going Manual!!!!

Methodical wrote in post #9899725
Are you talking about adding or subtracting exposure compensation?

If you point the camera at a scene, in manual mode, and you vary the ISO so that the meter says +2 at ISO 1600, +1 at ISO 800, 0 at ISO 400, -1 at ISO 200, and -2 at ISO 100, then the ISO 1600 will give the least noise, and ISO 100 the most. If the scene has a majority of area as the brightest tones, the 1600 at +2 might be usable.

The point is that increased absolute exposure is what makes ISO 100 images cleanest, not the ISO 100 setting itself, which is the noisiest in manual mode (IOW, with a fixed, given manual exposure). The tradeoff is noise vs highlight headroom.

I will have to give that a try. So at ISO 100 I want to expose to the right by 2 stops (overexpose a bit) Is this in affect reducing ISO. And at ISO 1600 I want to expose to the left by 2 stops (underexpose a bit)

John Sheehy wrote in post #9900096
If you point the camera at a scene, in manual mode, and you vary the ISO so that the meter says +2 at ISO 1600, +1 at ISO 800, 0 at ISO 400, -1 at ISO 200, and -2 at ISO 100, then the ISO 1600 will give the least noise, and ISO 100 the most. If the scene has a majority of area as the brightest tones, the 1600 at +2 might be usable.

The point is that increased absolute exposure is what makes ISO 100 images cleanest, not the ISO 100 setting itself, which is the noisiest in manual mode (IOW, with a fixed, given manual exposure). The tradeoff is noise vs highlight headroom.

Methodical wrote in post #9900350
I will have to give that a try. So at ISO 100 I want to expose to the right by 2 stops (overexpose a bit) Is this in affect reducing ISO. And at ISO 1600 I want to expose to the left by 2 stops (underexpose a bit)

I think you miss the point still. You're talking about 8 stops difference in absolute exposure there, impossible with a fixed absolute manual exposure (Av & Tv).

This is what I am saying: If you have set the Av and Tv values manually, and are now choosing the ISO, there is no repercussion from going higher, other than the possibility of clipping, if you go too high. The highest ISO which doesn't cause unwanted clipping will give the least noise (but you may only notice this in shadows). The normal rules about high ISO contribution to noise reverse when you have fixed Av and Tv values, in Canon DSLRs, many Nikons, and a few Pentaxes (but not at all with other Nikons, Pentaxes, and with P&S cameras or Olympuses- they have the same noise or very close, regardless of ISO setting).

Was reading something else and came across this.

Just did a quick test to check your assertions. RAW, ACR all default settings, all noise and sharpening to 0.

First series - Av f8, metering to brightest area gave 1/250 @ ISO 100, snapped a series from ISO 100-1600, severe clipping from 32 but that is because on the 550D, the shutter cannot go any faster then 4000.

2nd series - Av f8, metering to midtones gave 1/160 @ ISO 100, snapped just ISO 100 and 1600.

Import though ACR in PS5.

All pics at normal zoom look good.
Bright tone metered pics - zoomed to 400%, substantial noise at 1600 decreasing steadily to none visible at 100.

Midtone metered pics - none visible at 100, evident at 1600 but no color color noise compared to bright tone metered pics.

Either I am doing this wrong or there is something else going on ?

John Sheehy wrote in post #9900490
I think you miss the point still. You're talking about 8 stops difference in absolute exposure there, impossible with a fixed absolute manual exposure (Av & Tv).

This is what I am saying: If you have set the Av and Tv values manually, and are now choosing the ISO, there is no repercussion from going higher, other than the possibility of clipping, if you go too high. The highest ISO which doesn't cause unwanted clipping will give the least noise (but you may only notice this in shadows). The normal rules about high ISO contribution to noise reverse when you have fixed Av and Tv values, in Canon DSLRs, many Nikons, and a few Pentaxes (but not at all with other Nikons, Pentaxes, and with P&S cameras or Olympuses- they have the same noise or very close, regardless of ISO setting).

bobobird wrote in post #12212064
Either I am doing this wrong or there is something else going on ?

You tested something completely different than what I was talking about. You varied the actual sensor exposure, along with the ISO, so that the lower ISOs had more actual sensor exposure. What I am talking about would be tested as follows:

Set the camera on a tripod, and pick a subject. Set the camera to the highest ISO and choose a manual exposure that centers the meter (or even a little bit to the right), then, change only the ISO, bringing it down to ISO 100 eventually.

The tradeoff will be lower noise for the highest ISOs, but less highlight headroom.

Thanks for the quick response.

Will do as you suggested.

Forgive another newbie question but how could this be used in a practical situation ?

bobobird wrote in post #12212627
Thanks for the quick response.

Will do as you suggested.

Forgive another newbie question but how could this be used in a practical situation ?

It lets you know what you're trading off, when choosing an ISO in "M" mode. At the lower end of the ISO scale, leaving a lot of headroom comes at a big cost in extra shadow noise. At the higher end of the range, using a lower ISO has little or no cost in extra noise, while giving much more headroom.

Thanks John, tried it out this morning.

Bright day. Camera on tripod pointing out the window. In-camera noise reduction turned off. Spot metering to a midtone.

At ISO 6400, needle would only center at F20 for shutter of 1/3200. Took a sequence of shots for each ISO to 100.

The historgram did not hit either the left or right sides except at 100.

All the pics turned out very noisy perhaps a little less at 6400 but very soft. No noise improvements were seen in any of the images down to 100.

Doing it wrong again ?

sticking my nose in bobobird ...
Did you keep the shutter speed and aperture setting the same for all your pictures?

I think the idea is to vary ONLY the ISO. This way, the total amount of light is the same - only the camera amplification (ISO) is changing.

You should end up with a set of photos - ranging from very very bright .. bright ... normal ... dark ..etc.

Then in post, you adjust the photos to the same apparent brightness.

The dark shots at ISO 100 (which are brightened) should show more shadow noise than the higher ISO shots (initially too bright, then dimmed in post).

(Take what I say with a grain of salt ... i am often wrong)

Yes only changed ISO, bright to dark as you said.

Only looked at the noise levels in post but did not try to correct assuming I was doing it wrong and had to redo.

I will pp the ISO 200 as it appears a bit better and post back.

Thanks.

bobobird wrote in post #12217465
Yes only changed ISO, bright to dark as you said.

Only looked at the noise levels in post but did not try to correct assuming I was doing it wrong and had to redo.

I will pp the ISO 200 as it appears a bit better and post back.

Thanks.

Heh! You got the right idea, although you went farther than I typically advise people to go!

My test is: take a scene, not particularly bright, that you can get a "good" exposure at ISO 1600, in Manual, shooting Raw and with any in-camera settings for noise reduction off including HTP if you ever use it.

Take the shot, and you want few if any highlights clipped, few if any shadows clipped -- those things can skew your results. Don't worry about "exposing to the right" so much as to get a well-balanced scene with enough lighter areas as well as shadow/darker areas that can help you really evaluate things.

Then, all you really need to do is lower the ISO by four stops to 100. You are in Manual, so you will be taking in the same amount of light, the sensor will collect the same amount of light, the only difference is that the camera will then pass the signal off to the respective ISO electronic amplifier, it will be amplified and then the resulting Raw data will be saved to the Raw file (again, no in-camera "fiddling").

Now the acid test: we all learned that "A high ISO produces noise so it is always better to use a low ISO", right?

If that were true, then we can test that here. It's easy: take the ISO 100 image into your Raw processor of choice, and increase the overall exposure by four stops so that it is the equivalent of ISO 1600 -- it sounds scary, but this is exactly what we are talking about here -- the ISO 100 signal "should" according to the "urban photographer legend" be cleaner than the ISO 1600 shot. Well, you be the judge. It has actually been found that in at least most Canon models, presumably because of all the work that has been put into high ISO performance, the ISO amplifier produces a bit less noise than the lower ISOs.

Believe it, or not!

I am too impatient for Manual for most cases... I find Av mode every bit as useful, and usually do little adjustment if at all. I can look at a scene and expected what EC I would need, ISO, etc... Manual would be much slower, and for much of the candid stuff I do, I don't have time!

Manual I use for special occasions (and when using flash), but not often!

Yea I started to use manual two months ago and haven't looked back

Manual sure is a blast to use. I used it back in December when I was in Florida taking night shots, especially of fireworks, and manual really helped me achieve the results I wanted! This post makes it sound like I never use Manual, but since I use a prime, Im 99% in Av mode

Tony_Stark wrote in post #12218549
Manual sure is a blast to use. I used it back in December when I was in Florida taking night shots, especially of fireworks, and manual really helped me achieve the results I wanted! This post makes it sound like I never use Manual, but since I use a prime, Im 99% in Av mode

I use prime on manual. Are you saying it's better to be in av? I ask cause I'm new to manual.