DXO now has MF cameras

[TMARK]

I do not know you guys, but I can see the difference between a 31mp DB  file and a 21mp DSRL file printed at a3 size, not even mention if the DB is 39mp. It take a couple of sec to focus my eyes properly to image detail, but then there will be not doubt between which is which.

About 70% of fashion and portrait advertising on the world, is shot with Canon. If you go through  some well printed fashion magazine (ala Numero, POP, V, W etch) I can tell which is a photo taken with a DB and which one is a Canon. A part from the difference on crispiness, it is the poorer gradient of the skin tones that tells me immediately which one is a canon. I can see the same flatness on skin also on the Nikon shots at the Nikon web site that someone linked above. Especially the beauty shot ( appalling styling and make up by the way), probably because of the light as well, the image is very flat and pasty. To that pastiness it may contribute the lack of mid range contrast typical of some 35mm lenses. If you shot the same image with a 80mm mamiya, zeiss or hasselblad on 30mp back you immediately get more depth on the image. The difference will be like looking at a 35mm contact sheet next to a 6 by 6 made with an hassy on the old days.

Those that cannot discern these differences between images while working with their cameras, should stick to making between 30 to 50 posts a day on multiple photography web forums or looking at Doxo sites. Once you do that every day, when do you find the time to take photos properly and acquire real life experience?

I have to disagree with you on this when it comes to things printed on a web press.  The only way to tell a 5D shot from a P45 shot from drum scanned 4x5 film is in the look.  I base my opinion on my stuff that's been published in some of the better mags and on stuff I lit that was published in the better mags. There probably isn't more than three and a half stops between web press Ink Black (which aint black, by the way) and paper white.  The nuance of an MFDB file (which is there on screen and in a print) is simply lost when printed on any web press.

[lisa_r]

I agree, I shoot main fashion for one of the really big fashion magazines and I can easily see that my pictures are shot with a phase back while others are shot with a canon, it really isn't that hard at all. A well printed magazine page shows the difference just as well as a good inkjet print.

Woof, good work. Perhaps you would say which magazines and issues so we could have a look - this is where the rubber meets the road.

[woof75]

Woof, good work. Perhaps you would say which magazines and issues so we could have a look - this is where the rubber meets the road.

Thanks, I prefer anonymity though.

[ziocan]

I agree, I shoot main fashion for one of the really big fashion magazines and I can easily see that my pictures are shot with a phase back while others are shot with a canon, it really isn't that hard at all. A well printed magazine page shows the difference just as well as a good inkjet print.

Hi there! which one do you use?

[woof75]

Hi there! which one do you use?

P21, anything bigger and you have to down rez for editorial.

[ziocan]

I have to disagree with you on this when it comes to things printed on a web press.  The only way to tell a 5D shot from a P45 shot from drum scanned 4x5 film is in the look.  I base my opinion on my stuff that's been published in some of the better mags and on stuff I lit that was published in the better mags. There probably isn't more than three and a half stops between web press Ink Black (which aint black, by the way) and paper white.  The nuance of an MFDB file (which is there on screen and in a print) is simply lost when printed on any web press.

IMO on the magazines I mentioned above it shows also on the skin nuances. Also on good catalogues like the Borgdof & Goodman.
Especially if shot in studio with Umbrellas or Octas, I can see it around the arms and legs of the models, that Canon shots, have less "shades of pink" and depth, they are not as rich as the shots with a DB. On portraits and beauty, the highlights of lips and nose edge with shiny or wet make up, when shot with canons are more "sparkling" and the last bit of transition to white is steeper, contributing to a more digital look. especially with the older 5d. though the older 5d had a general better skin representations than the 1ds mark 2.
Also majority of fashion photographers and many retouchers simply use Lightroom and ACR, which contributes to average and smooth/clean up skins just on the straight conversion. that does not help either.
those things have definitively improved with the last batch of 20mp+ DSLR, but yet DB still shows its edge on reproducing skin.

cheers.

[woof75]

IMO on the magazines I mentioned above it shows also on the skin nuances. Also on good catalogues like the Borgdof & Goodman.
Especially if shot in studio with Umbrellas or Octas, I can see it around the arms and legs of the models, that Canon shots, have less "shades of pink" and depth, they are not as rich as the shots with a DB. On portraits and beauty, the highlights of lips and nose edge with shiny or wet make up, when shot with canons are more "sparkling" and the last bit of transition to white is steeper, contributing to a more digital look. especially with the older 5d. though the older 5d had a general better skin representations than the 1ds mark 2.
Also majority of fashion photographers and many retouchers simply use Lightroom and ACR, which contributes to average and smooth/clean up skins just on the straight conversion. that does not help either.
those things have definitively improved with the last batch of 20mp+ DSLR, but yet DB still shows its edge on reproducing skin.

cheers.

Glad someone else notices these things. I hate Lightroom and always use capture one. A really good example is work by inez and vinoodh (canon) and mert and marcus (phase).

[Guy Mancuso]

When there is real money and important jobs on the line i will grab the MF system every time, I will use the 35mm for other non essential shooting like events and such. Just that simple, these tests do NOT tell the artistic story and say nothing whatsoever of just downright good raw processing and technique which counts for a lot of the total image. Also says absolutely nothing on how the image renders just like MTF charts. I NEVER bought a lens on MTF charts and will never buy a camera on DXO scores.

[TMARK]

IMO on the magazines I mentioned above it shows also on the skin nuances. Also on good catalogues like the Borgdof & Goodman.
Especially if shot in studio with Umbrellas or Octas, I can see it around the arms and legs of the models, that Canon shots, have less "shades of pink" and depth, they are not as rich as the shots with a DB. On portraits and beauty, the highlights of lips and nose edge with shiny or wet make up, when shot with canons are more "sparkling" and the last bit of transition to white is steeper, contributing to a more digital look. especially with the older 5d. though the older 5d had a general better skin representations than the 1ds mark 2.
Also majority of fashion photographers and many retouchers simply use Lightroom and ACR, which contributes to average and smooth/clean up skins just on the straight conversion. that does not help either.
those things have definitively improved with the last batch of 20mp+ DSLR, but yet DB still shows its edge on reproducing skin.

cheers.

You know what, you got me thinking.  I went back and looked at tears.  Most everything I've shot for magazines has been with either an Aptus or a P30+, or on film.  Mainly film.  I can see a difference between the few 1ds2 shots, but only really in beauty.  Waist up.  Mainly what I see is the more pleasing rendering of the lenses on a larger sensor.  No noticeable resolution differences. But yes, some beauty I shot on a Leaf 75 a few years ago really is noticably better in terms of skin than a similar Canon shot. The retouching was also top notch.

Its funny you mention Bergdorfs, as the stuff I've shot for them was with a 1ds2, a 5d and 645 film.  Now I shoot video for them.

This DxO stuff is a tempest in a teapot.  Really.  MF has its place as does 4x5 as does 35.  Its that simple.  Photographers as opposed to gear collectors know this and use whatever gets the shot.

[BernardLanguillier]

Firstly, I can not see any difference whatsoever between the A900 and the D3x files, not at any print size, not at any ISO (I never shot at above 1600, and rarely at 800).

On prints up to 13X19" I can usually see the difference between the DSLR shots and the P65+ shots, but not always. In some cases it is moderately obvious and in a few is hugely obvious.

In print sizes above about 16X20" I can always see the difference. It is dramatic. Not just resolution (that's a given) but in terms of micro-contrast, smooth tonalities, and depth of colour rendition – those amazing iceberg blues.

Michael,

Thanks for the information. I don't believe that your P65+ vs DSLR will surprise anyone. What would be more fun is a comparison between a P25+ and these DSLRs, or a comparison between a 6 images stitch from a D3x and the P65+...

As far as the lack of difference in print between A900 and D3x, it makes total sense for the kind of scenes you have been shooting in Antartica, especially if ACR/LR was used for conversion. You would probably see a larger gap at A1 if the D3x files were converted with Raw Developper and optimally sharpened but the gap will mostly show in scenes with more dynamic range still.

Cheers,
Bernard

[douglasf13]

As far as the lack of difference in print between A900 and D3x, it makes total sense for the kind of scenes you have been shooting in Antartica, especially if ACR/LR was used for conversion. You would probably see a larger gap at A1 if the D3x files were converted with Raw Developper and optimally sharpened but the gap will mostly show in scenes with more dynamic range still.

Cheers,
Bernard

  This seems to echo the Borg's findings, and they own both cameras.  Iliah says the D3x is slightly better with shadow noise and in high DR landscape, where the A900 is better in lower DR landscape (better green separation,) and portraits of dark skin/haired people.  There is enough difference that Iliah is keeping both cams, although I think a lot of it has to do with the ZA lenses.  I wouldn't say either camera has better IQ, just different, and lens+raw converter selection is critical. Still, I think it's splitting hairs.

[ziocan]

This seems to echo the Borg's findings, and they own both cameras.  Iliah says the D3x is slightly better with shadow noise and in high DR landscape, where the A900 is better in lower DR landscape (better green separation,) and portraits of dark skin/haired people.  There is enough difference that Iliah is keeping both cams, although I think a lot of it has to do with the ZA lenses.  I wouldn't say either camera has better IQ, just different, and lens+raw converter selection is critical. Still, I think it's splitting hairs.

Honestly if the Borg bros process the a900 files with their convertor and using their own profile, I can tell you that they are not using the a900 at its best. their profiles delivers colors that are pretty much off and contrast is also a bit off. I have tested few files using their RPP software but with a color profile made by a third person and the results difference were quite significant.
I think the best way to go, as raw convertor, for the a900 is C1 or Raw Developer.
I doubt that a fashion or portrait/ beauty shot, taken with an a900 and a a 85mm or 135mm Zeiss or the 50mm and 100mm Sony, will look any worst than if it was taken with the nikon.
With those 4 primes, the final look on the images of the Sony, should have an edge on a down town Manhattan, Tokyo, Paris or Milan art director's palate blind taste.

For those who likes splitting hairs, here follow two links, where they tested both a900 and dx3 and with what ever lenses they used. the sony had an hair split advantage of about 50 lines per inch. I do not think those test are the absolute grail of the resolution test, since considering the cameras use the same sensor and have the exact same pixel count, the lens that will be used will dictate the difference on lines per inch. I doubt we are going to see any difference on detail at any print size especially if both are processed with Raw developer or C1.

Enjoy:
http://www.popphoto.com/cameras/5700/nikon...line-page3.html
http://www.popphoto.com/cameras/5624/sony-...tics-page3.html

[douglasf13]

Honestly if the Borg bros process the a900 files with their convertor and using their own profile, I can tell you that they are not using the a900 at its best. their profiles delivers colors that are pretty much off and contrast is also a bit off. I have tested few files using their RPP software but with a color profile made by a third person and the results difference were quite significant.
I think the best way to go, as raw convertor, for the a900 is C1 or Raw Developer.
I doubt that a fashion or portrait/ beauty shot, taken with an a900 and a a 85mm or 135mm Zeiss or the 50mm and 100mm Sony, will look any worst than if it was taken with the nikon.
With those 4 primes, the final look on the images of the Sony, should have an edge on a down town Manhattan, Tokyo, Paris or Milan art director's palate blind taste.

For those who likes splitting hairs, here follow two links, where they tested both a900 and dx3 and with what ever lenses they used. the sony had an hair split advantage of about 50 lines per inch. I do not think those test are the absolute grail of the resolution test, since considering the cameras use the same sensor and have the exact same pixel count, the lens that will be used will dictate the difference on lines per inch. I doubt we are going to see any difference on detail at any print size especially if both are processed with Raw developer or C1.

Enjoy:
http://www.popphoto.com/cameras/5700/nikon...line-page3.html
http://www.popphoto.com/cameras/5624/sony-...tics-page3.html

  Yeah, I use C1 myself, and it seems great with the A900.  The thing about the Iliah Borg is that he prefers a very flat image with desaturated color out of the RAW converter, because he does his contrast and color manipulation all in PS.  Although that may be the way to eek out the best in an image, I'm fine with doing the bulk of my work in the RAW converter.  I saw those popphoto tests, and the resolution of the two cams is pretty much identical.  I think it's interesting that they found the A900's color to be so off.  I haven't seen this complaint too often.

[elf]

Has anyone compared a 5-20 image mosaic or panorama shot with a DSLR to a similar shot with MF?  Will the IQ of the MF still be better when the resolution of the DSLR shot is much higher than the MF shot?

[BernardLanguillier]

Has anyone compared a 5-20 image mosaic or panorama shot with a DSLR to a similar shot with MF?  Will the IQ of the MF still be better when the resolution of the DSLR shot is much higher than the MF shot?

That is indeed exactly what is at stake here, but it seems to be very hard to get a clear answer on that one...

My current eductated bet is that A1 or A0 print size:

- D3x*1 = P25+
- D3x*4 >> P45+
- D3x*6 > P65+
- D3x*9 > drum scanned 4*5

Cheers,
Bernard

[yaya]

- D3x*4 >> P45+
- D3x*6 > P65+
Cheers,
Bernard

Bernard, I'm wondering how much sense does this make in practice and for which application?

[BernardLanguillier]

Bernard, I'm wondering how much sense does this make in practice and for which application?

Well, this makes sense within the confine of those applications that can be addressed well by stitching:

- The shooting takes longer (although I would argue that it is about as fast shoot a 6 frames pano with a D3x than a single one with a P65+ on an Alpa),
- The post-processing takes longer especially when HDR is part of the equation,
- There is always a risk that you messed up when shooting the pano and cannot come up with a usable image of perfect quality (less than 1% as far as I am concerned),
- Some types of moving subjects are not easy to deal with (that includes changing light)
- Critical sharpness is harder to reach in the outdoors at some shutter speeds
- Strobe shooting ends up being more difficult because you need one flash per pano frame
- Basically, it takes more logistics, skills and knowledge than a single capture on top of everything else

But when the conditions are met, my experience is that a stitched file is 100% impossible to distinguish from a single capture.

Stitching is progressing much faster than any other area of photography, both on the hardware (robotized head with fully scriptable control from a PDA/laptop) and software sides (totally automated stitches dealing with HDR and focus stacking).

Cheers,
Bernard

[Carsten W]

On prints up to 13X19" I can usually see the difference between the DSLR shots and the P65+ shots, but not always. In some cases it is moderately obvious and in a few is hugely obvious.

In print sizes above about 16X20" I can always see the difference. It is dramatic. Not just resolution (that's a given) but in terms of micro-contrast, smooth tonalities, and depth of colour rendition – those amazing iceberg blues.

Michael, how much of this difference would you attribute to the P65+, and how much simply to MF DBs? I am strongly considering picking up a Sinar eMotion 54LV, a 22MP back, and am curious...

[csp]

I agree, I shoot main fashion for one of the really big fashion magazines and I can easily see that my pictures are shot with a phase back while others are shot with a canon, it really isn't that hard at all. A well printed magazine page shows the difference just as well as a good inkjet print.

i would call this - self fulfilling prophecy - and i too very much doubt that you are able to see a significant differences between perfect processed  35mm and mf files . using files for printing with resolution  above what offset can handle leads to softer reproduction (rip interpolation ) and not increased detail. even under perfect conditions it is very hard to see a difference in a final offset print  between 180 or 300 dpi.

[woof75]

i would call this - self fulfilling prophecy - and i too very much doubt that you are able to see a significant differences between perfect processed  35mm and mf files . using files for printing with resolution  above what offset can handle leads to softer reproduction (rip interpolation ) and not increased detail. even under perfect conditions it is very hard to see a difference in a final offset print  between 180 or 300 dpi.

It depends what you call significant, if I can see any difference that contributes to the look I want then I call that significant.