A Question about Camera Noise

[Dave Leroy]

After reading a recent post here, " Canon 7D and 1D Mark IV Metering posted 09-15-2010" I began to wonder about the apparent advantage some cameras have over others in regard to use of high ISO's. Artie clarifies in this post that the 7D and 1DMkIV need considerably more light than a 40D or 50D in a low light condition.

If a 7D or a 1DMkIV need more light than a 40D or 50D in low light conditions then if all other things are the same, isn't some of the high ISO advantage of the 7D and 1DMkIV lost? Perhaps most of the advantage.

Dave

[Mike Milicia]

I take Artie's comments to mean that when using an autoexposure mode (e.g. Av) and Evaluative metering in low light, the 7D and Mk IV require more positive exposure compensation than the 40D/50D in order to arrive at the same correct exposure. This would not lessen the high ISO advantages in any way.

Hopefully Artie will chime in and confirm.

[Roger Clark]

Dave,

I think Artie was referring to metering.

But regarding camera noise, the noise we see in digital camera images is almost all due to what is called photon noise, the noise due to counting for a brief period (the exposure time) the random arrival of photons to the focal plane. These physical properties set basic performance of any sensor. For example, for a given sensor size (e.g. APS-C), as one increases the pixel count, each pixel must be smaller so it collects less light. The analogy is collecting rain drops in a rain storm: a smaller bucket will collect less water. As one boosts camera ISO, the sensor doesn't actually change sensitivity, the camera changes post sensor signal gain, and since we shorten the exposure time, the sensor collects less light. In reality, noise also drops, but signal drops more (noise drops by the square root of the signal, and signal drops linearly with exposure reduction). What we perceive as noise is signal-to-noise ratio.

So cameras like the 7D with its small pixels collect less light per pixel, but those smaller pixels can resolve more detail. If one collects enough light, the signal-to-noise ratio is high and the image looks good. People seem to like/dislike different levels of apparent noise, so some will like some high ISO images from a camera, while others will not.

More on sensor noise (probably more than most want to know):
http://www.clarkvision.com/articles/...mance.summary/

Roger

[Dave Leroy]

What i am wondering about is how camera's make use of available light.

In a low light situation, Av mode, and shooting wide open by adjusting exposure compensation (changing shutter speed) it is possible to get a "good" exposure. A greater + EC adjustment lowers the shutter speed and allows more light to reach the sensor.

To obtain a higher a shutter speed ,the only option left in such a situation is to change ISO settings which may leave more "noisy" or "grainy" looking photos.

It is commonly said that a 7D and a 1DMK IV produce less "grainy or noisy" looking photos at high ISO's when compared to a 50D.

But if shooting in the same light, same lense wide open, Av mode, it is necessary to use a higher EC with a 1DMK IV than a 50D, the shutter speed with the 1DMKIV will be less. To get the 1DMK I back to a higher ss, the ISO must be bumped, but not with the 50D.

Therefore when comparing camera "noise", say a 50D vs a 1DMKIV it does not seem right to compare them at the same ISO values unless the ss and aperture are the same also.

And thanks Roger. Although I don't clearly understand the sources of noise, I have read with interest your past comments about noise. I do understand that pixel size does play as role as well as characteristics of cameras.

I am just really wondering if the noise advantage of these more expensive cameras is all it is supposed to be. I also understand there are other advantages to more expensive cameras. :confused:

Dave

[Mike Milicia]

But if shooting in the same light, same lense wide open, Av mode, it is necessary to use a higher EC with a 1DMK IV than a 50D, the shutter speed with the 1DMKIV will be less. To get the 1DMK I back to a higher ss, the ISO must be bumped, but not with the 50D.
Dave

This where I think your logic is going awry. Yes, with the Mk IV, you would need to use a higher EC, but only so that you will end up at the same shutter speed as the 50D with less EC.
Using your scenario, let's say you need +1/3 compensation to get the correct exposure with the 50D which puts the shutter speed at 1/125.
With the Mark IV, you may need to use +1 compensation to get to that same correct exposure, i.e. 1/125. There is no need to bump the ISO. The difference in the amount of compensations you need to apply is due to differences in the Evaluative metering algorithms between the cameras, not because one sensor reacts to light far differently from the other.
The correct absolute exposure settings are based solely on the amount of light hitting your subject and will be the same regardless of what camera you are using (otherwise incident meters would require you to dial in the camera model!)
What may differ between cameras is the amount of EC you need to apply to achieve that same correct exposure when using Evaluative metering and an autoexposure mode.

[Dave Leroy]

OK. Thanks Mike. Now I understand, and that is what you were both telling me.

It is the difference in metering of the cameras.

Much appreciated.

Dave

[Roger Clark]

Dave,

If you pick up any old photography book (film days, a decade or more old), one of the first things most books say is go out an calibrate your light meter. That is still true today.

Regarding light, if you use the same lens, same f/stop, and same exposure on a 7D, 50D and 1D Mark IV, the 50D will collect the least amount of light per square mm, but only by a bit. This is because of the improved collection efficiency of the newer sensors in the 7D and 1DIV. What the meters read is an entirely different story. The variation between cameras of the same model is likely variable too, as well as between models.

Roger

[Dave Leroy]

Thanks Roger.

What a wealth of information on your website! It covers the range from good advice to very technical.

Dave

[Emil Martinec]

What i am wondering about is how camera's make use of available light.

In a low light situation, Av mode, and shooting wide open by adjusting exposure compensation (changing shutter speed) it is possible to get a "good" exposure. A greater + EC adjustment lowers the shutter speed and allows more light to reach the sensor.

To obtain a higher a shutter speed ,the only option left in such a situation is to change ISO settings which may leave more "noisy" or "grainy" looking photos.

It sounds like you are giving too much primacy to what the metering is saying and doing. For instance, ISO is not part of exposure. Exposure is governed by shutter speed and aperture, which dictate the light per unit area reaching the sensor during the capture. ISO is simply a way of putting that exposure in a particular range of raw values in the raw file (especially at high ISO, where electronic read noise tends to be independent of ISO relative to exposure).

The reason the shutter speed changes as the ISO changes in Av mode is due to metering, which is always trying to keep the center of the image histogram in a particular range of raw values; so change the ISO and the same exposure is pushed or pulled to a different raw value, so the exposure is changed to re-center the histogram.

A lower exposure leads to lower S/N ratio, which visually appears as more "grainy" in the image.

It is commonly said that a 7D and a 1DMK IV produce less "grainy or noisy" looking photos at high ISO's when compared to a 50D.

In part that is because the 7D and 1D4 have less read noise than the 50D. Also the 50D has a good deal of pattern noise which corrupts raw conversion. Note also that the 1D4 has more sensor area, so collects more light for the same exposure (another way to look at it is that the smaller crop factor trades shallower DoF for more light gathered and hence better SNR)

But if shooting in the same light, same lense wide open, Av mode, it is necessary to use a higher EC with a 1DMK IV than a 50D, the shutter speed with the 1DMKIV will be less. To get the 1DMK I back to a higher ss, the ISO must be bumped, but not with the 50D.

This is why I thought you might be giving too much credence to what the camera metering suggests is the "correct" exposure. Raw converters have and exposure slider, precisely because different cameras can have different suggested exposure according to their metering. All this means is that they like to center the histogram in a different portion of the raw range, which then leads to a different suggested exposure for the same ISO setting. You are always free (within limits) to override that suggestion, not by changing the ISO but by using the EC dial to change the exposure at the same ISO.

Therefore when comparing camera "noise", say a 50D vs a 1DMKIV it does not seem right to compare them at the same ISO values unless the ss and aperture are the same also.

Exactly so.

I am just really wondering if the noise advantage of these more expensive cameras is all it is supposed to be. I also understand there are other advantages to more expensive cameras. :confused:

Dave

Within a given generation of Canons, there is very little difference in high ISO noise, especially the 1.6 crop cameras tend to use the same sensor. The big difference comes with crop factor, since as I mentioned above one trades a shallower DoF for greater light gathering capacity when the exposure is kept fixed. If on the other hand the DoF is kept fixed (by stopping down in inverse proportion to the crop factor), then the light gathered during the capture is independent of the crop factor and the images will be more or less the same in terms of SNR ("noise" in colloquial photography terminology) irrespective of sensor size.

[Dave Leroy]

Thank you very much Emil for the detailed response.

Mike, Roger and yourself have clarified a lot for me and have been very helpful.

The reason I had posted my question was here near Vancouver BC we are entering a period with a lot of low light conditions. I had been considering different camera options, and after misreading Artie's comments in the other thread I was confused.

Thanks Again,

Dave

[Geraldo Hofmann]

After reading a recent post here, " Canon 7D and 1D Mark IV Metering posted 09-15-2010" I began to wonder about the apparent advantage some cameras have over others in regard to use of high ISO's. Artie clarifies in this post that the 7D and 1DMkIV need considerably more light than a 40D or 50D in a low light condition.

If a 7D or a 1DMkIV need more light than a 40D or 50D in low light conditions then if all other things are the same, isn't some of the high ISO advantage of the 7D and 1DMkIV lost? Perhaps most of the advantage.

Dave

Well for me the advantage is at least partly gone if i need + 0.7 with my 7D and 0 with the 40D I loose nearly one stop in terms of shutter-speed. I have no experience with the MKiV but the 7D really needs + compensations in situations my MkII or 40D survives without in particular if I want some details in dark shadows.

But the other benefits are more important for me so I live with that drawback. No free lunch I guess :-)

[Roger Clark]

The reason I had posted my question was here near Vancouver BC we are entering a period with a lot of low light conditions. I had been considering different camera options...

Dave,

It comes down to basic physics. If light drops, you need a larger diameter lens focusing on a larger pixel. So 1) buy the camera with the largest pixels, 2) choose the focal length you want for the detail you desire, 3) choose the biggest lens in that focal length that is available, 4) save the money to buy the camera and lens, and 4) lift weights while saving the money to gain the strength needed to carry the bigger system ;).

Seriously, the larger the diameter of the lens, the more light it collects. If a larger lens is available, it will give you more advantage than a new camera. For example, if using a 300 f/4 with 1.4x TC (420 mm f/5.6), switching to a 400 f/2.8 gives a 2-stop advantage with your same camera. The difference between camera models is smaller.
(Comment about saving and weight lifting still stands.;););))

Roger

[Mike Milicia]

Well for me the advantage is at least partly gone if i need + 0.7 with my 7D and 0 with the 40D I loose nearly one stop in terms of shutter-speed.

I'm afraid you are making the same mistake that Dave did in his original post. For a given amount of light, you do NOT lose ANY shutter speed going with the 7D vs. the 40D.
A detailed explanation can be found in my previous posts in this thread.

[Dave Leroy]

A very good point Roger about getting a larger diameter lens, and perhaps at times overlooked.

The weight issue I do understand. A camera, Canon 500mm f/4, Gitzo tripod and Wimberly II sit quite snugly on my shoulder.

I suspect I would be far better off spending time on Lighting, Composition, and Subject, as mentioned on your website, as well as PhotoShop then wondering about another camera. :)

Dave

[Geraldo Hofmann]

I'm afraid you are making the same mistake that Dave did in his original post. For a given amount of light, you do NOT lose ANY shutter speed going with the 7D vs. the 40D.
A detailed explanation can be found in my previous posts in this thread.

Mike I would be happy this would be the case I still have have the 40D, 7D, 1DsMkII and to get the same level of details in the dark shadow I need +0.7 with all three 7Ds and I am the limit to blow the whites. This is not the case with the other two camera bodies mentioned. And I definitely need a longer exposure time with the 7Ds to get clean details in the shadow. At least this is was the results of my tests with high ISO with low light and dark shadows. In bright light and without dark shadows it doesn't matter. It is not a matter of a different metering system. But maybe others have different experience when they compare their cameras.

[Roger Clark]

Geraldo,

What raw converter were/are you using with the 7D raw files? The 7D is a rather unique camera, having two green filters. Some raw converters did (maybe do) not treat them correctly. As pointed out by others on BPN, converting with the Canon raw converter produced images with less apparent noise.

On an equal area basis, the 7D collects more light than any other Canon DSLR given the same lens and exposure time, and about 50% more light than a 40D. This is shown in Figure 10 at: http://www.clarkvision.com/articles/...mance.summary/

So, on an equal (image) resolution basis, the 7D should have plenty of detail in the shadows. It also has lower read noise, helping to give more detail in the shadows.

Roger

[Flavio Rose]

Dxomark claim that ISO's match well on the 40D and 7D.

http://www.dxomark.com/index.php/en/...rand2%29/Canon

To me this means that if you set the same exposure manually on the 40D and 7D and photograph the same scene, you should get a similarly bright image. Now, as Mike Milicia says, that does not mean that if you are using Av with evaluative metering and the same f/ number and ISO on both cameras and photograph the same scene, you will get a similarly bright picture (or similar shutter speeds for that matter). The evaluative metering may evaluate the scenes quite differently and give you different shutter speeds (or so I believe comparing the XTi with the 7D).

[Geraldo Hofmann]

Geraldo,

What raw converter were/are you using with the 7D raw files? The 7D is a rather unique camera, having two green filters. Some raw converters did (maybe do) not treat them correctly. As pointed out by others on BPN, converting with the Canon raw converter produced images with less apparent noise.

On an equal area basis, the 7D collects more light than any other Canon DSLR given the same lens and exposure time, and about 50% more light than a 40D. This is shown in Figure 10 at: http://www.clarkvision.com/articles/...mance.summary/

So, on an equal (image) resolution basis, the 7D should have plenty of detail in the shadows. It also has lower read noise, helping to give more detail in the shadows.

Roger

I am using LR3.x for converting the raw files... yes with equal image size the 40D and 7D are pretty close but if I want to exploit the plus of resolution f.e. crop the 7D image to 10MP... the the problems show up... but it is no big problem for me I just wished the 7D would be a little more forgiving in terms of exposure ... would be more tolerant in terms of pushing the shadows

I hope I could explain it so that it is understandable ... I will try to upload two images when I am back from the rain forest :-)

[Geraldo Hofmann]

Dxomark claim that ISO's match well on the 40D and 7D.

http://www.dxomark.com/index.php/en/...rand2%29/Canon

To me this means that if you set the same exposure manually on the 40D and 7D and photograph the same scene, you should get a similarly bright image. Now, as Mike Milicia says, that does not mean that if you are using Av with evaluative metering and the same f/ number and ISO on both cameras and photograph the same scene, you will get a similarly bright picture (or similar shutter speeds for that matter). The evaluative metering may evaluate the scenes quite differently and give you different shutter speeds (or so I believe comparing the XTi with the 7D).

This is exactly what I did but see mey response above...