Pages: 1 2 3 [4]   Go Down

Author Topic: New MF platform from Leica or Sinar makes sense.  (Read 18976 times)

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: CCD vs CMOS still? Compare at equal sizes please
« Reply #60 on: January 11, 2014, 04:33:55 pm »

Hi,

Unfortunately I can't. But, I have a P45+ 39 MP MFD (CCD) and a Sony Alpha 99 SLT at 24 MP (CMOS). 17 x 24" prints, I cannot really make apart. Colour rendition is pretty close when processed in LR 5.3 using DCP profiles created at the same time at similar conditions. The Alpha is better on DR as the P45+ has noisy shadows.

In a match, I would say the Alpha wins, looses no rounds and wins a few.

I do think that the P45+ has a resolution advantage, I can measure it in Imatest but I may not see it in real world test. Why not? One factor is that both are good enough. THe other factor is that I print smallish, like 17x24", at that print size both are very sharp.

I am not in the testing business. I either shoot Sony Alpha or P45+, it is seldom I shoot comparable images. BUT, using both systems for something like half a year I can say that I don't see any magic advantage of MFD or CCD.

What I see is that the P45+ has an advantage in resolution/MTF and the Alpha's are better regarding DR. I like shooting with the P45+ and Hasselbald 555ELD, but when it really matters it is the P45+/ELD that stays at home.

Best regards
Erik


Can you show me a comparisons between a CCD and a CMOS sensor of the same sensor size and pixel size where the CCD comes out ahead?
Logged
Erik Kaffehr
 

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Sony CCDs vs Canon and Sony CMOS sensors
« Reply #61 on: January 11, 2014, 04:42:58 pm »

I also insist that there is no proven noise advantage of Cmos, but rather it's a makers decision since MF users, traditionally use their cameras under conditions that don't require LL action shooting.
The advantage in noise levels and high ISO, low light performance of modern active pixel CMOS sensors offer CCD is clear, both in tech specs and observations, and is not due only to issues of micro lenses and CFA design. Remember that many early DSLRs used CCDs from Sony and Kodak, and many of them did have micro lenses. Canon instead used CMOS sensors, and gained a big advantage in lower noise and better high ISO performance. Then Sony moved to its Exmor CMOS, and again the advantage in low noise over Sony's own CCDs and Kodak's CCDs was clear.

Spec's and measurements show one big reason why: far lower levels of dark/read noise in active pixel CMOS sensors. This is turn is probably due to the fact that these active pixel CMOS sensors amplify the signal very early, during a direct transfer from each photosite to the edge of the sensor, providing more robustness against subsequent noise sources. CCDs instead use passive transfer of the charge from each photosite to a corner of the sensor, in thousands of hops from one photosite to the next.
« Last Edit: January 11, 2014, 04:45:25 pm by BJL »
Logged

EricWHiss

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2639
    • Rolleiflex USA
Re: New MF platform from Leica or Sinar makes sense.
« Reply #62 on: January 11, 2014, 04:47:34 pm »

RE: CCD vs CMOS.  I can't think of a CMOS sensor that I've felt produced a better looking file than CCD sensors when I step back from the image.  I know there all sorts of quantitative tests that show how great the latest CMOS sensors are in terms of DR and so on, but at the end of the day you have to also look and feel the images (assuming your goal is to make compelling imagery).  

Edit: I'm not saying there will never be one or even some other technology that will best the look of CCD, just so far.
« Last Edit: January 11, 2014, 04:49:17 pm by EricWHiss »
Logged
Rolleiflex USA

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: New MF platform from Leica or Sinar makes sense.
« Reply #63 on: January 11, 2014, 05:23:42 pm »

Hi,

I have only a limited choice P45+ and some Sony CMOS sensor of recent design. I also have an experience going from CCD to CMOS on APS-C. The first CMOS camera I had was the Sony Alpha 700, and it is the camera that produced most of my best images.

My recent experience is shooting with a Sony Alpha 99 SLT for about one year and with the Phase One P45+ for about half a year. My objectives are to get good images and credible color. What I can say is that I cannot see a lot of difference. The P45+ is a bit sharper, but I cannot really see this in 17x24" prints, larger prints probably but 17x24" is my normal print size. Regarding colour the P45+ has yellowish greens you love it or you hate it. I had some success getting decent colour of P45+ and LR5.3 using home made DCP profiles. I have tested Capture One for something like nine months but I decided that we don't make friends, so C1 went to waste basket. I love Lightroom workflow and tools, so I do have bias for Lightroom, that program has been invented for me!

Obviously, different posters have different views. That depends on different factors. I always loved Velvia, well, except for the colours and I always hated Ektachrome, but a couple of friends of mine loved Ektachrome, without realising how bad that film actually was ;-) I originally liked Kodachrome, than Kodachrome was sort of out of question so I started looking for alternatives and wound up with Velvia. I am not sure I like the colours but I like a lot else, not to have to deal with Kodak was one of those things.

Best regards
Erik


RE: CCD vs CMOS.  I can't think of a CMOS sensor that I've felt produced a better looking file than CCD sensors when I step back from the image.  I know there all sorts of quantitative tests that show how great the latest CMOS sensors are in terms of DR and so on, but at the end of the day you have to also look and feel the images (assuming your goal is to make compelling imagery).  

Edit: I'm not saying there will never be one or even some other technology that will best the look of CCD, just so far.
« Last Edit: January 11, 2014, 05:25:23 pm by ErikKaffehr »
Logged
Erik Kaffehr
 

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: New MF platform from Leica or Sinar makes sense.
« Reply #64 on: January 11, 2014, 05:32:28 pm »

RE: CCD vs CMOS.  I can't think of a CMOS sensor that I've felt produced a better looking file than CCD sensors when I step back from the image.
I have to keep asking this question: are you looking at images from sensors of equal size, and better yet from the same sensor designer/maker, so as to have some chance of having the same CFA using approach? Maybe from Nikon or Pentax SLRs during their transition from Sony CCD to Sony CMOS?

Because discussions like this keep confounding multiple differences, partly due to looking mainly at the CCDs designed for medium format systems, ignoring the APS-C and 4/3" sized ones that were used before CMOS took over down there.  After all, the "MF look" is a thing with film too.
Logged

Theodoros

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2454
Re: New MF platform from Leica or Sinar makes sense.
« Reply #65 on: January 11, 2014, 05:42:53 pm »

RE: CCD vs CMOS.  I can't think of a CMOS sensor that I've felt produced a better looking file than CCD sensors when I step back from the image.  I know there all sorts of quantitative tests that show how great the latest CMOS sensors are in terms of DR and so on, but at the end of the day you have to also look and feel the images (assuming your goal is to make compelling imagery).  

Edit: I'm not saying there will never be one or even some other technology that will best the look of CCD, just so far.
I agree to all the above, my observation also gives the DR advantage to my MF 22mp Kodak sensor than either my 36mp or 16mp FF sensors… The two FF sensors do record a little more DR but in high contrast situations, there is no way to keep most of it so as to make the picture look realistic, I've come to care more of usable DR than recorded DR and there the 22mp sensor comes ahead and even the 16mp Cmos FF sensor beats my other 36mp Cmos sensor. Another thing to consider, is that I've find that the CCD I shoot it -1EV than what I shoot the two Cmos sensors. Another observation is that the makers dedicated program does develop the files clearly better than any other program (like LR), but again, MF benefits even more in both DR and colour. As far as colour is concerned I think there is simply no comparison.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: New MF platform from Leica or Sinar makes sense.
« Reply #66 on: January 11, 2014, 05:50:30 pm »

Here is a sample of mixed images from MFD and DSLR I have posted here on LuLa a while ago. Some images are P45+ and some Sony Alpha 99 SLT. All images are downsized to 4000 pixels width mostly to get relevant actual pixels views, screen is normally about 100 PPI, but we normally print at 180-360 PPI. The images are from an actual shot, this is not testing, just using both systems in the field.

Not the comparison you ask for but the best I have.

http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/RealWorld/

Here is info on which is which:

http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/RealWorld/Answers.html

Best regards
Erik

I have to keep asking this question: are you looking at images from sensors of equal size, and better yet from the same sensor designer/maker, so as to have some chance of having the same CFA using approach? Maybe from Nikon or Pentax SLRs during their transition from Sony CCD to Sony CMOS?

Because discussions like this keep confounding multiple differences, partly due to looking mainly at the CCDs designed for medium format systems, ignoring the APS-C and 4/3" sized ones that were used before CMOS took over down there.  After all, the "MF look" is a thing with film too.
Logged
Erik Kaffehr
 

Theodoros

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2454
Re: New MF platform from Leica or Sinar makes sense.
« Reply #67 on: January 11, 2014, 06:27:50 pm »

Here is a sample of mixed images from MFD and DSLR I have posted here on LuLa a while ago. Some images are P45+ and some Sony Alpha 99 SLT. All images are downsized to 4000 pixels width mostly to get relevant actual pixels views, screen is normally about 100 PPI, but we normally print at 180-360 PPI. The images are from an actual shot, this is not testing, just using both systems in the field.

Not the comparison you ask for but the best I have.

http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/RealWorld/

Here is info on which is which:

http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/RealWorld/Answers.html

Best regards
Erik

If I may suggest Erik, since you love LR for all your work, I propose to try the following for your P45+ files, …do some quick (but careful) development in C1P1 for the exposure/DR and colour temperature you want and then export the file as a DNG, use the DNG in LR along with your Sony files, I suspect that this may increase DR considerably for files from your P45+ and you may find that colour and sharpness may benefit too.
Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: New MF platform from Leica or Sinar makes sense.
« Reply #68 on: January 11, 2014, 08:23:43 pm »

BJL,

I have no problem admitting that files processed by Lightroom from CMOS and CCD are often equivalent. At this point I think the discussion is over.

 As for the remark about color coming from CFA, I think other design factors affect color like crosstalk in the sensor.

Edmund
« Last Edit: January 11, 2014, 08:35:23 pm by eronald »
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
minimizing extraneous differences when comparing CCD to CMOS
« Reply #69 on: January 11, 2014, 08:32:59 pm »

Here is a sample of mixed images from MFD and DSLR I have posted here on LuLa a while ago.

Here is a thought: since many current sensors in both 35mm format and DMF have 6 micron pixel pitch (24MP 35mm format CMOS sensors, and the 50MP and 60MP MF sensors) a comparison of crops to 6000x4000 from various "6 micron" sensors using lenses of the same focal length could be informative. That would not avoid differences in CFA design approach though (which is why I suggested "Sony vs Sony" comparisons).
Logged

synn

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1235
    • My fine art portfolio
Re: New MF platform from Leica or Sinar makes sense.
« Reply #70 on: January 11, 2014, 09:09:10 pm »

Are we still doing this CCD vs CMOS thing?

I really feel that some of you guys should take  a look at the big picture. Pun intended. I don't care how much superior CMOS is on paper, not a single CMOS camera has produced images that look as appealing as a CCD camera till date. The Leica M240 vs M9 example Ronald brought up is a good one.
Having started my digital journey with CCD APSC (Nikon D70s), I can certainly say that I had that look, lost it when I moved to CMOS and now have it again with MFD. I have shown several examples here before, but people tend to ignore images and keep going back to paper spec differences to prove their point.

I don't buy the "It's the filters" argument either. If it were just the filters, some manufacturer at some point of time would have attempted a CMOS based camera that puts color accuracy as top priority over everything else such as high ISO. Hasn't happened yet.

As a final note, here are two image shot yesterday. Same model, same location, same lighting.



The framing isn't exactly the same because:

 - I didn't have lenses with equivalent FoVs for both cameras
 - A model isn't a brick wall. She will keep moving

 Look at the tonal separation between the reds in the dress and the reds in the skintones in the Credo file. In the D800 file, no matter what I do, adjusting one changes the other (Without using masks etc). Of course, the Credo file also has that "3D" look that comes with MFDBs.
If you want to ignore the image and go back to paper spec discussions, feel free. But as an artist, I trust what I see and for me, there's no comparison.

Logged
my portfolio: www.sandeepmurali.com

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: New MF platform from Leica or Sinar makes sense.
« Reply #71 on: January 11, 2014, 09:22:25 pm »

Synn, it won't work, most engineers are guys with average "guy" color vision so they simply don't *see* that there is a problem.

Unfortunately, in your example, I like the D800 image because the model looks more .. 3 dimensional :)

Edmund

Are we still doing this CCD vs CMOS thing?

I really feel that some of you guys should take  a look at the big picture. Pun intended. I don't care how much superior CMOS is on paper, not a single CMOS camera has produced images that look as appealing as a CCD camera till date. The Leica M240 vs M9 example Ronald brought up is a good one.
Having started my digital journey with CCD APSC (Nikon D70s), I can certainly say that I had that look, lost it when I moved to CMOS and now have it again with MFD. I have shown several examples here before, but people tend to ignore images and keep going back to paper spec differences to prove their point.

I don't buy the "It's the filters" argument either. If it were just the filters, some manufacturer at some point of time would have attempted a CMOS based camera that puts color accuracy as top priority over everything else such as high ISO. Hasn't happened yet.

As a final note, here are two image shot yesterday. Same model, same location, same lighting.



The framing isn't exactly the same because:

 - I didn't have lenses with equivalent FoVs for both cameras
 - A model isn't a brick wall. She will keep moving

 Look at the tonal separation between the reds in the dress and the reds in the skintones in the Credo file. In the D800 file, no matter what I do, adjusting one changes the other (Without using masks etc). Of course, the Credo file also has that "3D" look that comes with MFDBs.
If you want to ignore the image and go back to paper spec discussions, feel free. But as an artist, I trust what I see and for me, there's no comparison.


« Last Edit: January 11, 2014, 09:24:30 pm by eronald »
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

synn

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1235
    • My fine art portfolio
Re: New MF platform from Leica or Sinar makes sense.
« Reply #72 on: January 11, 2014, 09:30:38 pm »

Haha well, I actually think the Credo file looks more 3 Dimensional. Maybe it's the downscaling. But at a decent madnification, the D800 file certainly looks a lot flatter.
Logged
my portfolio: www.sandeepmurali.com

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: New MF platform from Leica or Sinar makes sense.
« Reply #73 on: January 12, 2014, 12:17:47 am »

Hi,

I have tried Capture One for I guess nine months but deinstalled a couple of days ago, we never made friends. I definitively don't think C1 gives more DR at least not on my exposures. I expose for the highlights, so what limits my DR is shadow noise.

Thanks for the suggestion, anyway.

Best regards
Erik

If I may suggest Erik, since you love LR for all your work, I propose to try the following for your P45+ files, …do some quick (but careful) development in C1P1 for the exposure/DR and colour temperature you want and then export the file as a DNG, use the DNG in LR along with your Sony files, I suspect that this may increase DR considerably for files from your P45+ and you may find that colour and sharpness may benefit too.
Logged
Erik Kaffehr
 

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: New MF platform from Leica or Sinar makes sense.
« Reply #74 on: January 12, 2014, 02:26:00 am »

Hi,

My two first DSLRs had CCD, it was the Konica Minolta Dimage 7D (6MP) and the Sony Alpha 100. There might have been a significant rendering difference between the two. After that I got a Sony Alpha 700 which was my first CMOS camera and I made my best images using that camera (much depending on travel opportunities).

I have found two images shot same day, place and subject: http://echophoto.dnsalias.net/ekr/Articles/CMOS_VS_CCD/

Those images were processed to my taste the same time. The link above shows to both JPEGs and raw images.

Best regards
Erik




I have to keep asking this question: are you looking at images from sensors of equal size, and better yet from the same sensor designer/maker, so as to have some chance of having the same CFA using approach? Maybe from Nikon or Pentax SLRs during their transition from Sony CCD to Sony CMOS?

Because discussions like this keep confounding multiple differences, partly due to looking mainly at the CCDs designed for medium format systems, ignoring the APS-C and 4/3" sized ones that were used before CMOS took over down there.  After all, the "MF look" is a thing with film too.
Logged
Erik Kaffehr
 

Theodoros

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2454
Re: New MF platform from Leica or Sinar makes sense.
« Reply #75 on: January 12, 2014, 04:21:57 am »

Haha well, I actually think the Credo file looks more 3 Dimensional. Maybe it's the downscaling. But at a decent madnification, the D800 file certainly looks a lot flatter.
I would add some (maybe 8%) warming filter to the Credo file… Certainly the dress on the Credo is red, while in D800 it's orange but the face is over saturated pink… Same problems I get from my D800E with respect to my MFDB… The MFDB is well ahead in both tones and colour.
 
P.S: Not to mention the blue cast that has been developed on the D800 (which is what happens in my case too) in the deep shadows of the background which IMO makes the image unusable. Clearly, if one was not to "push" to expose LL information (which is what creates the blue cast) on the D800, the image would show much less DR than its MFDB counterpart. Therefore, if one seeks for"correct" development, MFDBs clearly have more "usable" DR than any DSLR…
P.S-2: Interesting part is that D4 also beats D800 at base ISO for "usable" DR, that said, it is still well behind my MFDB, but it does present the best IQ out of all Cmos sensor DSLRs I've tried. This, in combination with the fact that D4 has the fattest pixels among all DSLRs and the fact that many find some "magic" to the old Kodak 22mp 37x49mm sensor, leads me to the conclusion that more than CCD vs. Cmos differences, pixel size may have shrunk more than it should… Heck, I never missed resolution for single shot needs from any MFDB and whenever there was extreme resolution needed (like in reproduction of big size paintings), the resolution and colour of high-resolution single shot backs was never enough… Only 16x multishot (microstep), is in a class of its own for resolution, DR and colour.
« Last Edit: January 12, 2014, 05:55:02 am by T.Dascalos »
Logged

synn

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1235
    • My fine art portfolio
Re: New MF platform from Leica or Sinar makes sense.
« Reply #76 on: January 12, 2014, 10:17:08 am »

I would add some (maybe 8%) warming filter to the Credo file… Certainly the dress on the Credo is red, while in D800 it's orange but the face is over saturated pink… Same problems I get from my D800E with respect to my MFDB… The MFDB is well ahead in both tones and colour.
 
P.S: Not to mention the blue cast that has been developed on the D800 (which is what happens in my case too) in the deep shadows of the background which IMO makes the image unusable. Clearly, if one was not to "push" to expose LL information (which is what creates the blue cast) on the D800, the image would show much less DR than its MFDB counterpart. Therefore, if one seeks for"correct" development, MFDBs clearly have more "usable" DR than any DSLR…
P.S-2: Interesting part is that D4 also beats D800 at base ISO for "usable" DR, that said, it is still well behind my MFDB, but it does present the best IQ out of all Cmos sensor DSLRs I've tried. This, in combination with the fact that D4 has the fattest pixels among all DSLRs and the fact that many find some "magic" to the old Kodak 22mp 37x49mm sensor, leads me to the conclusion that more than CCD vs. Cmos differences, pixel size may have shrunk more than it should… Heck, I never missed resolution for single shot needs from any MFDB and whenever there was extreme resolution needed (like in reproduction of big size paintings), the resolution and colour of high-resolution single shot backs was never enough… Only 16x multishot (microstep), is in a class of its own for resolution, DR and colour.

I agree on all counts!

It's good to see a D800(e) user who accepts it for what it is (and understands its capabilities AND shortcomings) rather than defend it fiercely as if it can do no wrong.

The Credo file was indeed a tad cold for my liking (I blame it on the slightly cool strobe), so I did do some toning in the final version. You can see it here: http://www.sandeepmurali.com/p879687852/h1CBBE5C8#h1669a012

I have not shot with a D4 in the studio, but I have seen a shot from Joe McNally under strobes with the D4 that was rather impressive. I would still pick my "Old tech" CCD equipped MFD though.

I also agree about the usable DR part. I am a moron and shot a strobe lit shot right after shooting some available light shots at 800 ISO while the metering was for 50. 4 stops overexposed! There was nothing but white in the shot and I still managed to recover a perfectly usable image that holds up for everything except extreme pixel peeping. Do that with a D800 file and you'll end up with a lot of grey patches with no color info.
« Last Edit: January 12, 2014, 08:36:57 pm by synn »
Logged
my portfolio: www.sandeepmurali.com
Pages: 1 2 3 [4]   Go Up