black. I simulated this in Lightroom by increasing the Exposure by 4 stops, as we can see in this shot (above right).

See here too how the data on the histogram is mainly in the middle, but with the white bird spiking up the right side of the histogram? The data in the middle is what should be the black background, and that should be way over on the left, with the white crane on the right, but not touching the right side of the histogram. If you see this in the field, without knowing what you are overexposing, it’s time to pull your exposure back until the histogram is just before or just touching the right side.

Of course, a crane against a blue sky would be a different setting again, and the size of the crane in the shot changes how much exposure compensation we need too, so you end up fighting a losing battle, or bracketing. Unless you are shooting for HDRs, bracketing just shows a lack of understanding on exposure. I was initially told this by one of my mentors, a Japanese photographer called Hiroshi Yokoyama, a wonderful elderly gentlemen that used to be an Olympic winter sports photographer, and now lives in Hokkaido fulfilling his life’s work. I also heard Ansel Adams say this on a documentary too, so it’s not just me being outspoken here. Well, maybe not…

Why is this Important?

Exposing for whites aside, you probably are still wondering why I’m such a stickler for exposing to the right, even for scenes of average brightness and tonality. You’d be forgiven for thinking that if the scene is similar to a mid-grey, you could leave the exposure up to the camera, but that is not the case. I started exposing to the right intuitively, but after a few years, someone pointed out a great article on Michael Reichmann’s Luminous Landscape web site about Optimizing Exposure. This was actually a follow up from a 2003 essay called Expose Right, and both are well worth a read. The biggest take-away is how cameras distribute the image data when the image is saved to your memory card, which we’ll look at here.

Camera sensors have to convert the light that is captured by the photo-diodes on your sensor to a digital value. Most cameras these days save this data for each pixel with 14 bits of data, meaning we can have up to 65,536 tonal values per pixel. Many cameras at the moment record about 12 stops of dynamic range, some more, some less, and this is the range between a true black and a true white, and with a 14bit sensor, this means we’d have 65,536 tonal values between the two extremes.

Think of this as gradually filling a bucket with water, or even filling a photo-diode on your sensor with light. A totally empty bucket would be zero, or totally black, and then as you pour water into the bucket, or light into the photo-diode, you gradually fill it until you hit the maximum it can hold, which would represent a pure white. Anything after that is just overflow. The white can’t get any whiter, and that pixel is now over-exposed.

Because digital sensors are linear devices, if you double the amount of light that hits the photo-diode, you double the voltage generated by the sensor, or the amount of water that we pour into the bucket. The result is that data is literally halved with each stop or EV (Exposure Value) that is recorded, so the brightest stop of light the sensor can record has to be double that of the second brightest. This means that to use the full 65,536 tonal values across the entire image, we have to half the amount of data that can be used for each exposure value from pure black to pure white. For a 12 stop dynamic range sensor, this means your data is broken up into 12 steps, as follows…

  Darkest Mid Tones Brightest
EV  1  2  3  4  5 6 7 8 9 10 11 12
# Tones 16 32 64 128 256 512 1,028 2,048 4,096 8,192 16,384 32,768

This is why the darker parts of the image are noisier than the brighter parts. The noise is always there, but because we have less data to record the darker areas of our image, the noise is much more visible. I’m sure you’ve noticed that even a nicely exposed image often has a bit of noise in the shadows. This is why.

Exposing an Average Scene

We’ve looked at a white scene, or a dark background with a white subject, and for those scenes, it’s easy to understand why I set the exposure as I did. But what if you shoot a mid-tonal scene? Of course, if you leave the exposure up to the camera, it will record the scene in the middle of the histogram, so you will essentially recording your image with between say 256 and 2,048 tonal values per stop as opposed to between 4,096 and a massive 32,768 tonal values per stop in the brightest four stops of the image, and for a 12 stop dynamic range camera, this would be what is represented by the right-most third of the histogram. Here’s an example of such a histogram, from this shot of a Springbok in Namibia earlier this year (below).

Springbok Shot in Lightroom

Springbok Shot in Lightroom

See how the histogram data is over in the right third, again though noting that Lightroom is giving me an extra stop here. In the camera, the histogram was right up to the right shoulder. This of course means that despite me shooting this image at ISO 2500, even when viewed at 100% there is very little noise, as we see here (below).

Springbok at 100%

ISO 2500 Springbok photo at 100%

In an extreme case, you might find yourself shooting a very dark subject against a dark background. In the film days, you’d have dialed in maybe a stop or even two stops of negative exposure compensation, to ensure that the scene was recorded naturally dark, remembering that the camera wants to brighten it up to an 18% grey.

Darken Down in Post

To ensure that we record our images with as wide a range of tonal values and as little noise as possible, you would even expose a very dark scene to the right, and then reduce the Exposure of your image in your post processing workflow. I don’t have any examples this extreme, although many of the sand-dune photos from Namibia this year have had the exposure dialed down to some degree. My favorite shot of the camel thorn trees against the sunlit sand dunes at Deadvlei for example was dialed back by half a stop of exposure in Lightroom, to what we see here (below).

Deadvlei Silhouettes

Deadvlei Silhouettes

This was according to plan of course. I get the highest quality image possible by exposing to the right, then darken it down in post as much as necessary. Of course, I’m aware that many people think my images are high-key anyway, but that’s because I have my monitor brightness set very low, as part of the calibration process. Many people don’t do this, and that’s why my images seem bright.

The sad fact is though that most people’s images are actually under-exposed because they trust the camera too much, but they never know because they view them on a monitor with the brightness turned up full. Then they print and wonder why the prints are too dark. If this sounds familiar to you, I hope a few pieces of the puzzle are starting to drop into place here.

A New Exposure Mode

Part of the reason I decided to talk about this today was because of a conversation on This Week in Photo a few weeks ago. I’d mentioned shooting to the right, which raised some listener questions. After talking a little about shooting to the right though, we went on to talk about new features that we’d like to see in cameras, and I said that I’d love to see a setting on the camera that allowed you to just keep exposing the image until it starts to blow out, or over-expose.

Ideally we’d be able to set a parameter to tell the camera how much of the scene gets over exposed. For example we might want to stop at 0%, so there are no specular highlights overexposed at all. Or we could set it to say 1%, 3% or 5% etc. as we can now start to blow out the image in the JPEG preview on the camera, but Lightroom gets more detail from the RAW file, so you can go over a little without worrying about it too much.

There are times such as when I’m shooting the snow monkeys, when I actually blow out probably up to 30% or 40% of the scene, because I’m looking for a well exposed monkey, and don’t care too much about the background. It’s often out of focus anyway, and the slice of snow on the same focus plane as the monkeys has a bit of texture, which reduces it’s luminosity to the point that is doesn’t overexpose anyway.

You have to be careful doing this of course, as too much over-exposure can bleed or bloom into the darker areas, messing them up a bit, but for the snow monkeys this doesn’t really happen due to their fur, kind of protecting the highlights around them.

Reverence

Reverence

What I could see happening is a kind of hybrid Aperture Priority mode, where I tell the camera what aperture I want to use, as that’s my most important setting, but then I also want to control how slow the shutter speed gets, so Auto-ISO would also need to be turned on. I’m already doing this for some of my photography, although I still shoot in Manual mode more than 99% of the time, because I lose too much control in Aperture Priority and exposure compensation drives me crazy.

When using Auto-ISO the camera bases the shutter speed on the focal length, but you can set a minimum shutter speed as well, which is necessary with wildlife as the animals can move around so much that you need a fast enough shutter speed to avoid subject blur. This needs to be opened up a bit though. On my 5D Mark III 1/250 of a second if the fastest shutter speed I can select. This would be fine for the snow monkeys most of the time, but I need faster shutter speed for birds in flight. Ideally I’d like to be able to set this to any shutter speed the camera can use, though I’d probably set it to between 1/500 and 1/1000 of a second for birds in flight, depending on the size of the bird and how much wing movement I want.

All of this though is only relevant if I had the ability to tell the camera to keep on exposing until a certain percentage of the scene got blown out, so I imagine the algorithms would be somewhat complicated, but we already have the ability to create a live histogram in Live View, so the camera could adjust the Shutter Speed, ISO or even Aperture, depending on our settings, to achieve the optimal exposure, which we can already see in the histogram.

This would of course become more difficult for TTL or Through The Lens photography, because there’s no light hitting the sensor like there is in Live View. Once you start to expose the image though, light is hitting the sensor, so assuming I was in Aperture Priority mode, I imagine the camera would need to change the ISO as well as the shutter speed dynamically as it exposed the image, all in as fast a time as 1/1000 of a second.

Stopping the exposure as the image starts to over-expose is already possible with some flash photography though, and that is working at similar speeds, so I know it’s not impossible to do. It’s really just a case of the camera manufacturers spending the R&D dollars or more likely yen, to figure this out for regular exposure and get it implemented. I see a time though when we will expose all scenes to the right for the best quality image, and then the camera might even include a bit of information on the actual tonal values in the scene, and normalize it in post, or even in the camera, but without throwing away all the tonal values that it gained by exposing the image to the right.

Use the RGB Histogram

So, if you don’t already expose to the right, but I’ve convinced you to give it a try, there are a few things that you’ll need to bear in mind as you work. Firstly, if your camera has the ability to display an RGB histogram, turn it on. The standard grey histograms are displaying an average of the Red, Green and Blue channels, and this can give you a false sense of security. Depending on your subject, one of the channels can start to blow out before the others, and it’s sometimes only possible to see this with an RGB histogram.

Don’t Fear High ISOs

Also, if you are going to shoot dark scenes with up to three or four stops brighter exposure, do keep your eye on your shutter speeds. You may need to increase your ISO to achieve a fast enough shutter speed for the scene in front of you.

This shot of a cheetah for example is not a dark scene as such, but it was dark when I shot this. Just minutes before we could only see this cheetah with the lights of our safari vehicle, and there was just enough light from the sun, still below the horizon, for me to be able to get this shot by cranking up my ISO to 12800.

Cheetah at ISO 12800

Cheetah at ISO 12800

This gave me a 1/125 of a second exposure, which was as low as I dare go, but I was standing on my seat with my 300mm and 2X extender fitted for a 600mm focal length, but the lens was resting on my hand, and so relatively stable. The cheetah wasn’t moving either, so I got away with this shutter speed. As we can see from this 100% crop though, sure, there’s noise, but it’s totally acceptable in my opinion. What most people instinctively do though in low light, is to be too afraid to increase the ISO and shoot the scene darker, expecting to brighten it up in post.

Cheetah at ISO12800 100% Crop

Cheetah at ISO 12800 100% Crop

The irony here of course is that if you shot this at say ISO 3200, two stops darker, you’d record the image with fewer tonal values per stop of light, making it much noisier, then when you increase the exposure in post, you amplify the noise, giving you a total piece of crap of an image. So, the long and short of it is, if you want to capture the best quality images that you can, shoot to the right, even if you then dial the exposure down in software later.


Show Notes

Luminous Landscape Essay:
Optimizing Exposure: http://www.luminous-landscape.com/tutorials/optimizing_exposure.shtml
Expose Right: http://www.luminous-landscape.com/tutorials/expose-right.shtml

Music by UniqueTracks


Audio

Subscribe in iTunesSubscribe in iTunes for Enhanced Podcasts delivered automatically to your computer.

Download this Podcast in MP3 format (Audio Only).

Download this Podcast in Enhanced Podcast M4A format. This requires Apple iTunes or Quicktime to view/listen.