What I learned from Gollywop -- and what I wonder

Started Mar 26, 2013 | Discussions thread
Macx
Senior MemberPosts: 1,347Gear list
Like?
Re: What I learned from Gollywop -- and what I wonder
In reply to richarddd, Mar 27, 2013

richarddd wrote:

Macx wrote:

richarddd wrote:

Anders W wrote:

richarddd wrote:

Anders W wrote:

Practice makes perfect though. That's why it is important to think and talk about it. Some people are wont to say that "theoretical" discussions of the present kind are of no help in the field. I think it's exactly the other way. Only by thinking and talking about it will you eventually know exactly what to do in the field, and do so intuitively and instantly. For example, I know "without thinking" when it's the proper time to switch from ISO 200 to ISO 400 on my E-M5 when I am out taking pictures. But I do so only because I have exercised my thinking about the matter quite a bit beforehand.

There are two ways to figure out how to proceed in the field (1) have a good set of rules telling you how to behave and (2) have a good set of underlying principles from which you can derive such rules.  They are clearly not mutually exclusive.

As an illustration, please summarize when you would and when you would not increase from ISO 200 to 400 on your E-M5. For example, are there cases in which you would increase ISO to 400 even though you could increase exposure without sacrificing desired DOF or blur control, or cases when you would not increase ISO to 400 even though you are at maximum exposure consistent with desired DOF and blur control and could increase ISO without clipping highlights with detail you would like to preserve?

My answer to the two questions at the end would be no. So I think you have already indirectly summarized what the rule is: If you can't reach the highlight clipping point at ISO 200 without exceeding your f-stop and shutter speed requirements, go to ISO 400.

If, on the other hand, we have the same scenario but a choice between ISO 1600 and ISO 3200, I would normally prefer to remain at ISO 1600 since the read-noise gain from going higher would be close to zero and staying at 1600 means that I have to worry less about any highlight clipping

All of which appears to lead us back to a rather simple set of rules for the E-M5:

Start at base ISO

Set aperture and shutter speed to get the most light on the sensor, consistent with desired DOF and lack of camera or subject motion blur, so long as we don't clip highlights in which we wish to preserve detail

If we are not clipping desired highlight details, increase ISO until just before clipping point. Skip values between 200 and 400 and stop at 1600.

If we can't hit targets (e.g., desired DOF or fast enough shutter), then decide on best compromise, which may involve bracketing.

Use orange blinkies to determine clipping.  Alternatively, spot meter on brightest point with detail, then increase exposure (and then ISO) by 3.3 stops.

Post-process to taste.

Have I missed something?

Not really, but you bring up bracketing, and I suppose the next thing is how to best use multiple exposures to increase the dynamic range.

If we're lacking shadow detail, bracketing is probably not the best way to go to increase dynamic range, compared to simply doing multiple normal exposures, following the rule set you wrote above and then stacking them.

Bracketing use the old way of thinking, where we exposed for a target brightness in the final image, but when we shoot for raw, we're thinking about exposing for the least amount of noise, because that is what gives us the best image quality when we're done.

The lack of shadow detail is, unsurprisingly, mainly because of the lack of light. A normal bracketing procedure would be to take our baseline exposure and then make a longer exposure specifically for the shadows. This method works, but the exposure for the shadows obviously clips our highlight so it does nothing for the highlight details.

If we instead make multiple normal exposures and stack them, we'll both be giving the shadows the extra exposure they need to preserve detail, but we'll also give the highlights more exposure than we could before and thereby we'll also reduce whatever noise there might be there. By stacking exposures we're in effect deepening the sensor wells allowing for much more light before hitting the clipping point.

In other words, if presented with the option of either shooting a baseline, normal, ettr exposure with a shutter time of 1 seconds, and then "bracketing" by shooting a 4 second capture for the shadows, we're better off by shooting four or five 1 second exposures and then stacking them.

Which leads to a second point. If we're in a position where we can use multiple exposures to simulate a long exposure like above, there might be an advantage in using a higher ("true") ISO setting if it reduces read noise, for example ISO 800 on the E-M5, if the combined shutter time is still within our limits. We would still need to make sure that each exposure didn't clip the highlight details, but the main limitation of the higher ISO which is the shallower effective well, would be eliminated, and the main advantage in the form of the lower read noise would be retained.

This is obviously only really interesting for people shooting with tripods and shooting stationary scenes. For the most part a single exposure is of course the only feasible option.

Thoughts?

Anders, DM and I recently had a discussion of this issue.

Anders pointed out that if you bracket, you will get an image with more exposure in the shadows and therefore lower noise. If you can combine this image with the ETTR image, you can get the lower shadow noise from the shadow exposure image and the rest of the picture from the ETTR image.  This should result in lower noise than averaging stacked images.

Hmmm? But when talking about stacking, I'm not talking about averaging, I'm talking about adding to the sensor data. This way you get the same amount of extra exposure in the shadows that you would by bracketing, but with the added advantage of extra exposure for the rest of the tonal range too. And perhaps with a lower read noise than base iso, too, if you so prefer.

Stacking images should give you at least as good noise performance in the shadows as bracketing.

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
BothNew
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow