Thanks for taking the time and effort to go through this. I have a D600 and two D200 bodies, so if I weren't so lazy I would've tried this. My comparisons so far have been much less empirical. I have noticed the differences are (perhaps obviously) affected by the lens, so one lens looks "better" on the D600, while others are the opposite. This probably has something to do with what you found with the blue rendering. Describing the difference in words is indeed difficult. You used the word "texture" at one point, which I think is a very good word for it. Also, "color depth" and "color richness" seems to come to mind for some images. Yet that's not right either, because my D600 images are plenty deep, rich, and textured. But in a different, more clinical (sterile?) way. I'd use my D600 for product images, photos for eBay listings, etc. The D200 is for what I call "real life" and "art".
This is the best comparison of CCD vs CMOS sensor output that I have seen so far. It would also be worthwhile looking at early CMOS sensor output to understand when things changed in terms of colours and texture. The images from the D3 and D700 look a lot closer to their CCD predecessors than to later CMOS models.
I did consider that the older ones might render closer but. The cropping required to do the comparison was a bit much. Hmm if I bought a d100 to use as the CCD camera I think I would be better but even then there is some finagling.
@@sameerrao5834 I did consider doing something similar (don't currently own a d300) but the pixel pitch of the d200 and d300 are different and I didn't want to deal with scaling images. That said I am considering looking for a D2x (just because I want one more than the d300 haha) so if I get that I expect to revisit this. I may also repeat this using a canon camera of the same era and the D200.
Great comparison. I’ve been using the D100 since April, the colors from it are quite pleasing compared to the CMOS sensor cameras I’ve used. I recently purchased a D40X (10mp CCD) and the images are technically better but the D100 has more of a filmic look to it. The only other cameras that produce similar colors are the D4, DF and maybe the D700 a bit.
There are cameras that have "something" in it, don't matter the sensor tech too much. I mean, the D700, or Canon 5D, even older Fuji mirrorless, have a distinct rendering being CMOS. Maybe it could be because the IR filters in front of the sensor or the not so perfect sensor data processing, or maybe it was intentional trying to get a good pleasing image and today cameras want a technically perfect image. Some people say that CCD gives more blooming to the highlights and color due the adjacent pixel contamination, and it could be true. But for me sensor size, lens, and overall rendering are far more important than sensor tech in the results.
The D200 is an incredible good camera, and it is cheap to get. I think they sold insane volumes of this model back in the days.. It is easy to get on the used market, which is nice..
The warmth and saturation is obviously more pronounced in the CCD samples. I think this is the case, regardless of camera. I was skipping through some old photos and came across some and thought ah, these must be the D700, did a quick check of the exif and oops - Pentax K10d. It is what it is, and it's fine by me👍
I'm interested in this debate and I use both of these cameras. To me, once you adjusted the colour on the D600 output, there's very little difference between the sample pictures. I personally think this debate is based on a faulty correlation. I think there's nostalgia about the look of early digital among younger people whose childhood pictures were taken like that. And the, frankly, poor quality of the pre 2005 cameras is being enjoyed for its particular aesthetic and mistaken for a 'filmic' quality. It is not. And I know that because I kept shooting film during that phase _precisely_ because of how horribly unlike film those early cameras were - terrible dynamic range with bone white highlights (compared to negative film's wide latitude) and awful muddy skin tones (compared to the beautiful gradations of film). In 2005, cameras like the D200 demonstrated that digital could look really, really good. And cameras since then, most of them CMOS, have continued to do so - with wide variations in the colour balances that manufacturers have favoured (it's all fashion, folks). The most 'filmic' looking camera I've used (and I still use it regularly) is the Fujifilm XE1 with the Xtrans1 sensor (the Xpro1 uses it and it's famous for this film-like rendering). And it's CMOS. But pretty much all post-2005 cameras can be tweaked to give beautiful results - some more easily than others such as the D600 and the Canon 5D Classic just as examples. I think the 'clinical' look people are responding to atm is about colour depth and the sharpness and contrast of very precise modern lenses - combined with certain fashions in rendering and processing. I don't think CCD sensors in themselves have any special aesthetic characteristics. Rather it's the era in which they were common that had its set of technical challenges and aesthetic choices.
I think the clinical look is due to several factors combined. One major one in my opinion is digital sharpening. There was no sharpening with film it was a trait of the lens only and maybe film grain size. I prefer an optical sharp lens and as little digital sharpening as possible. Digital sharpening seems unnatural to me and I think this is what people really mean when they say clinical. With digital cameras you don't know what digital processing may be happening in the sensor or processor without your awareness as default by the manufacture. To understand why digital does not look like film you have to ask yourself what processes are happening now that were not in play with film. Digital sensors require at minimum some sharpening due to sensor construction then there is the bayer filter or x-trans processing as well. also I believe film is a color subtractive Chemical process while digital is colour additive. It's more profitable to sacrifice lens optical qualities now and use digital processing to compensate.
@@dwaynepiper3261 What a pleasure to read your comment because I have been thinking that for a while but I am not an expert in English and it is difficult for me to express it. I think it is very true that the first CCD and early CMS professional cameras wanted to look like film, it is logical since those users had to be justified in making the change. I remember that time and we went from seeing weird toys to digital cameras that looked great and then turned out to perform better. So it was a matter of changing the bodies and taking advantage of the optics inherited from film, at least in Nikon and Canon Olympus went a different route. Then the world of photography was shaken by the cell phone revolution. which was not just a digital camera but a camera that made it "prettier" without any editing effort. Cell phones and social networks invaded photography. And for me they imposed a way of seeing. If the first digital cameras wanted to look like film, the latest digital cameras want to do what cell phones do but better. I mean, RAW files that are already super cooked by the software. I think that in optics almost everything is invented, but not in software. Or at least, it is cheaper to make a mediocre optic and correct it in software than to make a decent optic that did not need as much software adjustment. I don't know if you saw the video and user comments in Nigel Danson's comparison between a D200 and a Z7 in landscape photography. I don't know if you saw the video and user comments in Nigel Danson's comparison between a D200 and a Z7 in landscape photography I love reading user feedback and I guess not everyone has the D200. I also think there is a generational issue and over time perhaps the influence of film on the look of digital photography will not even be a preset.
@@dwaynepiper3261 That may be so but it doesn't really affect the question of CCD vs CMOS in terms of 'filmic' looks - since both are digital and both were sharpened both in camera and typically in post.
The gradient texture thing you mention, I noticed it too when using CCD. It is almost like it dithers color transitions, rather than making them smooth as in CMOS. Yes, it is a kind of random "grain" (as distinct from simple ugly banding if it were only a matter of color depth) and I find it much more pleasing to the eye than "plastic" smoothness of CMOS. Older CCDs had this effect especially noticeable. I have Canon G10 and G12, and in the older G10 this is very pronounced compared to G12 which feels a bit smoother. And yes, I think we can call it "grain" - I've started photography back with analogue and this effect is very similar to what you get on film. Also, I've noticed that CCDs even when at higher ISO still preserve detail, contrasty bits... With CMOS when you get noise, then everything is noisy, and with CCDs it is the gradients and uniform areas which are more noisy ("textured" I'd say) than detail. Coming from film originally, I also find this more pleasing to the eye than CMOS where everything just looks as if someone placed a uniform "noise" filter over everything.
Often the images from a camera with a CCD sensor look more "alive" from every one of them even the older pocket sized snapshot models from different brands. Why this is i don't know but it has something to do with the fact that they "are" ccds instead of cmos sensors. I clearly remember the very old Canon A1 pocket camera with around 1.2 mpixels something but the photos from that camera looked absolutely stunning and very much so alive despite the low pixel count. So it's not always the amount of pixels that determines the quality of the pictures you take also. You can also compare f.ex different models from Nikon let say the in comparison inexpensive d800 and other semi pro models to the D4 model in the pro series from that manufacturer. Here you can also very clearly see a big difference in the "aliveness" in the photos coming out of the D4 body in comparison to the D800 despite they both have cmos sensors. And this goes generally for other models from that lower price range too. And this is also true between pricier and lower series from Canon f.ex. So it seems that it is "possible" to make any kind of impression from the photos from most cameras if you really WANT to not only "blaming" the sensor for being a bad or better one especially between the cmos sensors due to different programming of the software itself. But i have always regardless of camera brand and model experienced that the images from a CCD sensor always looked "better" and much more alive in comparison. How this can be i cannot explain BUT there clearly seem to be a difference between these two technologies, the cmos ones always looking more flat and lifeless in comparison. So clearly the color science also plays a BIG role in how the photos are perceived coming out or being rendered by the software and programming from that specific model. AND you can also see the same difference between different models in that Dx.. Pro range series from f.ex Nikon. Not all look "alive" in the same way despite they all have cmos sensors and all coming from that pro series of cameras. So again, it seems that it is possible to make different outcomes from most cameras in the way the rendering engine is programmed AFTER the sensor itself especially with the cmos models. But again as we could see between the D4 and D800 from Nikon it IS possible to make images coming out from a cmos sensor more "alive" if you want to, but in my opinion NOT as much as from a ccd sensor. How this can be maybe the engineers at the various camera manufacturers can explain, because clearly there is a difference between these two sensortypes despite the more "expensive" programming of the software between some models with the cmos sensors inside of them in the more expensive range of cameras !
I definitely prefer the CCD. I have both a Pentax 645D ( a 40MP CCD) and a Pentax 645Z (a 50MP CMOS) and I can see a difference, and prefer the CCD. CCDs seem a bit more like film, indeed, the way that they work is more like film. CCDs are exposed to the whole image simultaneously, whereas CMOS is scanned in a raster fashion. I don't know why this would make a difference, but it does.
645d looks absolutely better than the z. the 3d depth is the single most impotant element the cmos are not there. it misses the point to compare the two sensors with a flower closeup.
The 645d is amazing. I had a gfx 50s which I think have the same sensor as the Z. Sold it and got the D instead. I also have the E1, E300, E500, E400 from Olympus, and a couple of M8. You can never get too many Kodak ccd's😊
I had a Nikon D200; the closest full-frame Nikon camera in color rendition is the Nikon D810. P.S. The aperture value on a full-frame camera should be multiplied by 1.5 to match APS-C. So F4 on APS-C is equivalent to about F6 on full frame (to get a similar look).
I mean the blue just seems like a difference in WB accuracy. However in my experience with the CCD sensors is they just tend to have a smoother gradient from black to white. You also have to remember we’re not seeing a completely raw image. The camera manufacturers runs it through their processor and interprets the colors as they see fit. Nikon probably didn’t change their color science that much between these two cameras. I’m sure if you had the same sensor in a Sony A dslr or a canon dslr you’d see some differences but just because the manufacturers process the raw data differently. Also I know the d200 says “36 bit color depth” in the specs. I don’t know if that was just a marketing scheme by counting each color channel bit depth and adding them together or not. I know most cameras today say 12, 14 or 16 bit so idk if the d200 was just a 12 bit rgb and they just claimed 36 bit?
I am not sure if it would be a WB accuracy issue as both WB and tint were adjusted to match unless you mean the WB tweaks internally are different which I would interpret as differences in color science as they developed their sensor tech. You are right though when you look at 2 different cameras with the same sensor more prominent color science differences come out. This is very much the case with the Nikon D80 vs Sony A100 vs Pentax (don't remember the model number) But yeah the 36 bit think from Nikon comes from 3x 12 bit channels. I mean the math checks out haha. So not inaccurate, just the way it's normally presented. But that did bring up a point noted by someone else. The D600 does 14 bits per color output so I might need to see if I can noticed a difference when switch it into 12 bits. Either way it's a fun look at the differences. I'll probably see if I can snag a d2x down the line to compare the 2 and just work around the pixel pitch difference.
Mr 50mm l listened carefully to your explanation related to the model choices for the sensor comparison. It made sense to me as l have both Nikons. Of course it is a matter of availability too. The CCD v CMOS ‘imbroglio’ does have a dichotomy of ‘positions’ with confirmed adherents. Perhaps not unlike BETA v VHS, the former apparently had the superior quality but the later cost less. I do recall from my meanderings on this topic, that Leica persisted with the CCD sensor with one of its later models. Who knows where further development may have taken the CCD? Although it has persisted l believe in photocopy technology. In the visual image comparison you conducted the D200 didn’t perform too shabbily at all. What the histograms revealed about the colours was particularly of interest. So some evolution in the ‘colour science’ has occurred over the model generations. I do agree with the views suggesting it is not all down to a sensor ‘shoot out’, filter configuration, processors etc are all part of the bigger picture.
Yeah there is likely a lot at play rather than just what is being read out of the sensors. But indeed Lieca and Phase One had continued to use CCDs for a long time after. The Lieca M9 used a Kodak sensor and was made in 2009 to 2012. Phase One IQ 3 80 being introduced in 2016. Until the IQ 4 series which was entirely CMOS line up of digital backs.
CCD seems to have better micro contrast. CMOS looks flatter and and more processed (Digital). I think a monochrome photo comparison test would show it better if true. Much easier to see the differences in B&W. I've seen other videos where it shows more color variations in gradients. I think it just looks like more of an analog tonal curve in CCD. CMOS has come a long way and modern sensors give excellent modern looks. They have so much circuitry though in their processing (High transistor count). CCD has much less circuitry but then is speed capped to it's design. To get speed out of CCD's companies like Kodak and Sony split sensors into a multi tap architecture where its like having up to four sensors working as one. They had to be balanced together to make it look seamless. CMOS took over after that as it gave imaging technology much more potential. I wonder how far they could could have taken CCD's overall had CMOS not taken over all development.
It would have more sense to have compared the D200 (last CCD) with the D300 (first CMOS) in what is almost the same body except for the larger rear screen. Much of the CCD film look is really the processor colour science!
I don't see why it would be better to compare the 2 aside from their generational closeness. I wasn't specifically looking at the sensors in terms of technical capability. (Hence why most of the photos were shot their base ISO, in scenes where the Dynamic range wasn't pushed to the limit). Additionally the D200 and D300 resolutions could have caused more of a look difference than the near identical pixel pitch of the D200 and D600. Admittedly I actively made the choice to examine 2 sensors with the same pixel pitch rather than of similar generation because I personally though I'd have even more emphasis on the differences given most opinions are that modern CMOS look "too clinical". Finally the D200 was later in the life of nikon's CCD but it's years before they stopped using it the D3000 was Nikon's final CCD DSLR (in 2009, 3 years after the D200 release), the D300 is also depending on how you view it the second or third gen DSLR from Nikon with CMOS, the D2h (LBCAST) technology is a variant of CMOS technology, but if you want to exclude that then the D2x would be the first gen Nikon with a CMOS sensor (supposedly designed by Nikon, made by Sony). Although the D300 is the first Nikon DSLR to feature a CMOS that is wholly designed by Sony.
CIS has some huge benefits being so widely used as a sensor compared to CCDs, and becoming more and more so. CCDs have some inherent benefits but with softwares and better interpolation I think there isn't enough in using a CCD.
Ouu yes sigma cameras would be an interesting way to go. I occasionally look for them but the price of them especially the Merrill variants is a bit too high for me ATM. I may try some of the older ones though.
I had a D200. It was alright. Hated the small LCD. really didn't notice anything spectacular about the sensor. shortly after I got a D7000. That in my opinion walked all over the the D200 sensor wise. Better in every single way. Sold the D200 and Canon 40D and another canon Xsi. Never looked back.
Haha fair enough for me the ergonomics of the camera are what set that camera apart. The small screen resolution I can forgive since that was considered a fairly generous screen for the time. As far as the sensor it's just my preference but I do like it in compared to 10-16mp sensors of the era. Not sure exactly why but I do. But then again I'm not currently very picky about my cameras haha.
ccd vs cmos isnt all that different... if youre going CCD for the CCD look, its the Fujifilm Super CCD DSLR's and the Kodak CCD DSLR's, theres no comparison.
For me, Nikon model after around Expeed 3 era (after 2012) rendered picture more blue leaning and seem to have darker image at the same exposure compare to older camera (might not included some really old model like D1 series or D100 which have somewhat darker image at the same exposure as well though they're different from model with Expeed 3 and later processor), and I have to say their light meter also pick a value around 1 stop under what the older DSLRs or film cameras can metered in bright daylight and even when compensate 1 stop more exposure its picture still not look as bright as picture from film camera when exposed at the same value with the same ISO film. So I think the difference should be credited to how newer image processor and overall camera processing work rather than the different between CCD and CMOS but because newer cameras had been made in a way that they're not as sensitive as older model for some reason and there're somewhat noticable change in Nikon color science in Expeed 3 era and beyond compare to Pre-Expeed era (everything before 2007 which have rather high color intensity while not have so good dynamic range and the image didn't look as clear as newer model) and original Expeed era (D3, D300, D700, D90; which I found their color rendition brighter and clearer than the model before and after while still vibrant; Expeed 3 and later didn't render as vivid color by default anymore and by that time Canon's image look more vibrant which is opposited from before that time).
Yeah so there does seem to be a pretty big change in the way the color science is done on these cameras. Haha I might need to repeat this with a d2x down the line.
Pics from my old Canon 10D and Pentax K10D look a little more saturated than pics from my 5Dii and K20D, respectively. Only reason I noticed that was all the hoo-hah recently about the CCD nostalgia. Yes, there is a slight difference, but not one I'd go out of my way to find or buy another old camera to capture. It's mostly not worth bothering about. On balance, I'd say the lens used on the day could have more affect on image balance - I shoot a lot of old manual glass and some of the older pics might have got their very slightly more saturated look from the glass in use.
CCD is less sharp and more saturated in those images. Again you can manipulate this in post to a certain extent. If you feel the color transitions are better on the CCD, then maybe micro-contrast is better in your opinion?
The reduced sharpness might be less a consequence of the sensor directly and more related to the Anti aliasing filter on top as they were more aggressive with that. Yeah CCD micro contrast I think could be better or rather more noticeable due to the lower bit depth of the CCD to CMOS. I should have probably switched the D600 to 12 bit output. Maybe another video!
The main reason for the "why didn't I compare it to XX" cameras is don't have one, that was also the case with the D300. Additionally the D600 was an interesting candidate as the pixel pitch is almost exactly the same. I shot all the D600 photos in DX mode, which outputs a raw file at about the same resolution as the D200, this is nice for me since I didn't have to do any sort or scaling, and then I can't chalk up results of one or the other as the pixel density disparity. That said, I do plan to do something along those lines eventually.
2005 APS-C/DX 10MP CCD vs 2012 Fullframe (FX) 24 MP CMOS...seems to be unfair. Much better would have been - you took the D90 (2008, 12 MP, CMOS) as comparsion 1:1 with the D200...just saying. 🙂
Yeah, I did that mainly mainly do to the near 1 to 1 pixel scaling (D600 in DX mode is 10mp). Plus given the arguments about the look of CCDs vs CMOS extend beyond similar generation cameras and usually broadly comments go something like "modern CMOs are soo plastic (digital, whatever other comparison)" I figure it shouldn't matter what year the CMOS sensor I used would be. But otherwise I will probably revisit this in the future for sure! Haha if I can I'll see if I can get a D2x or something closer. Or maybe the Canon 40D.
@@Mr50mmish ok, sorry, i didn't notice that you've had used the DX crop mode. 40D is good, still have mine since 2008. But meanwhile, i do prefer other gear. The 18-135 STM is fine, as shoot it all zoom, btw. Still nowadays. As for the looks, i do always prefer a CCD Sensor vs. CMOS, as for the colors. Don't care about DXO Mark ratings - as long, it's pleasing for the eye. Fun fact: the Canon Sensors are (mostly) are being CMOS, but even the old ones have a more "filmic" look, compared to their current offerings. 🙂
@@marcp.1752 all good, haha yeah I just wanted to avoid scaling issues is all! But yeah I'll definitely revisit this with similar generation sensor to see if it's more a generational look! Cheers!
@@Mr50mmish if i remember correctly, back in the day cmos sensors were front-side illuminated, while ccd sensors were bsi, which made cmos sensors show a bit more of a vignetting effect since light came in slightly off-angle? also, there might have been something about higher luma noise but lower chroma noise for ccds compared to cmos sensors back in the day? maybe those two effects was what made the ccd bokeh have a more "textured" look with more saturated colours? separate question just out of my own curiosity: did you find dynamic range suffered? ie. some people find canon 5dm3 is more saturated straight out of camera RAW than an r6, cos it has less dynamic range. is that the case for ccds vs cmos?
@@vtct_ytaa that could be it. Giving the noise profile it's look. When it came to overall saturation I actually didn't find the D600 less saturation once I matched the blue channel to the D200. Despite the D600's higher DR. Maybe the R6 flattens the image more to keep images from appearing blown. I do notice that the D600 will just recover more than the D200 as a result. Haha I suppose they might have actively made that choice to keep the images a bit more pleasant!
@@Mr50mmish ahh gotcha, very good video mate. keep up the good work and i look forward to watching more videos from you! Cheers for all the testing you do!
They look fairly similar to me, I guess I should have tried a test chart or something for the video. Although I'm sure somewhere there has to have been a review where someone calculated the effective max resolving power of each sensor!
I use the 2 due to their very similar pixel pitch. Additionally the d600 was shot in DX mode and the image output was effectively the same as the d200 as a result.
The CCD version has somewhat nicer colors, the cmos seems a bit cooler, technical. But these are ancient cameras the CCD version being just different, not better. The cmos has far better noise performance. CCD is anything but film, it has ugly color noise. And is just as harsh as any old digital, modern cameras destroy these results. I think it`s wishfull thinking and pseudo science combined, if you want it to be better you`re definetly going to see it. Besides, only 2 demo photos, that`s pathetic.
I did say off the bat it's not a technical examination of the 2 sensors, there are plenty of websites that will quantify how much better modern CMOS to CCD is. I am aware that CMOS sensors have progressed a lot and of course modern sensor noise performance will be much better than old CCDs. As for lack of photos given that I only color saw differences in the blues across a few photo for my photos I didn't think I'd need display that multiple times to show it. If you want a technical comparison you are free to play with all the charts you want on DXO.
@@Mr50mmish Man, there really was much more going on with the colors than only the blues. And btw. 2 pics is still a lousy 2 pics. Far to little in any comparison. And how come you say there still some magical CCD anologueness going on when all you now say there is only difference in blues. Seems like a lot of bull to me. Oh and i hate DXO wit a passion.
Photography is art first and science second, at least for me. And I prefer CCD, the older the better. I find talking about "science" and "pseudo-science" in context of art utterly ridiculous. But that's just me. I'm simply not into forensic photography. And besides, it seems obvious that the conclusion here that the difference is not in the way it handles color alone, but the way its noise is structured and textured, particularly when it comes to uniform vs detailed areas. I am not aware of any "scientific" metric devised for that so far.
Thanks for taking the time and effort to go through this. I have a D600 and two D200 bodies, so if I weren't so lazy I would've tried this. My comparisons so far have been much less empirical. I have noticed the differences are (perhaps obviously) affected by the lens, so one lens looks "better" on the D600, while others are the opposite. This probably has something to do with what you found with the blue rendering. Describing the difference in words is indeed difficult. You used the word "texture" at one point, which I think is a very good word for it. Also, "color depth" and "color richness" seems to come to mind for some images. Yet that's not right either, because my D600 images are plenty deep, rich, and textured. But in a different, more clinical (sterile?) way. I'd use my D600 for product images, photos for eBay listings, etc. The D200 is for what I call "real life" and "art".
This is the best comparison of CCD vs CMOS sensor output that I have seen so far. It would also be worthwhile looking at early CMOS sensor output to understand when things changed in terms of colours and texture. The images from the D3 and D700 look a lot closer to their CCD predecessors than to later CMOS models.
I did consider that the older ones might render closer but. The cropping required to do the comparison was a bit much. Hmm if I bought a d100 to use as the CCD camera I think I would be better but even then there is some finagling.
@@Mr50mmish Perhaps a D200 vs D300? Both are crop sensor, one a CCD and the other, an early CMOS.
@@sameerrao5834 I did consider doing something similar (don't currently own a d300) but the pixel pitch of the d200 and d300 are different and I didn't want to deal with scaling images. That said I am considering looking for a D2x (just because I want one more than the d300 haha) so if I get that I expect to revisit this. I may also repeat this using a canon camera of the same era and the D200.
Great comparison. I’ve been using the D100 since April, the colors from it are quite pleasing compared to the CMOS sensor cameras I’ve used. I recently purchased a D40X (10mp CCD) and the images are technically better but the D100 has more of a filmic look to it. The only other cameras that produce similar colors are the D4, DF and maybe the D700 a bit.
There are cameras that have "something" in it, don't matter the sensor tech too much. I mean, the D700, or Canon 5D, even older Fuji mirrorless, have a distinct rendering being CMOS. Maybe it could be because the IR filters in front of the sensor or the not so perfect sensor data processing, or maybe it was intentional trying to get a good pleasing image and today cameras want a technically perfect image. Some people say that CCD gives more blooming to the highlights and color due the adjacent pixel contamination, and it could be true. But for me sensor size, lens, and overall rendering are far more important than sensor tech in the results.
Very interesting test. The CCD pics do seem to have more character.
The D200 is an incredible good camera, and it is cheap to get.
I think they sold insane volumes of this model back in the days.. It is easy to get on the used market, which is nice..
thanks for the video. definitely useful. i have a D200 and a D300 and there is a difference.
Thanks! I was thinking of maybe redoing this but with a D300! Maybe if I come across one for a good price!
ccd looks like lil drops of color vs the cmos rendering as a whole cohesive scene, like a seurat painting vs a vermeer
cmos have a creamier look and the CCD sensors have more of an etched look.
D200's got a smidge of warmish tone which I can only describe as CCD magic. Because even if you up WB a bit you won't get the same results.
Yeah I agree the WB tweaks and tint can get close but it doesnt look the same.
CCD secret sauce 🙂
The warmth and saturation is obviously more pronounced in the CCD samples. I think this is the case, regardless of camera. I was skipping through some old photos and came across some and thought ah, these must be the D700, did a quick check of the exif and oops - Pentax K10d. It is what it is, and it's fine by me👍
I'm interested in this debate and I use both of these cameras. To me, once you adjusted the colour on the D600 output, there's very little difference between the sample pictures.
I personally think this debate is based on a faulty correlation. I think there's nostalgia about the look of early digital among younger people whose childhood pictures were taken like that. And the, frankly, poor quality of the pre 2005 cameras is being enjoyed for its particular aesthetic and mistaken for a 'filmic' quality. It is not. And I know that because I kept shooting film during that phase _precisely_ because of how horribly unlike film those early cameras were - terrible dynamic range with bone white highlights (compared to negative film's wide latitude) and awful muddy skin tones (compared to the beautiful gradations of film).
In 2005, cameras like the D200 demonstrated that digital could look really, really good. And cameras since then, most of them CMOS, have continued to do so - with wide variations in the colour balances that manufacturers have favoured (it's all fashion, folks). The most 'filmic' looking camera I've used (and I still use it regularly) is the Fujifilm XE1 with the Xtrans1 sensor (the Xpro1 uses it and it's famous for this film-like rendering). And it's CMOS.
But pretty much all post-2005 cameras can be tweaked to give beautiful results - some more easily than others such as the D600 and the Canon 5D Classic just as examples. I think the 'clinical' look people are responding to atm is about colour depth and the sharpness and contrast of very precise modern lenses - combined with certain fashions in rendering and processing.
I don't think CCD sensors in themselves have any special aesthetic characteristics. Rather it's the era in which they were common that had its set of technical challenges and aesthetic choices.
I think the clinical look is due to several factors combined. One major one in my opinion is digital sharpening. There was no sharpening with film it was a trait of the lens only and maybe film grain size. I prefer an optical sharp lens and as little digital sharpening as possible. Digital sharpening seems unnatural to me and I think this is what people really mean when they say clinical. With digital cameras you don't know what digital processing may be happening in the sensor or processor without your awareness as default by the manufacture. To understand why digital does not look like film you have to ask yourself what processes are happening now that were not in play with film. Digital sensors require at minimum some sharpening due to sensor construction then there is the bayer filter or x-trans processing as well. also I believe film is a color subtractive Chemical process while digital is colour additive. It's more profitable to sacrifice lens optical qualities now and use digital processing to compensate.
@@dwaynepiper3261 very well said,
@@dwaynepiper3261
What a pleasure to read your comment because I have been thinking that for a while but I am not an expert in English and it is difficult for me to express it.
I think it is very true that the first CCD and early CMS professional cameras wanted to look like film, it is logical since those users had to be justified in making the change. I remember that time and we went from seeing weird toys to digital cameras that looked great and then turned out to perform better. So it was a matter of changing the bodies and taking advantage of the optics inherited from film, at least in Nikon and Canon Olympus went a different route. Then the world of photography was shaken by the cell phone revolution. which was not just a digital camera but a camera that made it "prettier" without any editing effort. Cell phones and social networks invaded photography. And for me they imposed a way of seeing. If the first digital cameras wanted to look like film, the latest digital cameras want to do what cell phones do but better. I mean, RAW files that are already super cooked by the software. I think that in optics almost everything is invented, but not in software. Or at least, it is cheaper to make a mediocre optic and correct it in software than to make a decent optic that did not need as much software adjustment. I don't know if you saw the video and user comments in Nigel Danson's comparison between a D200 and a Z7 in landscape photography. I don't know if you saw the video and user comments in Nigel Danson's comparison between a D200 and a Z7 in landscape photography
I love reading user feedback and I guess not everyone has the D200. I also think there is a generational issue and over time perhaps the influence of film on the look of digital photography will not even be a preset.
I have two X-E1 and I am planning to buy more of them
@@dwaynepiper3261 That may be so but it doesn't really affect the question of CCD vs CMOS in terms of 'filmic' looks - since both are digital and both were sharpened both in camera and typically in post.
The gradient texture thing you mention, I noticed it too when using CCD. It is almost like it dithers color transitions, rather than making them smooth as in CMOS. Yes, it is a kind of random "grain" (as distinct from simple ugly banding if it were only a matter of color depth) and I find it much more pleasing to the eye than "plastic" smoothness of CMOS. Older CCDs had this effect especially noticeable. I have Canon G10 and G12, and in the older G10 this is very pronounced compared to G12 which feels a bit smoother. And yes, I think we can call it "grain" - I've started photography back with analogue and this effect is very similar to what you get on film.
Also, I've noticed that CCDs even when at higher ISO still preserve detail, contrasty bits... With CMOS when you get noise, then everything is noisy, and with CCDs it is the gradients and uniform areas which are more noisy ("textured" I'd say) than detail. Coming from film originally, I also find this more pleasing to the eye than CMOS where everything just looks as if someone placed a uniform "noise" filter over everything.
Or we can call it noise structure to be more precise, because it’s not grain
CCD forever! 📸
Often the images from a camera with a CCD sensor look more "alive" from every one of them even the older pocket sized snapshot models from different brands.
Why this is i don't know but it has something to do with the fact that they "are" ccds instead of cmos sensors.
I clearly remember the very old Canon A1 pocket camera with around 1.2 mpixels something but the photos from that camera looked absolutely stunning and very much so alive despite the low pixel count. So it's not always the amount of pixels that determines the quality of the pictures you take also.
You can also compare f.ex different models from Nikon let say the in comparison inexpensive d800 and other semi pro models to the D4 model in the pro series from that manufacturer.
Here you can also very clearly see a big difference in the "aliveness" in the photos coming out of the D4 body in comparison to the D800 despite they both have cmos sensors. And this goes generally for other models from that lower price range too. And this is also true between pricier and lower series from Canon f.ex.
So it seems that it is "possible" to make any kind of impression from the photos from most cameras if you really WANT to not only "blaming" the sensor for being a bad or better one especially between the cmos sensors due to different programming of the software itself.
But i have always regardless of camera brand and model experienced that the images from a CCD sensor always looked "better" and much more alive in comparison. How this can be i cannot explain BUT there clearly seem to be a difference between these two technologies, the cmos ones always looking more flat and lifeless in comparison.
So clearly the color science also plays a BIG role in how the photos are perceived coming out or being rendered by the software and programming from that specific model.
AND you can also see the same difference between different models in that Dx.. Pro range series from f.ex Nikon. Not all look "alive" in the same way despite they all have cmos sensors and all coming from that pro series of cameras.
So again, it seems that it is possible to make different outcomes from most cameras in the way the rendering engine is programmed AFTER the sensor itself especially with the cmos models.
But again as we could see between the D4 and D800 from Nikon it IS possible to make images coming out from a cmos sensor more "alive" if you want to, but in my opinion NOT as much as from a ccd sensor.
How this can be maybe the engineers at the various camera manufacturers can explain, because clearly there is a difference between these two sensortypes despite the more "expensive" programming of the software between some models with the cmos sensors inside of them in the more expensive range of cameras !
I definitely prefer the CCD. I have both a Pentax 645D ( a 40MP CCD) and a Pentax 645Z (a 50MP CMOS) and I can see a difference, and prefer the CCD. CCDs seem a bit more like film, indeed, the way that they work is more like film. CCDs are exposed to the whole image simultaneously, whereas CMOS is scanned in a raster fashion. I don't know why this would make a difference, but it does.
Yeah the global shutter on a CCD is pretty awesome but now Sony's global shutter CMOS is a thing I wonder how those will compare!
645d looks absolutely better than the z. the 3d depth is the single most impotant element the cmos are not there. it misses the point to compare the two sensors with a flower closeup.
when watching photos from ccd i am in there as if i had been on that street. with photos from cmos i am watching a very very nice photo.
@@jackielo2898 maybe I'll do a follow up later with more varied subjects.
The 645d is amazing. I had a gfx 50s which I think have the same sensor as the Z. Sold it and got the D instead. I also have the E1, E300, E500, E400 from Olympus, and a couple of M8. You can never get too many Kodak ccd's😊
I had a Nikon D200; the closest full-frame Nikon camera in color rendition is the Nikon D810.
P.S. The aperture value on a full-frame camera should be multiplied by 1.5 to match APS-C. So F4 on APS-C is equivalent to about F6 on full frame (to get a similar look).
For FOV correction that's true, but instead I just shot the d600 in DX mode. So no correction needed.
New D40 owner here..
Let me just grab a coffee and some popcorn 🙂
Haha 😆
I mean the blue just seems like a difference in WB accuracy. However in my experience with the CCD sensors is they just tend to have a smoother gradient from black to white. You also have to remember we’re not seeing a completely raw image. The camera manufacturers runs it through their processor and interprets the colors as they see fit. Nikon probably didn’t change their color science that much between these two cameras. I’m sure if you had the same sensor in a Sony A dslr or a canon dslr you’d see some differences but just because the manufacturers process the raw data differently. Also I know the d200 says “36 bit color depth” in the specs. I don’t know if that was just a marketing scheme by counting each color channel bit depth and adding them together or not. I know most cameras today say 12, 14 or 16 bit so idk if the d200 was just a 12 bit rgb and they just claimed 36 bit?
I am not sure if it would be a WB accuracy issue as both WB and tint were adjusted to match unless you mean the WB tweaks internally are different which I would interpret as differences in color science as they developed their sensor tech. You are right though when you look at 2 different cameras with the same sensor more prominent color science differences come out. This is very much the case with the Nikon D80 vs Sony A100 vs Pentax (don't remember the model number)
But yeah the 36 bit think from Nikon comes from 3x 12 bit channels. I mean the math checks out haha. So not inaccurate, just the way it's normally presented. But that did bring up a point noted by someone else. The D600 does 14 bits per color output so I might need to see if I can noticed a difference when switch it into 12 bits.
Either way it's a fun look at the differences. I'll probably see if I can snag a d2x down the line to compare the 2 and just work around the pixel pitch difference.
FINALLY
Mr 50mm l listened carefully to your explanation related to the model choices for the sensor comparison. It made sense to me as l have both Nikons. Of course it is a matter of availability too. The CCD v CMOS ‘imbroglio’ does have a dichotomy of ‘positions’ with confirmed adherents. Perhaps not unlike BETA v VHS, the former apparently had the superior quality but the later cost less. I do recall from my meanderings on this topic, that Leica persisted with the CCD sensor with one of its later models. Who knows where further development may have taken the CCD? Although it has persisted l believe in photocopy technology. In the visual image comparison you conducted the D200 didn’t perform too shabbily at all. What the histograms revealed about the colours was particularly of interest. So some evolution in the ‘colour science’ has occurred over the model generations. I do agree with the views suggesting it is not all down to a sensor ‘shoot out’, filter configuration, processors etc are all part of the bigger picture.
Yeah there is likely a lot at play rather than just what is being read out of the sensors. But indeed Lieca and Phase One had continued to use CCDs for a long time after. The Lieca M9 used a Kodak sensor and was made in 2009 to 2012. Phase One IQ 3 80 being introduced in 2016. Until the IQ 4 series which was entirely CMOS line up of digital backs.
CCD seems to have better micro contrast. CMOS looks flatter and and more processed (Digital). I think a monochrome photo comparison test would show it better if true. Much easier to see the differences in B&W. I've seen other videos where it shows more color variations in gradients. I think it just looks like more of an analog tonal curve in CCD. CMOS has come a long way and modern sensors give excellent modern looks. They have so much circuitry though in their processing (High transistor count). CCD has much less circuitry but then is speed capped to it's design. To get speed out of CCD's companies like Kodak and Sony split sensors into a multi tap architecture where its like having up to four sensors working as one. They had to be balanced together to make it look seamless. CMOS took over after that as it gave imaging technology much more potential. I wonder how far they could could have taken CCD's overall had CMOS not taken over all development.
I have seen this comment twice here about "mice contest." I could be wrong but micro contrast is found in lenses not camera bodies.
It would have more sense to have compared the D200 (last CCD) with the D300 (first CMOS) in what is almost the same body except for the larger rear screen. Much of the CCD film look is really the processor colour science!
I don't see why it would be better to compare the 2 aside from their generational closeness. I wasn't specifically looking at the sensors in terms of technical capability. (Hence why most of the photos were shot their base ISO, in scenes where the Dynamic range wasn't pushed to the limit). Additionally the D200 and D300 resolutions could have caused more of a look difference than the near identical pixel pitch of the D200 and D600. Admittedly I actively made the choice to examine 2 sensors with the same pixel pitch rather than of similar generation because I personally though I'd have even more emphasis on the differences given most opinions are that modern CMOS look "too clinical". Finally the D200 was later in the life of nikon's CCD but it's years before they stopped using it the D3000 was Nikon's final CCD DSLR (in 2009, 3 years after the D200 release), the D300 is also depending on how you view it the second or third gen DSLR from Nikon with CMOS, the D2h (LBCAST) technology is a variant of CMOS technology, but if you want to exclude that then the D2x would be the first gen Nikon with a CMOS sensor (supposedly designed by Nikon, made by Sony). Although the D300 is the first Nikon DSLR to feature a CMOS that is wholly designed by Sony.
CIS has some huge benefits being so widely used as a sensor compared to CCDs, and becoming more and more so. CCDs have some inherent benefits but with softwares and better interpolation I think there isn't enough in using a CCD.
Yeah the market kind of took the CMOS tech and ran with it.
In regards to your observation on colour transitions are these 12 or 14 bit sensors?
The D200 does 12 bit NEFs the D600 is 14 bits. I didn't think of it at the time but I should have probably set the output to 12bits on the d600.
Thanks. Looks like the CCD has more microcontrast. Would be interesting to saa a comparison with a Sigma Merrill.
Ouu yes sigma cameras would be an interesting way to go. I occasionally look for them but the price of them especially the Merrill variants is a bit too high for me ATM. I may try some of the older ones though.
I had a D200. It was alright. Hated the small LCD. really didn't notice anything spectacular about the sensor. shortly after I got a D7000. That in my opinion walked all over the the D200 sensor wise. Better in every single way. Sold the D200 and Canon 40D and another canon Xsi. Never looked back.
Haha fair enough for me the ergonomics of the camera are what set that camera apart. The small screen resolution I can forgive since that was considered a fairly generous screen for the time. As far as the sensor it's just my preference but I do like it in compared to 10-16mp sensors of the era. Not sure exactly why but I do. But then again I'm not currently very picky about my cameras haha.
@@Mr50mmish Agree on the ergonomics. have a D810 now. that feels good. Also a D4. That has a relatively low Rez sensor. Nice look for portraits.
@@markkasick That sounds like an awesome pairing! Haha one day Imma have to try the D800 line!
ccd vs cmos isnt all that different... if youre going CCD for the CCD look, its the Fujifilm Super CCD DSLR's and the Kodak CCD DSLR's, theres no comparison.
That's fair, gives me an idea though, :D Maybe one day I'll do a comparison between CCDs, Kodak / Fujifilm / Sony / DALSA maybe.
Yes, the CCD matrix gives a more beautiful color, do you know that Panasonic has such matrices?
CCDs are great! I am not aware, from what I know Panasonic in their stills and hybrid cameras use their own NMOS which is also a variant of CMOS
For me, Nikon model after around Expeed 3 era (after 2012) rendered picture more blue leaning and seem to have darker image at the same exposure compare to older camera (might not included some really old model like D1 series or D100 which have somewhat darker image at the same exposure as well though they're different from model with Expeed 3 and later processor), and I have to say their light meter also pick a value around 1 stop under what the older DSLRs or film cameras can metered in bright daylight and even when compensate 1 stop more exposure its picture still not look as bright as picture from film camera when exposed at the same value with the same ISO film. So I think the difference should be credited to how newer image processor and overall camera processing work rather than the different between CCD and CMOS but because newer cameras had been made in a way that they're not as sensitive as older model for some reason and there're somewhat noticable change in Nikon color science in Expeed 3 era and beyond compare to Pre-Expeed era (everything before 2007 which have rather high color intensity while not have so good dynamic range and the image didn't look as clear as newer model) and original Expeed era (D3, D300, D700, D90; which I found their color rendition brighter and clearer than the model before and after while still vibrant; Expeed 3 and later didn't render as vivid color by default anymore and by that time Canon's image look more vibrant which is opposited from before that time).
Yeah so there does seem to be a pretty big change in the way the color science is done on these cameras. Haha I might need to repeat this with a d2x down the line.
Pics from my old Canon 10D and Pentax K10D look a little more saturated than pics from my 5Dii and K20D, respectively. Only reason I noticed that was all the hoo-hah recently about the CCD nostalgia. Yes, there is a slight difference, but not one I'd go out of my way to find or buy another old camera to capture. It's mostly not worth bothering about.
On balance, I'd say the lens used on the day could have more affect on image balance - I shoot a lot of old manual glass and some of the older pics might have got their very slightly more saturated look from the glass in use.
CCD is less sharp and more saturated in those images. Again you can manipulate this in post to a certain extent. If you feel the color transitions are better on the CCD, then maybe micro-contrast is better in your opinion?
The reduced sharpness might be less a consequence of the sensor directly and more related to the Anti aliasing filter on top as they were more aggressive with that. Yeah CCD micro contrast I think could be better or rather more noticeable due to the lower bit depth of the CCD to CMOS. I should have probably switched the D600 to 12 bit output. Maybe another video!
Why not compare the CCD D200 to its immediate successor, the CMOS D300? That seems vastly more enlightening in contrasting the two technologies.
The main reason for the "why didn't I compare it to XX" cameras is don't have one, that was also the case with the D300. Additionally the D600 was an interesting candidate as the pixel pitch is almost exactly the same. I shot all the D600 photos in DX mode, which outputs a raw file at about the same resolution as the D200, this is nice for me since I didn't have to do any sort or scaling, and then I can't chalk up results of one or the other as the pixel density disparity.
That said, I do plan to do something along those lines eventually.
D200 just takes better pics. I even like the 3200 grain on d200. Looks lofi filmic
2005 APS-C/DX 10MP CCD vs 2012 Fullframe (FX) 24 MP CMOS...seems to be unfair. Much better would have been - you took the D90 (2008, 12 MP, CMOS) as comparsion 1:1 with the D200...just saying. 🙂
Yeah, I did that mainly mainly do to the near 1 to 1 pixel scaling (D600 in DX mode is 10mp). Plus given the arguments about the look of CCDs vs CMOS extend beyond similar generation cameras and usually broadly comments go something like "modern CMOs are soo plastic (digital, whatever other comparison)" I figure it shouldn't matter what year the CMOS sensor I used would be.
But otherwise I will probably revisit this in the future for sure! Haha if I can I'll see if I can get a D2x or something closer. Or maybe the Canon 40D.
@@Mr50mmish ok, sorry, i didn't notice that you've had used the DX crop mode. 40D is good, still have mine since 2008. But meanwhile, i do prefer other gear. The 18-135 STM is fine, as shoot it all zoom, btw. Still nowadays. As for the looks, i do always prefer a CCD Sensor vs. CMOS, as for the colors. Don't care about DXO Mark ratings - as long, it's pleasing for the eye.
Fun fact: the Canon Sensors are (mostly) are being CMOS, but even the old ones have a more "filmic" look, compared to their current offerings. 🙂
@@marcp.1752 all good, haha yeah I just wanted to avoid scaling issues is all! But yeah I'll definitely revisit this with similar generation sensor to see if it's more a generational look! Cheers!
@@Mr50mmish All fine. 😀
You should compare them to Foveon :)
I actually do want to give it a ago. Haha if I can find the SD1 Merrill or maybe the quattro for a reasonable price you can definitely expect a video!
was it a blind study?
Haha, definitely not. I'm pretty biased
if not grain, is it noise?
Yeah, but it might be a specific type of noise. I'm not exactly sure what as there are lots of different noise sources between CMOS and CCDs.
@@Mr50mmish if i remember correctly, back in the day cmos sensors were front-side illuminated, while ccd sensors were bsi, which made cmos sensors show a bit more of a vignetting effect since light came in slightly off-angle?
also, there might have been something about higher luma noise but lower chroma noise for ccds compared to cmos sensors back in the day?
maybe those two effects was what made the ccd bokeh have a more "textured" look with more saturated colours?
separate question just out of my own curiosity: did you find dynamic range suffered? ie. some people find canon 5dm3 is more saturated straight out of camera RAW than an r6, cos it has less dynamic range. is that the case for ccds vs cmos?
@@vtct_ytaa that could be it. Giving the noise profile it's look. When it came to overall saturation I actually didn't find the D600 less saturation once I matched the blue channel to the D200. Despite the D600's higher DR. Maybe the R6 flattens the image more to keep images from appearing blown. I do notice that the D600 will just recover more than the D200 as a result. Haha I suppose they might have actively made that choice to keep the images a bit more pleasant!
@@Mr50mmish ahh gotcha, very good video mate. keep up the good work and i look forward to watching more videos from you! Cheers for all the testing you do!
seems like a stronger aa filter on d600
They look fairly similar to me, I guess I should have tried a test chart or something for the video. Although I'm sure somewhere there has to have been a review where someone calculated the effective max resolving power of each sensor!
d700 has much cleaner image
One day I'll get a d700 to review!
Apsc vs Fullframe? Not a fair comparison: you should have compared D200 vs D300.
I use the 2 due to their very similar pixel pitch. Additionally the d600 was shot in DX mode and the image output was effectively the same as the d200 as a result.
The CCD version has somewhat nicer colors, the cmos seems a bit cooler, technical.
But these are ancient cameras the CCD version being just different, not better. The cmos has far better noise performance.
CCD is anything but film, it has ugly color noise. And is just as harsh as any old digital, modern cameras destroy these results.
I think it`s wishfull thinking and pseudo science combined, if you want it to be better you`re definetly going to see it.
Besides, only 2 demo photos, that`s pathetic.
I did say off the bat it's not a technical examination of the 2 sensors, there are plenty of websites that will quantify how much better modern CMOS to CCD is. I am aware that CMOS sensors have progressed a lot and of course modern sensor noise performance will be much better than old CCDs. As for lack of photos given that I only color saw differences in the blues across a few photo for my photos I didn't think I'd need display that multiple times to show it. If you want a technical comparison you are free to play with all the charts you want on DXO.
@@Mr50mmish Man, there really was much more going on with the colors than only the blues. And btw. 2 pics is still a lousy 2 pics.
Far to little in any comparison.
And how come you say there still some magical CCD anologueness going on when all you now say there is only difference in blues.
Seems like a lot of bull to me.
Oh and i hate DXO wit a passion.
Photography is art first and science second, at least for me. And I prefer CCD, the older the better. I find talking about "science" and "pseudo-science" in context of art utterly ridiculous. But that's just me. I'm simply not into forensic photography.
And besides, it seems obvious that the conclusion here that the difference is not in the way it handles color alone, but the way its noise is structured and textured, particularly when it comes to uniform vs detailed areas. I am not aware of any "scientific" metric devised for that so far.