Leica Sensors - CCD vs CMOS

boojum

Top Veteran
I posted this in an old CD vs CMOS thread and I am not sure it was seen. If it was and was boring, skip this. If it was not, read on. I plan to do more tests.

I have been interested in this discussion and am, honestly, personally fond of CCD over CMOS. But, laboring under the weight of education in these forums I decided to test the differences. I live in an area pretty much starved of sunshine, especially this year. However we did get a break with some sunshine in the last few days, now gone.

I shot a series of three images for each lens, f/2.8, f/5.6 and f/11. The lenses were Canon 28mm LTM f/2.8, Canon 35mm LTM F/2.0, Canon 50mm LTM f/1.8 and Jupiter 85mm LTM f/2.0. The cameras were M9 (CCD) and M240 (CMOS). The images are uncompressed DNG files without in-camera lens correction and at ISO 200, aperture priority. I was honestly surprised at the outcome of the raw DNG images.

The house in the photos is in the Alderbrook section of Astoria, Oregon. It is an interesting community, off by itself with some artists in it. This house was once a worker's shotgun home for mill workers. There is a string of them along this road. This particular one caught my eye for its colors and the quality of the rehab. Twenty years ago this house was borderline derelict. It is raised off the ground over what is called here a "summer basement." The house is on poles and beneath the house is packed earth or concrete. The vertical boards enclose the empty space which has no use other than to keep the home off the damp ground. The water table is very high and there are small drainage canals along here. The Columbia river is a few hundred yards behind the home.

These are the links:

Proton #1 Sensor Test - Google Drive


drive_2020q4_32dp.png
drive.google.com

https://drive.google.com/drive/folders/1OserqhHlhaySfAAWBU-SGpPNCmBQJbvj?usp=sharing
 
I looked at the results of the 28mm lens- and my belief is that the CMOS sensor showed more vignetting. The CCD sensors used in the M9 are Back-Side-Illuminated, and have almost 100% fill-factor with offset microlens arrays to collect light off angle. The CMOS sensor used in the M240 is front-side-illuminated. The difference- collecting light coming in at steep angles. This is also a problem for Fast lenses. Longer focal length lenses, stopping down the aperture, and retro-focus wide-angles alleviate this problem as the light falls on the sensor almost directly. This is why cameras want to have the lens type "dialed in" to run non-uniformity-correction on the image.
 
Hmm. That explains some of the unexpected vignetting that I get with my M240.

It does leave me wondering which of my pile is retro-focus and which is not, which will benefit most from stopping down, etc.

When you say "dialed in" is that the lens detection? Or is it something else I can do on my M240?

Thanks!
 
Dialed in- lens detection, either sensed or manually entered.
This raised a problem for me. I bought the LTM to M adapters with the grooves for 6-bit encoding and coded to what I thought was the most appropriate Leica lens codes. But I got dark, murky and sometimes dramatic results, more results than I had hoped for. Because of this I have stopped encoding lens profiles. Perhaps I can manually encode each lens to see what is the best choice for encoding and then paint the 6-bit grooves to reflect that. Something to consider short of buying a lot of Leica lenses.

One thing to point out is that the lens profile affects both JPG and DNG. On the one hand that makes good sense as both files are affected by the lens vignetting. On the other hand I thought that raw, in this case DNG, meant pristine and unmodified. I was wrong. The devil is in the details.
 
The saturation, contrast and vibrancy is really cranked up on the M240 compared to the M9. Especially in the yellows on the railings. . I'm comparing image 317 to 146 in DXO PL.

I've long preferred CCD to the Cmos sensors for their more "organic" or "textured" look , whether that's real or imagined on my part doesn't really matter, it's just how I feel. But i do think later cmos sensors are reverting to more of a ccd look in some cases. I plan to one day buy an M8 , pair it with a Voigt 28mm Ultron "vintage" lens and be happy shooting CCD when the mood takes me.
 
i do think later cmos sensors are reverting to more of a ccd look in some cases.
Really? What cameras are we talking about here? I always thought the early cmos sensors in the Nikon cameras from, say, D700 up to D800 had more of that look, certainly nicer to my subjective eye than later and more ‘capable’ cmos cameras generally.
 
My M10-R with the CMOS sensor derived from a medium format leica sensor has a great color separation, fine details, superb shadow depiction. It is a big step beyond my previous CMOS-M240. But I always looked with awe at the organic, plastic quality of my first M8 CCD pictures.

The M10-R might compete picturally with the M9. But LRc has to do a lot of work, and uses its M10-R profile. But I am afraid the M9 with its CCD is still having the edge - images are more brilliant.

A problem I see is that these modern sensors - though measured as having low noise - still exhibit so much noise that the plasticity of objects is reduced, i.e. they become flat again.
 
Last edited:
I don't know. Truly, you could pour all I know of camera technology into a thimble and still be able to jump in for a swim. I was surprised to see the better saturation of the M240 over the M9 in those two images. I use 308 from the M9 and 137 from the M240, both f/11 on the Canon 28mm f/2.8. The colors look more saturated and more as I remember them in the M240 photos. But the mind is a tricky organ. Maybe they look more like I want them to look. Were I to want a print I would go with 137 just for the color saturation.
 
I looked at the results of the 28mm lens- and my belief is that the CMOS sensor showed more vignetting. The CCD sensors used in the M9 are Back-Side-Illuminated, and have almost 100% fill-factor with offset microlens arrays to collect light off angle. The CMOS sensor used in the M240 is front-side-illuminated. The difference- collecting light coming in at steep angles. This is also a problem for Fast lenses. Longer focal length lenses, stopping down the aperture, and retro-focus wide-angles alleviate this problem as the light falls on the sensor almost directly. This is why cameras want to have the lens type "dialed in" to run non-uniformity-correction on the image.
Interesting.
As I moved to the M10-R I noticed that the vignetting is less too. I even have the impression lenses are sharper (not only due to the higher resolution) but the way they depict is nicer (the main bokeh circle for some lenses). maybe I should sit down and show some images.
The Lumix S5 with its BSI is really a marvel in use, such as for the Jupiter-9. It looks "totally" different now.
 
Not a fair comparison?

First a good standard old-fashioned CMOS shot:
LR013577.jpg
Join to see EXIF info for this image (if available)

Shadows increased a bit, to be honest. And some other things in the M10-R profile of LR.

Then a modern BSI-sensor
PS5_1207.jpg
Join to see EXIF info for this image (if available)

The same picture with the first camera showed it was muddier, but of course, the setting sun changes every minute.
Again and again, I see a BSI sensor does SHINE. More by itself, without processing. Is my experience. (Don't look at sharpness; with my profile settings here that is gone after uploading.)
And it does look more like a CCD.
 
The hard point here is the difference between "more like" and "just like." I favor the M9 sensor a lot. I just posted a string of five shots I took with an M9 and an Amotal, an unbeatable combo for me. But it may just be me. For those times when I want the best color I will go with the M9. When I want to be sure I get a pic I have an auto everything Sony with good color, but not quite as good as the M9, that I can use.
 
OK, M9 vs M11.


The sensor is the thing. At least in my eyes (intentional double meaning). I am very fond of my M9 and really love how it performs with some old lenses from the late 40's (Amotal) and late 50's (KMZ J8). The M240 has nice color but it is not as "sculpted" as the M9, in my eyes. The Pixii is quite good and I will get David Barth to come on-line and get mine up and running again as it does make nice photos.

But back to the M9 vs the M11. There is a lot of hoo-ha about the sensor. Release flak reminds me of Donna Summer's Fame: "I'm gonna live forever, I'm gonna learn how to fly . . . " That it can live in three different pixel counts is interesting. That the camera has new processing software is interesting. And all the rest is interesting but can the image compare to the image of the M9? How do trey compare side by side? Has anyone done any tests on this? I'd love to see some tests. Have they been done? Anyone??

NB - I posted this on RFF so those who read both boards are not seeing double.
 
The M11 sensor is Back Side Illuminated, as are the CCD's used in the M9 and M Monochrom. It should do better accepting light coming in at sharp angles than the M10 and M240.

But- no data sheets have been published. No details of the pixel binning, etc.
If you are that curious- rent one.

The M9, M Monochrom, and even M8 are well matched to the lenses I use. The Df is well matched to the lenses I use. Getting an M11, even an M10m or M10r- no interest. Bigger files to resolve blur.
 
I have a 5 MPx Olympus E-1 (CCD), which was renowned for its colour signature.

I also have a 10 MPx E-510 (CMOS) which has an all but identical colour signature.

However, my 12 MPx E-30 (CMOS) has native colours like a Nikon - i.e. greener than green grass and cyan skies ...

My mFTs collection all have accurate colour signatures.

Go figure ...
 
The M11 sensor is Back Side Illuminated, as are the CCD's used in the M9 and M Monochrom. It should do better accepting light coming in at sharp angles than the M10 and M240.

But- no data sheets have been published. No details of the pixel binning, etc.
If you are that curious- rent one.

The M9, M Monochrom, and even M8 are well matched to the lenses I use. The Df is well matched to the lenses I use. Getting an M11, even an M10m or M10r- no interest. Bigger files to resolve blur.
Thanks for that BSI info and relation to CCD. We both prefer CCD, and I guess we both prefer them with old lenses. I sure like what I am getting. The M11 could be an interesting comparison and maybe, just maybe, may slay the CCD dominance of image quality. But I need to see it shown and proven before I will believe it. Side by side would be nice. One can hope.
 
I've long preferred CCD to the Cmos sensors for their more "organic" or "textured" look , whether that's real or imagined on my part doesn't really matter, it's just how I feel. But i do think later cmos sensors are reverting to more of a ccd look in some cases.
I don't know if this is what was meant, but in terms of "textured" looks the M4/3 sensors, from about the Olympus EM5 and Panasonic GX80/85 onwards, have a gritty look that reminds me of how my CCD Nikon could look at times... And the high resolution Ricoh and Pentax sensors in the KP, K-1 and GR III cameras, with the pre-processor which reduces noise in RAW files, have an unobtrusive grain at high ISO which isn't as toothy as the smaller sensors but does seem pretty organic. I think a lot of it has to do with higher luminance noise and low chroma noise. Generally that produces what I'd call "organic" looks. Colors, though, at or near base ISO, are seldom quite as brilliant as they were with CCD sensors... but remember how easy it was to overfill certain color channels? Specifically the reds with Nikon.
 
I was hoping that the new model’s files might bear a close comparison to those of the M9 but to my subjective eye that doesn’t appear to be the case, well from what I’ve seen so far at least. I can only imagine that it’s the use of the vintage lenses that might bring that about. Maybe. I’d certainly look forward to seeing some proper comparisons.
 
my belief is that the CMOS sensor showed more vignetting. The CCD sensors used in the M9 are Back-Side-Illuminated, and have almost 100% fill-factor with offset microlens arrays to collect light off angle. The CMOS sensor used in the M240 is front-side-illuminated. The difference- collecting light coming in at steep angles.

I really should pull out the M9 and take some shots with it. I haven't used that or the M240 for a long time because they don't have connectivity with the Fotos app which is how I edit/process my photos these days.

But the reason why I want to try the M9 again is because I've always got heavy vignetting on my M10. Less so on the M240.
 
Back
Top