The look of a CCD versus a look of the CMOs in medium format

[ErikKaffehr]

Hi,

Interesting! This effect you noted, it is a the pixel level?

Would be nice if you post a sample.

Best regards
Erik

That is certainly my experience. I shoot with a Hasseblad H3D31ii (microlensed). Aliasing often shows up a "rainbow" speckling on skin micro-highlights where pores are hit by hard light, for example.

I absolutely LOVE the resulting shimmer- it is like applying a very, very fine subtle rainbow glitter makeup with an airbrush. (Which we've tried, and which is a damn sight harder to create and photograph for real). 

Even when sent through a skin smoothing pass and downsampled in post, an impression of the shimmer remains - contributing to what others have called the skin texture.

This is an out-and-out technical ERROR. This colour detail is aliased, not real.

Allied with the extra apparent sharpness and micro-contrast that shooting without an anti-alising filter makes, plus the post-production diffusion techniques which artificially preserve the fine detail compared with the way you'd record it with an AA filter or on-lens diffusion, the net result is my preferred rendition of the skin of pretty girls. (Which is 98% of my professional photography).

The sum of errors and imperfections can often get you to a more interesting creative place than a technically accurate rendition.

More than anything it is the way the camera renders skin texture that makes me choose it over my Canon when I have enough light to work with. The Canon gives nice enough results, just not the "fairy dust turned up to 11" look that I particularly like.

By contrast, to get the look I like best from motion photography I tend to use diffusion and smoke and blurry old stills lenses when shoot motion- I think the Hasselblad "sparkle" would probably look ghastly with shifting aliasing in motion. It just looks stunning in a single frame

The only answer is, as always, use the tool that gets you the result you want. And hire before you buy!

  Cheers, Hywel

[Hywel]

Here we go. Not safe for work, contains nudity:

http://www.restrainedelegance.com/preview/reh_20140424_1293222.jpg

It's fighting with over-sharpening and JPEG compression (I quickly did the export to full-sized JPEG, normal deliverable is a smaller JPEG for me) but you can see the rainbow shimmer effect on her face.

Cheers, Hywel.

[eronald]

Here we go. Not safe for work, contains nudity:

http://www.restrainedelegance.com/preview/reh_20140424_1293222.jpg

It's fighting with over-sharpening and JPEG compression (I quickly did the export to full-sized JPEG, normal deliverable is a smaller JPEG for me) but you can see the rainbow shimmer effect on her face.

Cheers, Hywel.

I like the picture. Nothing wrong with it as an image. Technically speaking you need to underexpose by a couple of stops and play with the tone curve, but probably you'll lose exactly the effect which makes the image interesting. And I think there are bunch of non linear effects to consider here, micro speculars, anti-blooming cicruitry, flare on the sensor protection glass, clipping and oversharpening in post etc. But who cares as long as you manage to make the result look alive?

Let me use a cooking analogy although I'm not much of a cook. We are all aware that sometimes one takes good steak and chars the edges, drowns it in sauce or turns it into minced meat. But to make a decent meal one still wants the original steak to be decent. The question on the table here is whether Cmos is as good a point of departure for "cooking" as CCD. I say CMOS is ok for most imagery, eg landscape, architecture and interiors, but if one wants to do images such as yours CCD is actually a better place to start from.

Edmund

[ErikKaffehr]

Hi,

Thanks a lot. I guess I see what you mean.

Best regards
Erik

Here we go. Not safe for work, contains nudity:

http://www.restrainedelegance.com/preview/reh_20140424_1293222.jpg

It's fighting with over-sharpening and JPEG compression (I quickly did the export to full-sized JPEG, normal deliverable is a smaller JPEG for me) but you can see the rainbow shimmer effect on her face.

Cheers, Hywel.

[Hywel]

Yeh, I know it isn't a great shot, it has clipped (my inexperience with African skin tones is to blame there, I was concerned about losing detail in the shadows and over-compensated). And I just exported it quickly at full-size so it's over-sharpened, the JPEG compression is too heavy. I just had it to hand on a card as an example of the rainbow shimmer effect I was talking about.

Cheers, Hywel.

[bcooter]

My feeling is we're chasing our tail.

Yes high iso is great for photojournalists some wedding photographers and sports guys, but overall, it's just a lot of smoothing and detail killing to the point I don't think any camera does more than 1000 iso and holds everything and if it does it becomes rather generic and digital looking.

These are three images.  Two from the past (one just retouched).

the first is from an Aptus 22 and a Contax, window light no retouching, small jpeg out of camera raw just color corrected in photoshop, no layers, no masking.


The second from a 12mpx Canon 1ds, cropped to about 2/3's of the frame.   We just retouched it for a reference video and decided to work in high resolution.
This was shot flash and obviously has intense retouching, but very little added to the look other than the beauty look of face and hair and wardrobe.


The third recently shot with the Leica S2 see see d version, using hmi's I think at 320 iso, but maybe 640 I don't remember.
It has layers of working shadows, some clean up and obviously masking the background to black.
Other than that very little skin or face retouching.


Now to me, as I went down the line in cameras, from 1ds, 2, 3 and x, Nikon D3, Leica's S2 and M8, and my 30+ and 21+ phase backs, I've noticed how the cmos cameras get newer, the less film like look, the more digtial or plastic look of the image and honestly I think rather generic.

But that's just one person's opinion but these samples represent cameras from 10, 7, and 3 years old and they all work close to the same.  

I don't think it's always cmos vs. ccd, I just think the makers get caught up in megapixels and ultra high iso and something changes in the look and the makers are no longer trying to emulate film, they're trying to make smooth, broad lattitude images.

I personally think a ccd file works deeper, more film like, more texture, but I also use ccd cameras at higher isos, never under 200 iso, usually at 320 to 640/800.



IMO

BC

[amsp]

Funny enough, the only digital cameras I've held onto over the years is the original (12MP) Canon 1Ds and the P25, exactly because of the look I get from them. I could care less if it's pixel size, CCD or magic dust that creates it, all I care about are the results I see.

[billthecat]

I like CCD better and I assume it is due to CMOS sensors having pixels that reduce noise early. Older CMOS didn't do that and it had a better look. The 1Ds was like a kind of film. (I sold my 1Ds for newer Canons.)

With audio, aggressive efforts to reduce noise takes feeling out of the sound. So I assume with images the early noise reduction squishes a thin layer of information off the image making it less alive.

Bill

[BJL]

I like CCD better and I assume it is due to CMOS sensors having pixels that reduce noise early. Older CMOS didn't do that and it had a better look. The 1Ds was like a kind of film. (I sold my 1Ds for newer Canons.)

With audio, aggressive efforts to reduce noise takes feeling out of the sound. So I assume with images the early noise reduction squishes a thin layer of information off the image making it less alive.

Bill

I am fairly sure that Canon has been using "early noise reduction", include correlated double sampling, from its first DSLR, and so in the 1DS; see this quote from http://www.canon.com/technology/approach/history/digital_tech.html, with my underlining, which indicates that CDS was used already in the D30, before the 1Ds.

Subsequently, Canon decided to use the CMOS sensor the company developed in-house as the image sensor in the digital SLR camera planned for release in 2000. At the time, imaging elements in Digital SLR Cameras requiring high image quality and high sensitivity were primarily CCDs, and the incorporation of a CMOS sensor was revolutionary.
Compared with CCD image sensors, although CMOS sensors generally offered the advantages of low power consumption, fast reading speeds, and low cost, their high levels of noise and poor sensitivity were pointed out as disadvantages at the time. To overcome these shortcomings, the company thoroughly reviewed all processes required for manufacture and developed a 4-transistor pixel structure and a correlated double sampling noise-cancellation system, thereby successfully reducing noise.
Meanwhile, it also became necessary to produce clean transistors with a leak current approximately 1/1,000th of that of transistors used in standard PCs and memory elements. Leaking of current is caused by heavy-metal contamination during the manufacturing process and irregularities in the structure of silicon crystals; however, the establishment of thorough cleaning and processing technologies for the removal of metal contamination paved the way for the launch of the EOS D30.

The other early noise reaction method is the very early amplification (as part of the signal transfer from photosite to sensor's edge) which is a defining feature of active pixel CMOS sensors, which is the only type of CMOS sensor ever used in DSLRs as opposed to the older "cheap and noisy" type that Canon refers to in the quite above.


I suspect that what you and many others see and prefer is not due to "CCD vs CMOS" or "CMOS with early NR vs CMOS without", but instead due to effects of difference like larger photosites: the CCD vs active pixel CMOS comparisons also have the CCDs with large pixels and in larger formats and/or with more expensive prime lenses, and of course those older CMOS sensors had smaller photosites than the new ones.  This of course produces a bias if images are compared at "100%" so that the images from newer CMS sensors are being viewed at a greater degree of enlargement, but it might also be the case that smaller photosites have inherently more problems of inter-site light leakage and related color errors, perceptible even in fair comparisons at equal degree of enlargement.

[MrSmith]

they all look the same on an iPad or a phone.

[eronald]

AFAIK the 1Ds was both the first Canon fullframe and also last EOS 1 chip of its technology; afterwards a major transition took place. I remember this from my interviews with the Canon people in charge of the EOS1 series development. Then came the 1DII, the 1Ds2 etc, which had completely different looks and were subject to "plastic skin", the 5D which was nice, the 1Ds3 which I own which is ok, and the 5DII which is very uneven.

Edmund

I am fairly sure that Canon has been using "early noise reduction", include correlated double sampling, from its first DSLR, and so in the 1DS; see this quote from http://www.canon.com/technology/approach/history/digital_tech.html, with my underlining, which indicates that CDS was used already in the D30, before the 1Ds.The other early noise reaction method is the very early amplification (as part of the signal transfer from photosite to sensor's edge) which is a defining feature of active pixel CMOS sensors, which is the only type of CMOS sensor ever used in DSLRs as opposed to the older "cheap and noisy" type that Canon refers to in the quite above.


I suspect that what you and many others see and prefer is not due to "CCD vs CMOS" or "CMOS with early NR vs CMOS without", but instead due to effects of difference like larger photosites: the CCD vs active pixel CMOS comparisons also have the CCDs with large pixels and in larger formats and/or with more expensive prime lenses, and of course those older CMOS sensors had smaller photosites than the new ones.  This of course produces a bias if images are compared at "100%" so that the images from newer CMS sensors are being viewed at a greater degree of enlargement, but it might also be the case that smaller photosites have inherently more problems of inter-site light leakage and related color errors, perceptible even in fair comparisons at equal degree of enlargement.

[BJL]

AFAIK the 1Ds was both the first Canon fullframe and also last EOS 1 chip of its technology; afterwards a major transition took place. I remember this from my interviews with the Canon people in charge of the EOS1 series development. Then came the 1DII, the 1Ds2 etc, which had completely different looks and were subject to "plastic skin", the 5D which was nice, the 1Ds3 which I own which is ok, and the 5DII which is very uneven.

I have never read anywhere about such a major transition, so if you have any published sources I would be interested; otherwise, I remain skeptical: didn't you previously claim that the earlier CMOS sensors did not use CDS, and blame the "plasticky look" in part on CDS?

The confirmed facts including yours above fit better with my hypotheses about photosite size and related differences in the degree of enlargement(*) at which evaluations are done: you seem to put the original 5D in the "nice" category and the 1DsII in the "plastic skin" category, but the 5D came with a new 12.8MP sensor almost a year after the 1DsII with its 22MP sensor [August 22 2005 vs September 21 2004], so what ties the claimed 5D look to the look of the earlier 11MP 1Ds and distinguishes it from the 1DsII with its "completely different look" is larger photosites, not the different eras of the sensor technologies.


(*) I use "degree of enlargement" in its standard optical sense of the ratio between final displayed image size and the size of the image formed on the sensor, not some goofy "pixel percentage".

[synn]

they all look the same on an iPad or a phone.

Subjective, I guess. Some people can, some people can't; even on a 40" print.

I have a whole bunch of images in the opening slideshow on my site: http://www.sandeepmurali.com/
Taken with everything from MFD to a 1" sensor with everything in between (Some CCD, some CMOS). Some people could tell the MF (CCD)  images apart at the first glance.

[ErikKaffehr]

On the other hand, anyone having slightest amount of residual vision can tell a bad picture from a good picture even on an iPhone.

Best regards
Erik

Subjective, I guess. Some people can, some people can't; even on a 40" print.

I have a whole bunch of images in the opening slideshow on my site: http://www.sandeepmurali.com/
Taken with everything from MFD to a 1" sensor with everything in between (Some CCD, some CMOS). Some people could tell the MF (CCD)  images apart at the first glance.

[synn]

On the other hand, anyone having slightest amount of residual vision can tell a bad picture from a good picture even on an iPhone.

Best regards
Erik

Indeed.
Especially, test charts and aperture series.

[MrSmith]

I have my doubts.

Me too, on the whole most people are visually unaware, including some photographers.

[Fine_Art]

I dug up an old CCD shot to remind me of why I liked it.

[The View]

There was a different look between my Kodak sensor based digital backs and my dalsa sensor digital backs and both are CCD.  There are a lot of people saying the CMOS and CCD chips have got to be the same from some theoretical point of view - maybe, but in practice I don't find that IMHO.  I have two monitors on my setup a NEC and an Apple and no matter how hard I work with the spectra view calibration, I can't make them have the same look.  I can match the colors mostly, but the look is still different.

Getting back to cameras and CCD vs CMOS - its a complete system - not just the sensor chip but all the on board processing plus the rest - and of course even the lenses can affect color and look.  

I did lots of tests with the d800e vs my AFi-ii 12 and while the D800e detail in the luminous channel is very good, the color is flat as if it were averaged.   The color tonality does impact the look and is still an area where I think CCD chips excel.  Does it have to be that way? No, probably they could be made more similar but every designer using the CMOS seems to be looking for higher ISO performance instead of optimizing IQ at base.  The CFA choices and other on board processing probably account for some of the differences in look (without talking about lenses). I suspect whatever profiles the manufacturer use also impact the look.

I have the same impression that choosing a certain camera alone will give you a certain look just like choosing a certain film would.

And I can totally understand that you can't even match the looks of two systems if you tried hard. (There's a reason why some people love Leaf backs, and others prefer Hasselblad or Phase One - and I don't think it just boils down to how much they like the camera, the lenses, the company. I think many photographers choose a system by choosing a look).

Even with the same type of sensor.

And even when you process an image with Lightroom and Capture One, there is a certain difference of look.

And I also feel the CCD feels it's capable to a clearer and cripser and more natural looking image that a CMOS sensor.

[The View]

That is certainly my experience. I shoot with a Hasseblad H3D31ii (microlensed). Aliasing often shows up a "rainbow" speckling on skin micro-highlights where pores are hit by hard light, for example.

I absolutely LOVE the resulting shimmer- it is like applying a very, very fine subtle rainbow glitter makeup with an airbrush. (Which we've tried, and which is a damn sight harder to create and photograph for real). 

Even when sent through a skin smoothing pass and downsampled in post, an impression of the shimmer remains - contributing to what others have called the skin texture.

This is an out-and-out technical ERROR. This colour detail is aliased, not real.

Allied with the extra apparent sharpness and micro-contrast that shooting without an anti-alising filter makes, plus the post-production diffusion techniques which artificially preserve the fine detail compared with the way you'd record it with an AA filter or on-lens diffusion, the net result is my preferred rendition of the skin of pretty girls. (Which is 98% of my professional photography).

The sum of errors and imperfections can often get you to a more interesting creative place than a technically accurate rendition.

More than anything it is the way the camera renders skin texture that makes me choose it over my Canon when I have enough light to work with. The Canon gives nice enough results, just not the "fairy dust turned up to 11" look that I particularly like.

By contrast, to get the look I like best from motion photography I tend to use diffusion and smoke and blurry old stills lenses when shoot motion- I think the Hasselblad "sparkle" would probably look ghastly with shifting aliasing in motion. It just looks stunning in a single frame

The only answer is, as always, use the tool that gets you the result you want. And hire before you buy!

  Cheers, Hywel

This is what I mean: there are intricate differences of how an image comes out, sometimes intended, sometimes a side effect.

It is only reasonable that any chain of processing elements not only create different looking results - it is also very hard to actually match the results coming from different camera systems (as mentioned by the poster whom I quoted just above this post).

So, difference of looks from different sensors (and the systems they are a part of) are a given.

In regards to CCD vs CMOS, one probably cannot have all the advantages without any downsides: the high ISO capability of the CMOS seems to come at the cost of crispness and texture and quality of light.

[The View]

I don't think it's always cmos vs. ccd, I just think the makers get caught up in megapixels and ultra high iso and something changes in the look and the makers are no longer trying to emulate film, they're trying to make smooth, broad lattitude images.

I personally think a ccd file works deeper, more film like, more texture, but I also use ccd cameras at higher isos, never under 200 iso, usually at 320 to 640/800.

IMO

BC

I'm surprised that the Leica S2 with its CCD sensor is also considered plasticky (I have never used one, but I noticed that I was disappointed when I clicked from the Leica page to "images shot with the S2" - none of them had a look that couldn't be achieved with a 5D III).

In regards to older with newer cameras: I consider the 5D III files far superior to the 5D II files.

The fact that camera manufacturers try hard to develop cameras that sport great megapixels or high iso values and test well in pixelpeepr labs as dxomark is defenitely a horrible childhood desease of digital photography.

I hope that the industry soon goes beyond the teenage bragging and develops cameras in cooperation with artists or have people who don't design by technical charts only. We're living in a time where marketing people have much power due to oversupply of most goods - so the marketing people seems to have an unhealthy influence on camera development.