A Test on "exposing to the right" and ISO

I shot a Band Concert last Monday night during their rehearsal period. I had my 430ex flash and shot away some close(r) up shots of various musicians. All at ISO 400, f/5.6 to f/8 ... Then I removed the Flash in order to take larger group shots. I changed my c.fn code to "double" my allowable ISO to 3200 ... I had never done this before but this setting was a perfect test for me... as I had no choice in the matter given lack of ambient light. So I just shot away at 3200 ISO and the largest apeture I could get (I think it was like f/3.5 or f/4, but I'd have to check.) and at a SS of 100 - all in Manual mode & RAW of course ... Anyway, I was quite impressed how they turned out given this "test". In fact when I converted them to B&W in PP, the "grainy" look was really only seriously apparent on the drapes and screen behind the band. The Color shots did need some PP, but I was able to adjust them pretty well after all. Yea, I would love to have an f/1.8 lens ... but I don't .... yet.

I'm not sure I'm getting the original intent of the test. On the surface it seems to me the real result is that proper exposure from the beginning is best, not so much any differences between ISOs. Am I missing something?

Is this test saying that exposing to the right is better than a proper exposure? Or that exposing to the right at higher ISO is better than an underexposed picture at a lower ISO?

I might suggest that exposing to the right is proper exposure, sans blinkies.

rc13k wrote in post #5581662
Is this test saying that exposing to the right is better than a proper exposure? Or that exposing to the right at higher ISO is better than an underexposed picture at a lower ISO?

Mainly that exposing to the right with a higher ISO is better than an underexposed lower ISO but you could also conclude that a higher ISO exposed to the right is about the same as a lower ISO exposed properly if you think the 1600 image looks similar to the 800.


I was really just messing around for my own understanding since I used to be so reluctant to go to 1600 ISO or above but I though people might find my results useful.

a_kraker99 wrote in post #5581069
E-K, wouldnt the fact that most of an images dynamic range is stored in the right side of the histogram make the image look a little washed out when bringing the exposure back down? Since so much information is packed on the right hand side, when you bring the exposure down wouldnt it be normal for there to be a little less contrast?

I wouldn't expect much of a difference if your RAW converter's exposure adjustment is doing a simple linear scaling. A -1 stop adjustment would simply divide the values by 2. So you would have a +/- 0.5 DN potential error - not too important to the mid-tones but more of an issue with the shadows.

e-k

rc13k wrote in post #5581662
Is this test saying that exposing to the right is better than a proper exposure? Or that exposing to the right at higher ISO is better than an underexposed picture at a lower ISO?

YES - at an ISO of 1600, I could not obtain a SS of anything faster than about 1/30 sec with a large f/stop (2.8). And they were not "posing" for me ... So bumping to 3200 saved me ... at least I got the shot and it was not blurry at all. the histogram was somewhere around centered, as I recall .. but I'd have to recheck that to verify.

a_kraker99 wrote in post #5581695
Mainly that exposing to the right with a higher ISO is better than an underexposed lower ISO but you could also conclude that a higher ISO exposed to the right is about the same as a lower ISO exposed properly if you think the 1600 image looks similar to the 800.


I was really just messing around for my own understanding since I used to be so reluctant to go to 1600 ISO or above but I though people might find my results useful.

If that's the case, wouldn't that suggest that exposing to the right at ISO 1600 is better than a properly exposed shot at ISO 1600? So what's better? A properly exposed shot at ISO 1600 or one that's slightly over exposed to the right?

Cody21 wrote in post #5581794
YES - at an ISO of 1600, I could not obtain a SS of anything faster than about 1/30 sec with a large f/stop (2.8). And they were not "posing" for me ... So bumping to 3200 saved me ... at least I got the shot and it was not blurry at all. the histogram was somewhere around centered, as I recall .. but I'd have to recheck that to verify.

Yes I knew that already. I was suggesting only IQ not lost shots due to motion blur. Obviously you should use higher ISO to keep a faster shutter.

rc13k wrote in post #5581827
If that's the case, wouldn't that suggest that exposing to the right at ISO 1600 is better than a properly exposed shot at ISO 1600? So what's better? A properly exposed shot at ISO 1600 or one that's slightly over exposed to the right?

Well if a 1600 ISO image exposed to the right looks about the same as a 800 ISO image that is properly exposed I would assume that the 1600 ISO image exposed to the right would definately look better than a properly exposed 1600 ISO image. I will see if I can test that tonight by going 1 and 2 stops higher with the shutter speed at 1600 ISO and bringing the exposures back down.

rc13k wrote in post #5581827
If that's the case, wouldn't that suggest that exposing to the right at ISO 1600 is better than a properly exposed shot at ISO 1600? So what's better? A properly exposed shot at ISO 1600 or one that's slightly over exposed to the right?

When comparing at the same ISO, exposing to the right will be better assuming you:

1. Didn't blow any channels; and
2. Were able to ETTR without any additional ramifications (e.g. driving your shutter speed to low causing motion blur ).

The fuzzy area, at least to me, is whether it is better to ETTR at a higher ISO rather than use a proper exposure at a lower ISO.

e-k

I said it several times, and I'll repeat it - I use ISO 640 on a regular basis when doing flowers.

Noise can be seen if I pixel peep, but for some reason I can't find the noise in the 11 x 14 prints.

In fact, I can't remember when I used ISO 100 or 200.

for an example of ISO 3200 almost exposed to the right

https://photography-on-the.net/forum/showthre​ad.php?t=507784

Cody21 wrote in post #5581794
YES - at an ISO of 1600, I could not obtain a SS of anything faster than about 1/30 sec with a large f/stop (2.8). And they were not "posing" for me ... So bumping to 3200 saved me ... at least I got the shot and it was not blurry at all. the histogram was somewhere around centered, as I recall .. but I'd have to recheck that to verify.

I am still of the opinion that when shooting RAW with a 30/40D you are better of underexposing at ISO 1600 and boosting exposure in the converter rather than setting the camera to ISO 3200. The 'H' ISO is done by software manipulation, so if you up exposure while converting you are doing exactly the same. But the exposure boost done by the camera is always one stop. What if you can get the needed shutter speed with only a 2/3 stop underexposure or 1/2 stop? If I do it myself I can keep the push to a minimum and therefore get less noise. Also, noise is most evident ion the shadows. With a custom tone curve I can push the highlights more than the shadows. I trade off DR and shadow detail for less noise. It depends on the image, but sometimes it's worth it.

I am sure ISO 3200 is great for jpg shooters, but as a RAW shooter it is just another item on the list of things that the 40D has but I never use, like HTP, HINR and 1/3 ISOs.

Palladium, please help me with your reasoning.

Why did you use ISO 3200 during the day with good light?