The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

If CCD rendering can be achieved with CMOS where are the examples?

ErikKaffehr

Well-known member
Hi,

Several good points.

Best regards
Erik

I actually find this a really interesting question. I am wondering with the seemingly endless question for DR, for example, we as photographers are not considering the significance on the aesthetics. If you had an image with a DR of infinity, it would look horrible. I am not saying a limit to camera DR is needed, it is not so simple, but rather as we process those images, how do we maintain a "natural" look? It seems the conversation leads to that point that one camera look "digital" (unnatural) and one looks "film-like" (natural). Ironically, film was not that "natural." I also have a feeling that the folks see this are not all seeing the same thing, which makes the "look" hard to narrow down.

BTW, my favorite camera sports a CCD. But I think that is more to do with the brilliance of Kodak and Pentax.
 

ErikKaffehr

Well-known member
Hi Ray,

Elaborate please, my understanding is that noise is dominated by photon statistics (shot noise) mostly, except in darkest areas.

Best regards
Erik


No-one has mentioned this, but the signal to noise curve output by the hardware is also a huge factor in the "look".

Ray
 

tjv

Active member
I haven't read the entire thread, but my opinion is that the noise pattern of CCD is more random and often results in a subjectively nicer "look" in that it presents as a kind of grain structure, even at base ISO. Sometimes, like with film grain, this can result in a bit more of a feel of organic structure to otherwise flat areas of tone.
Anyone else agree?
 

ErikKaffehr

Well-known member
Hi,

Midtone noise is essentially shot noise, that is photon statistics, so that would be identical for CMOS and CCD. Modern CMOS (read Sony) has less readout noise than CCD. Readout noise is quite ugly usually, more like salt and pepper.

The two samples below were taken with a P45+ and Sony Alpha 99, the closes I get to a CMOS/CCD comparison with the stuff I have. They were exposed almost identically (ETTR) according to RawDigger, but the shadows are brighter on the Sony, which I attribute to more veiling flare in the more complex Sony lens. Both images were pushed +4EV in raw conversion.

Phase One P45+


Sony Alpha 99


Best regards
Erik




I haven't read the entire thread, but my opinion is that the noise pattern of CCD is more random and often results in a subjectively nicer "look" in that it presents as a kind of grain structure, even at base ISO. Sometimes, like with film grain, this can result in a bit more of a feel of organic structure to otherwise flat areas of tone.
Anyone else agree?
 

glenerrolrd

Workshop Member
Sure there are a lot of factors that effect the resulting raw file regardless of the sensor (CCD verse CMOS) . in general CMOS files tend to be flatter and have greater DR . Each manufacture creates there own proprietary look thru the firmware used to create the raw file from the sensor capture . Nikon for example opens the shadows and compresses the mid tones (generalization ) this creates superior high ISO performance . Sony does the opposite which creates the high contrast deep saturation at the expense of higher ISO .

Its easiest to see the differences in the comparison of the M9 and M 240 ..because you have the same lens and the same manufacture . The look is created by higher contrast and compressed lights and highlights . This is where the brilliance comes from .

Color is also not the same right from the sensor …the CMOS files tend toward yellow green and the CCD files tend toward magenta blue . Peach verse Pink skin tones . These get corrected thru the camera profiles (look at a before and after file rendering ) or thru an embedded profile created by the camera firmware . This is of course greatly affected by the cameras white balance settings …so if you are looking at this you need to either do a custom white balance or establish a standard like 5000K (which may not be consistent ).

Now consider that this is all holding the light constant . If you enjoy the aesthetic of the files right out of the camera ( like a film type you enjoyed) …then its much easier to avoid all the calibration and profiling required to achieve your desired aesthetic .

Personally I very much needed the increased ISO performance (need ISO1600) for street and travel work ….so as much as I preferred the M9 files ..I moved to the M 240 . Its more work and aesthetic will never be exact …so I will build over time a new aesthetic that starts with a better file .

For slower work with MF ..I stuck with the S 006 /S2 …for me ..MF is about maximum IQ and frequently tripod work . The S2 files have been tuned to an aesthetic I enjoy and I can get it out of LR with a straightforward profile .
 

ErikKaffehr

Well-known member
This thread on LuLa discusses a comparison made by Doug Peterson just when the IQ-250 arrived. Doug compared IQ-250 (CMOS) with IQ-260 and IQ-280 (CCD). The best comparison between CMOS and CCD I have seen so far.

This image tells the story on shadow noise (IQ-260 to the left)


Best regards
Erik
 

ErikKaffehr

Well-known member
Hi,

All indications I have ever seen is that raw files, out of camera, are quite linear. So camera firmware doesn't do the kind of manipulation you suggest.

Out of camera JPEGs are of course a different thing.

Best regards
Erik


Nikon for example opens the shadows and compresses the mid tones (generalization ) this creates superior high ISO performance . Sony does the opposite which creates the high contrast deep saturation at the expense of higher ISO .
 

fotografz

Well-known member
Science does not want anything in particular, messy, tidy, or otherwise. The CCDS/CMOS problem could simple be confirmation bias, and you can test for that. If there is a "look," that can be tested as well. If the science can actually create the machines and get them to work it really tight tolerances, it can be used to analyse the same machines. Naturally, there are a whole bunch of variables that would be more a factor in this question than just the sensor architecture.

Now, I can imagine that someone could have a preference for a camera and the look from that camera or even cameras, and there may appear to be a correlation to sensor type. But I have yet to find any real analysis to find out why or to even define the qualities associated with it--and they can be quantified.

I actually find this a really interesting question. I am wondering with the seemingly endless question for DR, for example, we as photographers are not considering the significance on the aesthetics. If you had an image with a DR of infinity, it would look horrible. I am not saying a limit to camera DR is needed, it is not so simple, but rather as we process those images, how do we maintain a "natural" look? It seems the conversation leads to that point that one camera look "digital" (unnatural) and one looks "film-like" (natural). Ironically, film was not that "natural." I also have a feeling that the folks see this are not all seeing the same thing, which makes the "look" hard to narrow down.

BTW, my favorite camera sports a CCD. But I think that is more to do with the brilliance of Kodak and Pentax.
Thanks for this. I can't get my "Like" button to work, but I do find this thinking to reveal many excellent points to ponder.

Your notion that we look for some improved characteristic in sensor/architecture response, but when we get it are possibly faced with changes in other characteristics that alters our perception of the results … negative biased or otherwise (Roger's M9 vs M240 experiences).

This isn't limited to perceptions of CCD vs CMOS … it just as easily can be found in CMOS vs CMOS, (FF 24 meg Sony chip used in the Nikon D3X vs Sony A900 being an oft used example).

This "show me" type thread has come to the forefront here because of the move to CMOS in MFD. The CMOS train in smaller formats left the station years ago.

The promise of CMOS in MFD is expanded functionality not easily implemented with CCD, if at all. However, if one has been working with CCD MFD for any length of time, the understanding of the results is already there and expectations are formed by a longer experiential base and evolution of CCD over some years … often in very defined applicational circumstances.

It is exceedingly difficult to not be influenced by initial response of anything new that deviates from a longer term experience from something else one has come to "subjectively" prefer. After all, continuity of a photographer's aesthetic style, or whatever we call it, can be in part due to the preferred choice of tools from the camera to optics and how they approach PP (or how skilled they may be or want to be in PP).

That folks "aren't seeing the same thing" could be THE core issue. It suggests we all look for something different even if it is hard to define … and when found to be intuitively expressive of one's subjective aesthetic tastes then becomes the standard one measures everything else against.

Frankly, who rationally wouldn't want the added functionality that CMOS provides? The initial question then becomes "at what aesthetic expense"? Even IF it could be analyzed and quantified (which I do not doubt), is it worth the effort? Is it essential in the big picture because (IMO) it's inevitable that CMOS is the future of MFD if it is to survive.

In the interim, we have choice … it is one I recently struggled with when Leica announced the CMOS S(007) … and I chose to secure the CCD S(006) to replace my S2P. I'll wait and see how others fare with the new camera over the coming years. Meanwhile I've preserved my continuity without all the angst associated with yet another variable and associated learning curve … which is just another interruption in the day to day task of making images and securing enough work to pay for all this stuff:)

- Marc
 

yaya

Active member
One area mentioned by Doug and that has been discussed before is related to the sensor's design, the shape of the pixels and the gaps between them.

Dalsa CCD's are known for having relatively small gaps between the pixels, smaller than in Kodak CCDs and smaller than in CMOS sensors. In certain scenes and on certain subjects this translates to higher levels of sharpness and micro-contrast and also to smoother colour gradations (the latter is also affected by DR, A/D conversion and RAW processing pipeline).

Over the years this has always been one of the strong selling points for Leaf backs when compared to Kodak-chipped backs, especially when using the backs on technical or view cameras with movements as the smaller gaps and shallower pixels lead to smaller amounts of sharpness and luminance falloff and overall "rounder" images.

Also, until the D800E came out, the majority of CMOS chipped cameras employed AA filters, "contributing" to softer looking images that had to be sharpened more and creating that plasticky look.

BR

Yair
 

torger

Active member
There's really only one CMOS sensor in MF, it's the first shot. I'd expect that it will improve from here.

I think the CCD vs CMOS is more about MF companies vs 135 companies in terms of how they prefer to design look rather than technology. There's lot of hand-tuning of the color profiles. Even if a sensor is technically worse at separating colors a hand-tuned color profile with deliberate increase of separation in some color range can make it appear to be better than the other.

When MF companies get used to working with the CMOS technology and any differences there might be they will be able to produce their look. Some think they've already succeeded.

I don't feel I'm capabable of judging that sort of thing as I've never been able to sort out which is which. I like the results of the CCDs but I also like the results of the CMOS.

I think CCD is dead also in MF but it doesn't really know it yet. In three MF generations from now I don't think there will be any CCD backs left, except possibly for some narrow speciality system. So if there is a difference lets hope they sort it out before then.

With CMOS I think the future of MF looks bright now, and Pentax is part of it. CMOS provides the feature set that makes it attractive to a wider audience, with that you can succed with a volume product which I think Pentax 645Z is becoming. With volume there will be some business for Sony and others to actually make CMOS sensors of this size and the format will not need to lag behind in technology.
 

MGrayson

Subscriber and Workshop Member
FWIW, I find the "look" of the Canon 1D (CCD) closer to the Canon 1DsII (CMOS) than either is to the Oly EM-5. I'll go further and put down my totally subjective and unscientific CCD/CMOS look scale.

--IQ,M9-----1D-1DsII-----------EM-5----------------------------- iPhone--

;),

Matt

PS. I think this is just how much I like them.... (No, I don't prefer the iPhone to the IQ backs...)

PPS. I know a professional portrait photographer who happily switches between an S2 and an M(240) without ever noticing or mentioning a difference in the look. They're just tools and he uses them. The look of his work is much more distinctive than any of the differences above.
 

torger

Active member
One area mentioned by Doug and that has been discussed before is related to the sensor's design, the shape of the pixels and the gaps between them.

Dalsa CCD's are known for having relatively small gaps between the pixels, smaller than in Kodak CCDs and smaller than in CMOS sensors. In certain scenes and on certain subjects this translates to higher levels of sharpness and micro-contrast and also to smoother colour gradations (the latter is also affected by DR, A/D conversion and RAW processing pipeline).

Over the years this has always been one of the strong selling points for Leaf backs when compared to Kodak-chipped backs, especially when using the backs on technical or view cameras with movements as the smaller gaps and shallower pixels lead to smaller amounts of sharpness and luminance falloff and overall "rounder" images.

Also, until the D800E came out, the majority of CMOS chipped cameras employed AA filters, "contributing" to softer looking images that had to be sharpened more and creating that plasticky look.

BR

Yair
Interesting to hear from the "source"!

I was thinking that gap-less microlenses is "old" in CMOS. Even with a smaller photo diode what counts should be what the microlens manages to collect?

I was actually thinking it would be the other way around that CCDs usually had larger gaps, and that would separate adjacent pixels more as they sample a smaller point, and thus increase micro contrast (at the cost of aliasing). The Kodaks (with a few exceptions) don't have microlenses at all, and also have light shields as they happen to care about suppressing crosstalk and thus they do get a bit small window into the photo diode.

In any case it seems like we're talking about pixel-peeping now, differences only visible at peeping level should not be meaningful to many, right? Just kidding, actually I think people choose based on pixel-peeping a lot more than they dare to say :)

I also was of the impression that Kodak sensors have lower stack height than the Dalsa? At least if looking at the 6um technology, but perhaps it was different with the 7.2um Dalsa vs 6.8um Kodak technology (ie Aptus 75 vs P45). With Dalsa 6um having microlenses on top and Kodak 6um none it will surely be lower with Kodak? The Kodak does have pixel vignetting in tech cams, but I rather have pixel vignetting than crosstalk, microlens ripple and tiling like with the Dalsa. That's why I'm more and more looking at Hasselblad with a 50 megapixel Kodak chip as my next back to my tech cam.
 

ErikKaffehr

Well-known member
Hi,

What Anders Torger writes makes some sense. Point sampling will increase aliasing while area sampling reduces it.

This was very obvious in Doug Petersons library shots, where the IQ-250 had much less aliasing than the IQ-260 and the IQ-280 shot at the same time.

Best regards
Erik


Interesting to hear from the "source"!

I was thinking that gap-less microlenses is "old" in CMOS. Even with a smaller photo diode what counts should be what the microlens manages to collect?

I was actually thinking it would be the other way around that CCDs usually had larger gaps, and that would separate adjacent pixels more as they sample a smaller point, and thus increase micro contrast (at the cost of aliasing). The Kodaks (with a few exceptions) don't have microlenses at all, and also have light shields as they happen to care about suppressing crosstalk and thus they do get a bit small window into the photo diode.

In any case it seems like we're talking about pixel-peeping now, differences only visible at peeping level should not be meaningful to many, right? Just kidding, actually I think people choose based on pixel-peeping a lot more than they dare to say :)

I also was of the impression that Kodak sensors have lower stack height than the Dalsa? At least if looking at the 6um technology, but perhaps it was different with the 7.2um Dalsa vs 6.8um Kodak technology (ie Aptus 75 vs P45). With Dalsa 6um having microlenses on top and Kodak 6um none it will surely be lower with Kodak? The Kodak does have pixel vignetting in tech cams, but I rather have pixel vignetting than crosstalk, microlens ripple and tiling like with the Dalsa. That's why I'm more and more looking at Hasselblad with a 50 megapixel Kodak chip as my next back to my tech cam.
 

yaya

Active member
Torger, I can only suggest that before you make the move that you try your Aptus 75 alongside the 50MP as you might opt to save a packet and stay with the Aptus...just forget the specsheets for a moment and take some real world pictures.

Two things to remember are that Kodak 6nm sensors offer dual readouts (potential centrrfolding and that some of them (40MP) also have "trafitional" microlenses making them less suitable for movements with LF lenses.

And in my experience with 9nm and 7.2/6.8nm sensors, the Dalsas always performed better in terms of colour uniformity.

Yair
 

archivue

Active member
i prefer the colors from my Aptus 22 over my DM33 (still quite similar)
i prefer the colors from my DM33 over a P45+ (kodak)
i've just test a credo 60, and sometimes it's even better than my Aptus 22... it depends on light... Flare, lenses, exposure time...
From a recent trip in Japan, i was surprise to see how good was the D800 in some pics... in some pics only...
So, an example of CCD vs CMOS won't tell you all the story !
You can obtain perfect images from a P45+, but it requires more work than a Dalsa's sensor !

Anyway, i'm dreaming of a real live view with my RM3D... with long exposure capabilities as good as P45+, Colors of the P25+, completely usable from 25iso to 1600, friendly with technical camera... and all if that for cheap... ;-)
And on the top, if we can have the support of really nice people as Yaya, Doug, Steve... it will be a real plus ;-)
 

torger

Active member
Torger, I can only suggest that before you make the move that you try your Aptus 75 alongside the 50MP as you might opt to save a packet and stay with the Aptus...just forget the specsheets for a moment and take some real world pictures.

Two things to remember are that Kodak 6nm sensors offer dual readouts (potential centrrfolding and that some of them (40MP) also have "trafitional" microlenses making them less suitable for movements with LF lenses.

And in my experience with 9nm and 7.2/6.8nm sensors, the Dalsas always performed better in terms of colour uniformity.
While I do study spec sheets (which generally don't document tech wide performance that well by the way, the angular response graph if available don't say that much), I have lots of real-world shots too, with the lens line I own, with Leaf, Phase One and Hasselblad products. As developer of raw software I've looked at quite many pictures in general. I totally agree that 6um Dalsa is the better performer in many aspects, and my own 7.2um has some pretty strong aspects too. The gap is not huge as I see it though.

The Hasselblad 50 megapixel backs don't have the centerfolds or tiling, at least not the models I have good analysis material for. Probably only one read-out channel. I don't think they ever made a microlens version of that sensor, Hasselblad don't have one. I'm not going to get a P30+ :)

I'm not the typical customer though. It's not all about maximum image quality under ideal conditions.

To me it matters to have good performance with the SK wides with large shifts. It also matters that the files have good archivability and to me that means that the files should not contain any special artifacts that require sensor-specific algorithms to clean up. Capture One is almost as good as my own Lumariver HDR at cleaning up Dalsa 6um wide angle artifacts, but I'm not sure if either of those softwares will exist in say 30 years. I'm sure that generic flat-field correction will exist in mainstream software though so I'm not worried about normal LCC. I also find it valuable to be able to use generic software with my files today and thus prefer not to be dependent on software that has special artifact removal algorithms.

That's pretty odd requirements I know, but they're important for me. They matter to me more than 0.5 stop more dynamic range or a tiny bit better tonality.

The weakest part of the Hasselblad's is however their screens and I've kind of got used to having a 100% focus check with my (nowadays) trusty Aptus 75... so it's not end of story, I might end up with that Aptus-II 10 in the end anyway... I do like the Aptus series in many ways.

And a another scenario is to stay as long as I can with the Aptus 75 I have now (which gives good results!) and hope for a new generation of CMOS backs that actually support the SK wides, and that there will be an economical option like today with the CFV-50c. Had the CFV-50c supported the SK wides I would probably have bought it within a year or so together with an SK28 to compensate the loss in sensor size (my widest today is SK35).

I do struggle a bit with color currently. I don't want to use the default looks provided with the default converter. I like following and controlling the process according to my own taste and ideals as close as I can from raw to print. I use own profiles, and I'd like to develop my technique there a lot more than I have today. It's is an interesting endeavour, and I think with better technique in that area the choice of sensor will matter less in terms of color.
 

Wayne Fox

Workshop Member
I am wondering with the seemingly endless question for DR, for example, we as photographers are not considering the significance on the aesthetics. If you had an image with a DR of infinity, it would look horrible. I am not saying a limit to camera DR is needed, it is not so simple, but rather as we process those images, how do we maintain a "natural" look? It seems the conversation leads to that point that one camera look "digital" (unnatural) and one looks "film-like" (natural). Ironically, film was not that "natural."
agreed, I don’t see film as “natural” - and I don’t mean that in a negative way. remarkable engineering, each film has/had a different look based on the engineering priorities, and the reason for so many types of film was to offer different renderings because there was no ability to do it after capture. Personally I always thought Astia was fuji’s most “natural” film type yet it’s gone, but we still have the over saturated colors of Velvia.

As far as dynamic range, a real challenge. Because we look at a scene much differently than we look at a photograph of the scene, dynamic range becomes an issue. As we look at the original scene we “build” the scene by moving our eyes quickly over it, with our eyes adapting to the luminance level of the area of the scene we are looking at. Shadows rarely go totally black in the scene to our eyes yet frequently do so when captured. So how do we achieve this balance when creating an image?

Personally I think it’s easier using a high dynamic range sensor (ala a phase back or a nikon d800) and then pulling the shadows up and highlights down. I think LR does a pretty fair job of pulling up the shadows yet keeping the colors a little muted, similar to what we would see. HDR software seems to miss this point and personally I think LR is a little better at it than C1 (although 8 is better than 7).

As far as the “look” of CCD, I better not go there, because I’m pretty happy with what I get out of my d800, a7r and my IQ180. For me the choice is about the glass/focal lengths/resolution I’m after. Maybe it’s because I do only landscapes but I don’t see that much difference.
 

synn

New member
I thought my old D70s had horrible DR and was craving for more headroom back in the day.

That is, until I re-acquainted myself with film and started shooting Velvia 50. Shooting slide makes one think smarter when shooting digital, IMO. I now see the DR of my D800 and Credo as a bonus, not a necessity.
 

GregMO

Member
That is, until I re-acquainted myself with film and started shooting Velvia 50. Shooting slide makes one think smarter when shooting digital, IMO. I now see the DR of my D800 and Credo as a bonus, not a necessity.
Everything is just a rendering more or less & comes down to our own personal subjective taste to determine what we each use.

With Velvia, just being off by 1/3-1/2 stop makes a big difference. For scenes that require more range, I find Ektar 100 can handle everything I photograph with a single shot. Portra has more DR then I need.
 

Paratom

Well-known member
I think it doesnt mind.
One has to shoot a camera and look at the results, colors, tonality etc. and decide if one likes it or if one doesnt like it.
Personally I couldnt make up my mind between the M9 and the M240 color and have the luxury to be able to keep both. The M9 sometimes shines but sometimes it seems really of. I am leaning toward the M because it seems more consistent.
I find the Oly EM1 to deliver more consistent and better color than bothe the M9 and M, and also better than the Sony A7 series...even though the Oly is CMOS.
I like the color from the S a lot but it doesnt look allways neutral.

So no, I cant see the difference between ccd and cmos, I have my preference which cams output I do like though.
However it seems that there are few if any ccd cameras where I dont like the output....mmhhh?
 
Top