CCD vs CMOS

[Jeff S]

Advertisement (gone after registration)

On 11/10/2023 at 10:39 AM, Al Brown said:

One is Audi and the other is a BMW. 

Only when set to auto mode.

Jeff

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[adan]

Short version - There are a huge number of steps between light falling on a sensor package (layer of silicon plus other layers) - and a digital image we can see and evaluate. ALL those steps will heavily affect what we eventually see from a sensor, probably far more so than the simple (and simplistic) CCD/CMOS dynamic.

 Extended version (NB: this is a simplified "schematic" description - quibblers take note 😁 ) -

ALL of the following influence the final image - and ALL of them are made by human engineering choices. You can't wander through the woods and find a "natural" CCD or CMOS chip. Someone designed the sensor, and the firmware/software that supports it, to make the specific sensor for the product in hand.

- cover-glass wavelength transmission capability (recall the M8 - leaked a lot of infrared)
- Bayer RGB color filter wavelength spectral transmission range. *
- charge-to-voltage and voltage-to-digital conversion algorithms (CMOS generally does the first on the sensor in each pixel, CCD dumps the charges for the camera to convert)
- the inherent dynamic range of the sensor - how much charge or voltage, per pixel, can fit in, between "empty" and "overflowing."
- tone curves that convert the linear sensor output (1 photon > 1 electron) to a more film-like S-shaped contrast curve. And yes, even for .raw/.DNG images there can be a tone curve specified. **
- for color images, de-mosaicing algorithms *** to blend pure RGB pixels into one another to get all colors in every pixel.

Keep in mind that we never actually see a "raw" picture (except with specialized software). If we could, a color image would look like a chessboard of grays (just brightnesses). It must be converted into a full RGB of HSL image for most software to even display it. What we see on the camera or computer screen are "on the fly interpretations" that our eyes and brains can deal with.

- Color profiles chosen as the default, by Leica (Embedded) and by post-processing software engineers (Adobe, C1, you name it). Which specify the amount of orange or purple (Hue) in a given "red," for example, and how intense it is (per-channel Saturation). Blues more purple or more cyan, greens more yellow or more cyan.
- White balance, chosen by settings, or the camera ("as shot") or the post-processing engineers (Adobe "Auto").

I've noticed, checking some M11 .DNGs, that they come into Adobe Camera Raw with a massive amount of magenta tint in the default "As shot" WB. Which will certainly make the images look "M9 CCD" to some extent. But are not the native color rendering of the M11 CMOS sensor.

- Whether post-processing software uses the "embedded" tone curves or color profiles by default - or whether the user has to intervene (if possible) to get the "most native" camera/sensor output..

On the whole, all those factors will have far more to do with determining the apparent character of a sensor, than whether it is CCD ot CMOS.

Except maybe the DR component. At one time, early CMOS sensors had less DR and less ISO capability (more noise) than CCD, due to the non-light-sensitive circuitry cluttering up the front of each pixel (poor "fill factors"). But CMOS pulled ahead somewhere around 2010 - some of the "puch" of CCDs comes from their lower DR. Just a Kodachrome had more "punch" than color negs - at the cost of inpenetrable shadows at times.

______________________

* Reducing the "purity" of a Bayer filter can improve ISO response by letting more light of other colors reach a given pixel. Same reason a Monochrom can reach higher ISOs than a Bayer-filtered M9/10/11. But will reduce the color discrimination - again, the Monochrom is "color-blind" on its own. And a 2009 CMOS (e.g. Canon 5DI/II) has duller colors than a 2009 CCD (e.g M9).

Of note, Leica changed the red filter specs slightly, in between the M8 and M9 sensors (both identical Kodak CCD architecture otherwise). An engineering choice.

**This image shows the default tone mapping LEICA chose to apply for the M10 vs. the M(typ240) - (both CMOS), compared with the curve Canon chose for the 6D (also CMOS).

Sadly, I've not yet found an M8/M9 CCD camera for comparison testing. I hypothesize it would look more like the M10/Canon curves - or even steeper and "punchier," and with less DR (less leeway/steps between pinning out at black (0) or white (255).

Y = over/underexposure of a gray card in stops, X = actual 256-bit brightness output via the algorithm.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

*** An "artistic interpretation" of how a raw pixel array (a chessboard of gray "brightness values," at left ) gets resolved by an algorithm into a full-color picture.

[M11 for me]

Thank you very much for this explanation. Great to read.

[Adam Bonn]

I would say as I think is already mentioned is that CCD is old tech and CMOS is not only newer but still being developed.

So one V the other is more about old V new products... 

My hunch is that if CCD was being as developed as much as CMOS is then the story would be different...

Older cameras tend to have files with less ISO performance, more signal noise, less malleability, less pixels and less DR.

IMHO this tends to give older (CCD) files a more organic noisy look, the lower DR manifests itself as greater shadow contrast, giving a crisper look to the colours and a faster (well lower in the DR range) roll off into the highlights.

This leads to files that often look very pleasant when first imported, that pick up nicely with minimal edits, then one tends to stand back and say ah they don't make 'em like they used too which is true, you get more these days - but you have to work harder to get it in post.

There's a great many folks selling profiles and presets and even applications where you can design your own camera profiles for anyone looking to get a leg up in the colours and tones of their images

[overexposed]

On 11/8/2023 at 1:11 PM, Anthony MD said:

So much discussion on this subject with more supporters of CCD than CMOS.

CCD supposedly has a more filmic appeal than CMOS.  Is all this attributed to the sensors only or do other factors determine final image rendering…?

i hated the Sony A7RII files, i hated the M240-P files, i hated the Fuji GFX100 files(not really hated but well..) i loved my P45+ files, my IQ180 files and currently my P65+ files.

I wonder why?

even though i dont hate my M10-R files, i dont love them. there is always something different about CCD files.

i always wondered if its the BSI sensors that do something different, but then the M240-P sensor was not BSI

 

no no no, CCD all the way!

[Anthony MD]

9 hours ago, overexposed said:

i hated the Sony A7RII files, i hated the M240-P files, i hated the Fuji GFX100 files(not really hated but well..) i loved my P45+ files, my IQ180 files and currently my P65+ files.

I wonder why?

even though i dont hate my M10-R files, i dont love them. there is always something different about CCD files.

i always wondered if its the BSI sensors that do something different, but then the M240-P sensor was not BSI

 

no no no, CCD all the way!

Thorsten Overgaard had said in his review on the Leica M-D 262, the M-D had a different sensor than the M240 but the same as the M60!

He talked with the product manager in Germany to find this information.

I love the results from the MD…!

[overexposed]

Advertisement (gone after registration)

9 hours ago, Anthony MD said:

Thorsten Overgaard had said in his review on the Leica M-D 262, the M-D had a different sensor than the M240 but the same as the M60!

He talked with the product manager in Germany to find this information.

I love the results from the MD…!

i HIGHLY doubt its a different sensor!

[Anthony MD]

4 hours ago, overexposed said:

i HIGHLY doubt its a different sensor!

When Thorsten compared his images from his M240 with his MD 262 he realized they were different.  So he asked the product manager at Leica Germany why.  The answer was the MD has a different sensor than the 240.  The MD 262 and M60 have the same sensor…!

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Edited December 15, 2023 by Anthony MD

[Anthony MD]

37 minutes ago, Anthony MD said:

When Thorsten compared his images from his M240 with his MD 262 he realized they were different.  So he asked the product manager at Leica Germany why.  The answer was the MD has a different sensor than the 240.  The MD 262 and M60 have the same sensor…!

 

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Edited December 15, 2023 by Anthony MD

[overexposed]

any chance we can get RAW files?

240 and 262 @ iso 100 and 1600?

 

i still cannot imagine its a different sensor, maybe they adjusted the processing line, but if the noise looks identical i would say its the same

[Anthony MD]

57 minutes ago, overexposed said:

any chance we can get RAW files?

240 and 262 @ iso 100 and 1600?

 

i still cannot imagine its a different sensor, maybe they adjusted the processing line, but if the noise looks identical i would say its the same

Do you think the M240 has the same sensor as the M60…?

[overexposed]

1 minute ago, Anthony MD said:

Do you think the M240 has the same sensor as the M60…?

yes

[Anthony MD]

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Anthony MD]

26 minutes ago, overexposed said:

yes

Why would a camera that cost almost four times as much as another have the same sensor?
Doubt that very seriously…!

Edited December 15, 2023 by Anthony MD

[overexposed]

you are new to the leica game it seems

[Anthony MD]

30 minutes ago, overexposed said:

you are new to the leica game it seems

I did a lot of research before my purchase of the Leica M-D 262…!

[Anthony MD]

32 minutes ago, overexposed said:

you are new to the leica game it seems

What game is Leica playing…?

Edited December 15, 2023 by Anthony MD

[Anthony MD]

2 hours ago, overexposed said:

any chance we can get RAW files?

240 and 262 @ iso 100 and 1600?

 

i still cannot imagine its a different sensor, maybe they adjusted the processing line, but if the noise looks identical i would say its the same

I don’t have a 240…!

Edited December 15, 2023 by Anthony MD

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[Anthony MD]

1 hour ago, Anthony MD said:

I don’t have a 240…!

I purchased the MD 262 since I only shot film for many years. The MD shoots like a film camera.  If it had a screen I wouldn’t have bought it…!

Edited December 15, 2023 by Anthony MD

[pippy]

For many years I had an M9-P and loved the quality of the files it provided and, in turn, the quality of the prints which I could make from those files. Then, a number of years later, I bought an M-D 262 and, to this day, love the files it provides (...etc...). Tangentially I also have an M Monochrom and love the files that it provides (...etc...).

Where am I going right?

Philip.

Edited December 16, 2023 by pippy