going Manual!!!!

Thanks Tim.

There is a lot of information around the forum regarding 'Expose to the right', 'ETTR', or 'HAMSTR'.

Not sure if anyone posted yet .. lot of good info

https://photography-on-the.net/forum/showthre​ad.php?t=744235

Wilt wrote in post #12228046
I am, nevertheless, failing to grasp the logic which he presents, that it is better to shoot at higher ISO than at lower ISO. "The point is not that 1600 is a better ISO to shoot at; the point is that if your chosen manual exposure doesn't blow highlights at 1600, you will be better off set to 1600 than something lower."

Perhaps that is because your noise model is fabricated, and has nothing to do with Canon reality.

The electronic noise in the camera does not all originate before or during amplification, as your model suggests. There is a lot of noise incurred in the system after amplification. At ISO 100, it is most of the noise (other than photon shot noise, which is only related to the actual absolute exposure on the sensor). At ISO 3200, it is only a small part of the added noise. So, depending on the model, there is 5x to 10x as much added absolute noise at ISO 100 as at the highest ISOs. The steepest part of this curve occurs right at the bottom; ISO 100 adds almost twice as much absolute noise as ISO 200.

John Sheehy wrote in post #12233593
Perhaps that is because your noise model is fabricated, and has nothing to do with Canon reality.

The electronic noise in the camera does not all originate before or during amplification, as your model suggests. There is a lot of noise incurred in the system after amplification. At ISO 100, it is most of the noise (other than photon shot noise, which is only related to the actual absolute exposure on the sensor). At ISO 3200, it is only a small part of the added noise. So, depending on the model, there is 5x to 10x as much added absolute noise at ISO 100 as at the highest ISOs. The steepest part of this curve occurs right at the bottom; ISO 100 adds almost twice as much absolute noise as ISO 200.

Since you seem to have insight into this, could you please explain for us the post-capture noise introduction as embodied within the Canon implementation? I know that I have a practice of using ISO 100, 400, and 1600, but it would seem that ISO 200 would be a better 'floor', and I want to understand why that might be true.

Came across this in my travels. May be of interest to the discussion ?

http://www.northlight-images.co.uk …ras/canon_1ds3_​noise.html

Wilt wrote in post #12233616
Since you seem to have insight into this, could you please explain for us the post-capture noise introduction as embodied within the Canon implementation? I know that I have a practice of using ISO 100, 400, and 1600, but it would seem that ISO 200 would be a better 'floor', and I want to understand why that might be true.

It's just the nature of CMOS sensors with analog signals, most likely.

The only sense in which 200 is unique is that 200 vs 100, (IOW, 200 with ETTR) gives the greatest savings in the added noise. 400 w/ETTR vs 200 is slightly less of a savings, but 6400 w/ETTR vs 3200 does nothing but lose highlight headroom.

I have the 20D values memorized, so I'll use them; measured in 12-bit RAW standard deviations of black frames, the noise is as follows:

ISO Read noise
100 2.07
200 2.2
400 2.45
800 3.2
1600 4.7
3200 9.4

As you can see, the read noise floor is not proportional to ISO, except at the upper end of the range.

This pattern is typical of all Canon DSLRs except the original 1D, which uses a Panasonic CCD sensor. It is also the way some Nikon-designed (non-Sony) DSLR sensors like the one in the D3 and the one in the D3s.

Almost all other cameras out there work pretty much in the simple linear way that you assumed, except that some cameras add a tad more absolute noise at base ISO as double the base, with no change above that, and some compact cameras have slightly more absolute read noise at higher ISOs, because of very poor amplifiers. For compacts with RAW, in general, it is always beneficial to do all your "high ISO" work from base ISO, with so-called "under-exposure", because even if it does not improve the noise, you get smaller RAW files and/or more highlight headroom.

Does anyone have any side by side comparison shots under identical lighting. If not I'll do it later with my 50D to see the differences in noise at various ISO from low to high.

In the mean time, what it boils down to in the end is that you want to use the ISO that will be more of an advantage and give you the best IQ for the situation. In a low light situation which is better? A lower ISO and longer shutter speed, or a higher ISO and shorter shutter speed. Both ways are going to produce noise in one fashion or another. Personally I would rather deal with a little ISO noise that I can easily fix if necessary.

When I use my camera for astrophotography I have found the optimal ISO shooting dark skies is ISO 800. Above that is too noisy and would take too many light frame shots stacked to improve the SNR to an acceptable level.
The background colors as well as highlights and shadows also make a considerable difference in how visible the ISO noise will be in each picture.


in case anyone is interested here is some information on high iso and high iso noise reduction from Canon.
http://www.learn.usa.c​anon.com …e_reduction_art​icle.shtml

http://www.learn.usa.c​anon.com …eorge_lepp3_art​icle.shtml

And a couple more
http://www.digital-photography-school.com/iso-settings

http://www.digital-slr-guide.com/iso-and-image-noise.html

bobobird wrote in post #12233815
Came across this in my travels. May be of interest to the discussion ?

http://www.northlight-images.co.uk …ras/canon_1ds3_​noise.html

This would have been a little nicer if the histograms had been zoomed into, or gave some noise numbers, but you can still see the overall effect. The 3200 is clearly double the noise of 1600, but certainly not 32x the noise of 100.

You also see a secondary effect here, with the 1/3-stop ISOs. The way the original 5D and the 1D-series cameras do these ISOs is to use 1.25x and 1.6x analog gain just before or in the ADC (too late a stage to be of any real value), multiplying the read noise by these same values. At the lower ISOs, this means more read noise at the 1.25x and 1.6x gains than at the next higher ISO (more read noise at ISO 160 than at 200). At higher ISOs, where read noise is more proportional to ISO, these 1.25x and 1.6x gains do not cause extra noise (but waste RAW headroom).

dawnwithacamera wrote in post #12236134
One thing that's really important to consider, is how to meter. Once you get the concept of how the buttons and dials interact that should be the next thing you concentrate IMO.

That's the one thing I am not quite getting the gist of. For example, I have a shot set, but my EV meter is at -2, what do I need to change to get it back to 0? Is the the shutter speed, the f/stop or the ISO?? Or is that dependent on what I am trying to achieve?

Just thought of the way I really wanted to ask my question--if my meter says -2, what way to I need to go with my f/stop? Smaller or larger, or my shutter, faster or slower? That is the part that I am confused on!!!

The thing to bear in mind is that the "cure" for noise at any ISO is light. Underexposing at any ISO is, from my experience, inviting visible noise in your images.

But, of course, using a high ISO implies that your image would be underexposed at a lower ISO due to low light. What I've seen is that the "native" ISOs (which are electronically amplified) are of a better handling than trying to boost amplification in post.

When setting up a shot, my rule of thumb is to get the best exposure possible using the aperture and shutter speed and the parameters of your scene. If at ISO 100 you can get a great exposure, then that will be the best ISO. But if not, and you can't slow the shutter speed or open the aperture more without messing up your shot, then it is better to go a stop up in native ISO as long as you don't blow needed highlights.

Dawn, so you know, if you are using Spot metering and Manual Exposure, you can use a subject like that, or other parts of the scene, to set your exposure. It's more tricky in the "program" modes (Av, Tv and P) but people have learned to apply the same principles to get the right approach.

Basically, if you meter a subject and it shows a meter reading, it is up to you to determine if that is the "correct" reading for that subject in the scene/lighting you are dealing with. A very dark subject could in fact have an "accurate" reading at -2 EV and you could leave that stand. Or, if that subject has important detail you want to capture, you could nudge your aperture or shutter speed to get things "To The Right" as long as you keep an eye on highlights that you don't want to lose. Or, if your overall scene is "too dark" but you can't slow the shutter speed down without getting motion blur, or you can't open the aperture any more, well then it's time to raise the ISO -- if you can go to the next stop up of native ISO and not blow important highlights, assuming you are shooting Raw, then that will get you the best "starting point" -- in your Raw processor you can then get what could be a bit of a bright image toned down while keeping the shadows from being dragged too low.

This whole thread and links has validated my choices while using my new-to-me 50D for the first time. I was competing in a shotgun only practical shooting match. Most of the stages were "jungle runs" which are basically set up along trails through the woods with the targets placed out in the woods. Starting out I was getting frustrated because in Av the shutter speed kept bouncing all over as the shooter moved through the different sunlight and shadowed areas. After two shooters' runs and basically 0 keepers I decided to change to manual and set the EV on the competitor in the sun and keep it there. It worked out great and I learned more during that 4 hours of shooting than I thought possible. Now I need to start learning how to work with RAW files and get a bigger CF card.

Curious. How fast were the targets going from sunlight to shade? What method did you employ to adjust for such changes? Was this some kind of contest or training or something?

Al

ALShooter wrote in post #13807403
...Starting out I was getting frustrated because in Av the shutter speed kept bouncing all over as the shooter moved through the different sunlight and shadowed areas...

Methodical wrote in post #13807722
Curious. How fast were the targets going from sunlight to shade? What method did you employ to adjust for such changes? Was this some kind of contest or training or something?

Al

The targets are static with the exception of a few flying clays. The change from sunlight to shade pretty much happened anytime the competitor (my subject) moved. They are only shooting for 15 seconds if they are real fast to a minute or so if they have issues reloading or have malfunctions. To adjust for major changes I was cranking the wheel to adjust shutter speed and keep the EV somewhat centered. I did mistakenly shoot most of the match at too high an ISO. The main thing I learned is that I have a LOT to learn. I'll post a picture once I get home to show the environment.

It is a shooting match or competition. The goal is to hit the targets faster than others do so it involves running with guns. To start a stage a competitor "makes ready" by loading the gun and getting into the defined start position. When the buzzer sounds the timer starts and you can start shooting. The longest stage included about 40 yards of movement through the woods. This match was put on primarily for 3-gun competitors to get some shotgun only practice in.

This was one of the stages. You can see a target to the right.

HOSTED PHOTO
please log in to view hosted photos in full size.

>>The use of the reflected light meter in the camera makes the scene seem to be forever changing, when the source of light is constant. Exposure should NOT be keyed to the brightness of the subjects, it should be keyed to the brightness of the light falling onto the scene.>>

Thanks for this info. I have a question though. If the reflected light is changing, why do the settings not have to change? Spot metering (or partial) may be more effective at a very dark subject on a very sunny day and spot metering allows for this, correct?

Thanks,

Tim