Jump to content
lct

Clean CL at 3200 iso?

Recommended Posts

I’ll see if I can dig up a good, general reference that isn’t too technical. In the meanwhile here is a brief description.

 

There are several sources of noise in any image. The first is read noise—the ability of the camera to accurately read out the contents of the CMOS chip at each pixel. It is typically fixed at somewhere between 2 and 20 electrons depending on ISO. If you are trying to read out 20,000 electrons of signal (plenty of light) even 20 electrons is immaterial. As the light gets dimmer, though, 20 electrons are more relevant. If your signal was 100 electrons, that read noise matters—a lot. This was the situation we were in, say, ten years ago. Since then, read noise levels have been dropping. Now it’s not at all unusual to see less than 2 electrons of read noise. Even with very low light, 2 electrons is usually pretty minor.

 

The next source of noise is thermal noise. This really only affects long exposures. I’ll ignore this for now.

 

The interesting one for this discussion is shot noise. Light itself is quantized, and the flow of photons from your subject into the camera is not perfectly even. This is shot noise and is inherent in the light. Google Poisson distribution if you want to read about this. Basically, it’s the random distribution of counts that are bounded at zero. Like a bell curve for measuring light. Even a theoretically perfect detector can’t do anything about this. There just isn’t enough light at each pixel. It’s an upper wall for all detectors.

 

So why did I mention “pixel pitch” and “sensor size?” If you have a larger sensor or larger pixels (or both), you get more light hitting each pixel so the “wall” moves up. This is why the Sony A7s does so well, for example, compared to other full frame chips. Only 12 megapixels, so many more photons per pixel vs a higher resolution camera. Better signal-to-noise in the light itself.

 

We are literally getting to the point where sensors won’t be able to improve any more. Quantum efficiency on many back side illuminated sensors is approaching 80%, so there is no more light to grab. Read noise is already down around 2 electrons. Thermal noise is all but irrelevant for exposures up to several seconds. What’s left is the physics of light itself

 

I disagree with the "getting to the point where sensors won't be able to improve anymore" assertion.

There are many sensor improvements I'm waiting for to become commercially available:

 

curved sensors (already patented by Sony, Microsoft, Canon, etc)

native ISO <100

16-bit ADCs

non-Bayer filter image sensors (Foveon, etc)

 

There's probably more if I spent more time thinking about it.

 

Toshi

Share this post


Link to post
Share on other sites

Sensor improvements I'd like to see in APS-C and FF are:

 

Deeper wells to hold more electrons within the same surface area -- this would make true 14 bit 4 micron pixels happen, but for Leica's non-telecentric lenses would probably only work with microlenses on the cover layer.

 

16-bit ADC (already available on big MF chips like the 100 MPX Sony's that Phase One and Hasselblad use).

 

Global electronic shutter so that there is no "jello" effect from a 1/20 sec scan of an image taken locally at 1/250 sec.  This may require "back side" technology to separate the extracting circuits from the picture-taking circuits.

 

I know enough of the hassles that Foveon has to overcome to get any decent color out that i'm not holding my breath on stacked 3-color technology, nor am i convinced that we need it.  It is a cool idea, though, and some very smart people worked on its hardware and software (Carver Meade, Dick Lyons, & ...).

 

scott

Share this post


Link to post
Share on other sites

Just shot a couple hundreds indoor pics last night mostly at 3200 isos. I did not take my usual low light gear (Sony a7s mod, Lux 35, 50 & 75) to avoid the bulk of the 75 and to test the CL at the same time so i had the CL with a Lux 35/1.4 v2, a ZM 50/1.5 and a Cron 28/2. I would have preferred a small Lux or Cron 18 or 21 instead of the 28 but the CL worked like a charm i must say. EVF not always responsive in dim light but i did not miss more than a couple of shots because of that. As far as noise is concerned, i will have to use Define or same for the first time since i don't remember when. Magic C1 where are you?

 I will bring my usual gear next time with a 75/2 instead of the 75/1.4 but the little CL is very good otherwise. And progress won't stop suddenly in 2017 hopefully.

Edited by lct

Share this post


Link to post
Share on other sites

ISO 3200

 

aldi.jpg

 

 

I feel rather spoilt by the CL too...

Noise reduction zeroed in ACR.

I'm starting to do some comparisons CL + 35 Summicron-TL vs. α7R III + 50 Apo Summicron-M and 50/1.4 Zeiss Planar FE. What is ISO 2500 for the α7R III is ISO 1600 for the CL. I didn't go beyond 1600 on the CL today and they are really good. But first at ISO 800 and ISO 1250, respectively. 

 

Uncompressed JPEGs here: https://www.smugmug.com/gallery/n-qDx8fd/

 

α7R III + 50 Apo Summicron-M, opened in LR and only WB adjusted a bit.

ISO 1250 f/2.0 @1/160 sec.

 

CL + 35 Summicron-TL, opened in LR and WB adjusted to try to match the Sony

ISO 800 f/1.4 @1/160 sec.

Edited by Chaemono

Share this post


Link to post
Share on other sites

And now with the Shadows lifted by +100

 

Uncompressed JPEGs here: https://www.smugmug.com/gallery/n-qDx8fd/

 

α7R III, Highlights -11, Shadows +100, NR +30, LR default sharpening

ISO 1250 f/2.0 @1/160 sec.

 

CL, Exposure -0.20, Highlights -13, Shadows +100, NR +30, LR default sharpening 

ISO 800 f/1.4 @1/160 sec.

Edited by Chaemono

Share this post


Link to post
Share on other sites

And now the crops

 

Uncompressed JPEGs here: https://www.smugmug.com/gallery/n-qDx8fd/

 

α7R III

ISO 1250 f/2.0 @1/160 sec.

 

CL

ISO 800 f/1.4 @1/160 sec.

Edited by Chaemono

Share this post


Link to post
Share on other sites

And now CL and α7R III at ISO 1600 and ISO 2500, respectively. ISO 2500 on the Sony is really ISO 1600 on the CL. CL with the wonderful 35 Summicron-TL and α7R III with its native 50/1.4 Zeiss Planar FE.

 

Uncompressed JPEGs here: https://www.smugmug.com/gallery/n-qDx8fd/

 

 

α7R III, from ARW opened in LR and WB adjusted a bit to try to match CL

ISO 2500 f/3.2 @1/320 sec.

 

 

CL, from DNG opened in LR and WB adjusted a bit to try to match Sony

ISO 1600 f/3.2 @1/320 sec.

Share this post


Link to post
Share on other sites

And now with the Shadows lifted by +100

 

Uncompressed JPEGs here: https://www.smugmug.com/gallery/n-qDx8fd/

 

 

α7R III, Shadows +100, NR +25, LR default sharpening

ISO 2500 f/3.2 @1/320 sec.

 

 

CL, Exposure +0.40, Shadows +100, NR +25, LR default sharpening

ISO 1600 f/3.2 @1/320 sec.

Share this post


Link to post
Share on other sites

And now the crops of Baby Jesus. CL with a tad more noise.

 

Uncompressed JPEGs here: https://www.smugmug.com/gallery/n-qDx8fd/

 

 

α7R III

ISO 2500 f/3.2 @1/320 sec.

 

 

CL

ISO 1600 f/3.2 @1/320 sec.

Share this post


Link to post
Share on other sites

I must confess that, as I have no intention of moving to Sony -horrid UI- these “tests” are of no interest whatsoever to me. With modern sensors performance is good enough from any brand. It is the camera that counts.

Share this post


Link to post
Share on other sites

I must confess that, as I have no intention of moving to Sony -horrid UI- these “tests” are of no interest whatsoever to me. With modern sensors performance is good enough from any brand. It is the camera that counts.

+1

 

Camera body is not just a sensor.

Share this post


Link to post
Share on other sites

And back to topic, next I’ll do ISO 3200 on the CL but adjust the aperture on the α7R III to the APS-C equivalent for depth of field. In the Baby Jesus picture above, exposure was pushed a bit on the CL as aperture on both cameras was f/3.2. So it will be ISO 3200 f/1.4 on the CL and ISO 5000 f/2.0 on the α7R III. That should give similar exposure and then we’ll compare noise.

Edited by Chaemono

Share this post


Link to post
Share on other sites

Interesting thank you but i'm not sure to follow your comparison with different isos and apertures. The crop factor does not come into play here. If the available light requires f/2 at 3200 iso on FF cameras the same light requires f/2 at 3200 iso on APS-C as well. Am i missing something?

Share this post


Link to post
Share on other sites

Interesting thank you but i'm not sure to follow your comparison with different isos and apertures. The crop factor does not come into play here. If the available light requires f/2 at 3200 iso on FF cameras the same light requires f/2 at 3200 iso on APS-C as well. Am i missing something?

 

Edit

Edited by Chaemono

Share this post


Link to post
Share on other sites

Interesting thank you but i'm not sure to follow your comparison with different isos and apertures. The crop factor does not come into play here. If the available light requires f/2 at 3200 iso on FF cameras the same light requires f/2 at 3200 iso on APS-C as well. Am i missing something?

Agree. Too many variables make for meaningless tests.

Share this post


Link to post
Share on other sites

Interesting thank you but i'm not sure to follow your comparison with different isos and apertures. The crop factor does not come into play here. If the available light requires f/2 at 3200 iso on FF cameras the same light requires f/2 at 3200 iso on APS-C as well. Am i missing something?

Theoretically, you may be right. But if one wants an equivalent depth of field, same fall off from the focus point, same lush, beautiful bokeh to compare APS-C with FF, one needs to adjust for the crop factor in aperture as well. But I'm sure you know this. Be patient. Many pictures to follow soon. I was generous and gave the α7R III the 50 Apo Summicron in order not to handicap it

. They are not tests. They are comparisons of a similar look between APS-C and FF. We need to keep everything constant in order to solve for noise. 

Edited by Chaemono

Share this post


Link to post
Share on other sites

First, only WB adjusted on the α7R III to try match the CL

 

Uncompressed JPEGs here: https://www.smugmug.com/gallery/n-qDx8fd/

 

α7R III + 50 Apo Summicron-M

ISO 5000 f/2.0 (wide open) @1/3200 sec.

 

 

CL + 35 Summilux-TL (but it's really more like a Summicron in the FF world

 )

ISO 3200 f/1.4 (again, wide open) @1/3200 sec.

Share this post


Link to post
Share on other sites

And now with the Shadows lifted by +100

 

Uncompressed JPEGs here: https://www.smugmug.com/gallery/n-qDx8fd/

 

α7R III, Exposure +0.10, Shadows +100, Blacks +20, NR +30, LR default sharpening

ISO 5000 f/2.0 (wide open) @1/3200 sec.

 

 

CL + 35, Highlights +10, Shadows +100, NR +30, LR default sharpening

 

ISO 3200 f/1.4 (again, wide open) @1/3200 sec.

Share this post


Link to post
Share on other sites

And now the crops to look for that ugly ISO 3200 CL noise.

 

I'd say the α7R III sensor is terrific but it doesn't come close to the best FF sensors like in the D850 (Sony sensor). In fact, it's close to the best APS-C sensors when it comes to tonal range and color sensitivity at high ISO like the one found in the CL (probably also from Sony 

) IMO.

 

Uncompressed JPEGs here: https://www.smugmug.com/gallery/n-qDx8fd/

 

α7R III

ISO 5000 f/2.0 (wide open) @1/3200 sec.

 

 

CL

ISO 3200 f/1.4 (again, wide open) @1/3200 sec.

Edited by Chaemono

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...