Analog VS High-End-Digital: 80 Mpx, 4x5 Film, Mamiya Macro, IQSmart 2

[PaulSchneider]

Hi guys,

I'm writing this because I've been eagerly following the discussing ensuing the article from Mr. Zuber on the front page.

At some point, I found the results didn't match my experience with both media. I can say that since I'm in the lucky position to both own complete 4x5, 8x10, 80 Mpx, Alpa setups as well as an IQSmart 2 scanner. I also own the arguably best lenses for these systems such as Sironar-S's and SSXLs in different focal lengths, the Phase One LS lenses, the mamiya macro lens and a SK digitar lens for the Alpa.

Basically my impression was that the supposed advantage of digital was emphasized way too much and that many members of the forum who owned significant digital outfits jumped all over the results feeling comforted in their decision to buy into such expensive systems and shoot with it. That's why I feel that I could throw in a few of my findings for what they're worth in order to maybe put another angle into this discussion.

It is important to mention that I work a lot and that photography to me is just a way of escaping and having a passionate time doing something I really like. Hence my goal with this post is to uniquely discuss the technical aspects of the media and not my photography per se which I want to keep private. I'm quite curious and I love the technical aspects of photography, which is also the reason why the article of Mr. Zuber piqued my interest to start analyzing the differences of analog and digital more profroundly. First and foremost I want to know an answer for myself and I hope my findings might be of interest for others too who do not have the means to compare both media as I do at this moment in time.

I sincerely think that this forum has a heavy pro-digital bias and that this skews the objectivity of some comments. It is clear that working professionals who shoot fashion or architecture will laud the advantages an 80 MPX kit affords; but this shouldn't detract from the fact that a well-executed LF image can yield great results that can still exceed digital in terms of feel, special something and, yes, resolution. This is my personal opinion and this being said I would like to throw in my preliminary findings for what they're worth.

WHAT I COMPARED

Equipment used:

Analog: SSXL 150 @ f22, Expired Kodak Ektachrome shot on a Linhof TA 45s
Digital: Scanned @ 4000 DPI with IQsmart 2, non-wet
Shot with Mamiya Macro 120mm @ f16 on a light-table @ 1:1 and so as to cover the whole 4x5 as an 80 Mpix shot
Digital Back: Leaf Aptus 12 @ ISO 50 with a cable release on a DF body

The comparison: I took 4x5 chrome of a night scene I had lying around and compared different means of digitzing it as well as a digital shot of a similar scene in order to visually appreciate the resolution differences.

What I found, purely subjectively:

1. Digitizing film with a 80 Mpx back on a light table is almost as good as my IQSmart! Extremely easy and fast. When buying the IQSmart I compared it to a Flextight and found the Flextight to be less good. So basically if one shoots 4x5 digitizing a a negative or chrome with a newest generation back almost completely eradicates the need for a scanner that maxes out at 2000 DPI. Attaching the back to a mac workstation would let your digitize 30-40 4x5 in an hour with great quality.

2. The IQSmart does yield higher quality overall, even compared to a 1:1 shot of the film but the advantage really is slight. If one prints a 320 MPX 4000 DPI scan at 300 DPI it is more practical to use the IQSmart, because the grain structur looks nicer, the image is slightly sharper and because you don't need to stitch. Scanning at that resolutions take long though, never counted the exact minutes.

You could stitch multiple 1:1 shots and get similar resolution, but it is more tedious.

Uprezzing the full-frame shot of the 4x5 yields, I think very, very acceptable results, especially when printed at 300 DPI.

3. The sharpness of the full-frame shot of the 4x5 is not that far off from a shot of a 80 MPX back in purely digital means. 8x10 ist twice the linear resolution, I hence sincerely believe that 8x10 still can create unsurpassed image quality.

4. I find the analog picture, the colors a lot more pleasing. I think at high enlargements the analog print would look amazing.

5. I love both media, but analog in now way is as low-rez as it has been put forht in that article.

If I find time, I will conduct a controlled test of the same scene with both 8x10 and 80 MPX and I'm sure that 8x10 will hold its own.

The images that follow are:

1. The full scene



2. A comparison of a sharp detail of the scene digitized and sharpened pleasingly for me with the different media (scanner and macro setup)



3. A comparison of an 80 MPX detail shot, sharpened with a 100% crop of a full frame macro shot of a 4x5 chrome.


Shot at 2011-10-04

To conclude, the 4x5 is possibly a little less sharp and detailed than the 80 MPX shot; but 8x10 has 4x times the area, so in my opinion it is safe to say that an expertly executed 8x10 shot with the best lenses and scanned on a high-end scanner will offer more resolution than high-end digital. This is my finding, regardless of all the mathematic approaches or other tests conducted.

Also I find the grain structure and the colors a lot more pleasing. The grain structure allows for enlargements that look nicer to me. Especially the IQSmart produces an exquisit 320 MPX file I would love to print at 1.75m wide. I believe that for ultra large prints in an fine art setting I would always prefer analog over digital. Digital in my eyes makes perfect colors, and all to easily has a clinic look to it. It is too perfect. Of course there are curve tweaks, noise plugins etc. but it still isn't the same. Especially the last linked image above exemplifies this. I love the digital back for its strengths: i.e., speed, sharpness etc. But the look of analog gives it a painterly look which immediately reminds me of that special something I feel when I see art and not a glossy prospectus. Also be aware that I missed focus and only really the signs and the post are sharp on the analog pic and the house is already a bit out of focus...

Kindest regards to you all!

Paul

[PaulSchneider]

The linked images are of higher resolution but the forum makes them smaller. Any way I can make the forum display them at native size?

[ErikKaffehr]

Hi,

I'd say that the best way of publishing large size images is to use attachements. But you need to make them relatively small.

http://www.luminous-landscape.com/forum/index.php?topic=58067.msg468976#msg468976

Another comment, in my view, the best way to compare images is to scale them to same size. So, if I presume that I would print say 70x100 cm at 200PPI I would scale both images to 70x100 at 200PPI,
open both in Photoshop, zoom in to actual pixels, tile the images and than goto to Window->Arrange->MatchAll to get same crop. This is extreme pixel peeping. Zooming 1:2 may be more correct, as screen is normally around 100PPI.

Best regards
Erik

The linked images are of higher resolution but the forum makes them smaller. Any way I can make the forum display them at native size?

[ErikKaffehr]

Hi,

So you essentially find that a macro shoot of the 4x5 Kodachrome using the Aptus comes close to a scan using you IQSmart 2, but the IQSmart surpasses the Aptus macro shot for detail?

Regarding the night shot I'd suggest that I would like to see the film image and the Aptus image scaled to same size. I get the impression that you show both images at different size?

Apart from those comments I find your comparison informative.

I was surprised at the findings of the original article by Mr. Zuber. I'm hoping that they will rescan the 8x10 images. But I don't think that you would get twice the (linear) resolution from the 8x10 compared to 4x5", if you are using the best technique and lenses on both if stopping down to f/32 on 8x10" and 4x5" at f/11. Diffraction would play a very significant role at f/32. What I have seen on smaller formats, stopping down to f/f22 essentially halves linear resolution, turning a 40 MP MFDB into a 10MP one!

Best regards
Erik

Hi guys,

I'm writing this because I've been eagerly following the discussing ensuing the article from Mr. Zuber on the front page.

At some point, I found the results didn't match my experience with both media. I can say that since I'm in the lucky position to both own complete 4x5, 8x10, 80 Mpx, Alpa setups as well as an IQSmart 2 scanner. I also own the arguably best lenses for these systems such as Sironar-S's and SSXLs in different focal lengths, the Phase One LS lenses, the mamiya macro lens and a SK digitar lens for the Alpa.

Basically my impression was that the supposed advantage of digital was emphasized way too much and that many members of the forum who owned significant digital outfits jumped all over the results feeling comforted in their decision to buy into such expensive systems and shoot with it. That's why I feel that I could throw in a few of my findings for what they're worth in order to maybe put another angle into this discussion.

It is important to mention that I work a lot and that photography to me is just a way of escaping and having a passionate time doing something I really like. Hence my goal with this post is to uniquely discuss the technical aspects of the media and not my photography per se which I want to keep private. I'm quite curious and I love the technical aspects of photography, which is also the reason why the article of Mr. Zuber piqued my interest to start analyzing the differences of analog and digital more profroundly. First and foremost I want to know an answer for myself and I hope my findings might be of interest for others too who do not have the means to compare both media as I do at this moment in time.

I sincerely think that this forum has a heavy pro-digital bias and that this skews the objectivity of some comments. It is clear that working professionals who shoot fashion or architecture will laud the advantages an 80 MPX kit affords; but this shouldn't detract from the fact that a well-executed LF image can yield great results that can still exceed digital in terms of feel, special something and, yes, resolution. This is my personal opinion and this being said I would like to throw in my preliminary findings for what they're worth.

WHAT I COMPARED

Equipment used:

Analog: SSXL 150 @ f22, Expired Kodak Ektachrome shot on a Linhof TA 45s
Digital: Scanned @ 4000 DPI with IQsmart 2, non-wet
Shot with Mamiya Macro 120mm @ f16 on a light-table @ 1:1 and so as to cover the whole 4x5 as an 80 Mpix shot
Digital Back: Leaf Aptus 12 @ ISO 50 with a cable release on a DF body

The comparison: I took 4x5 chrome of a night scene I had lying around and compared different means of digitzing it as well as a digital shot of a similar scene in order to visually appreciate the resolution differences.

What I found, purely subjectively:

1. Digitizing film with a 80 Mpx back on a light table is almost as good as my IQSmart! Extremely easy and fast. When buying the IQSmart I compared it to a Flextight and found the Flextight to be less good. So basically if one shoots 4x5 digitizing a a negative or chrome with a newest generation back almost completely eradicates the need for a scanner that maxes out at 2000 DPI. Attaching the back to a mac workstation would let your digitize 30-40 4x5 in an hour with great quality.

2. The IQSmart does yield higher quality overall, even compared to a 1:1 shot of the film but the advantage really is slight. If one prints a 320 MPX 4000 DPI scan at 300 DPI it is more practical to use the IQSmart, because the grain structur looks nicer, the image is slightly sharper and because you don't need to stitch. Scanning at that resolutions take long though, never counted the exact minutes.

You could stitch multiple 1:1 shots and get similar resolution, but it is more tedious.

Uprezzing the full-frame shot of the 4x5 yields, I think very, very acceptable results, especially when printed at 300 DPI.

3. The sharpness of the full-frame shot of the 4x5 is not that far off from a shot of a 80 MPX back in purely digital means. 8x10 ist twice the linear resolution, I hence sincerely believe that 8x10 still can create unsurpassed image quality.

4. I find the analog picture, the colors a lot more pleasing. I think at high enlargements the analog print would look amazing.

5. I love both media, but analog in now way is as low-rez as it has been put forht in that article.

If I find time, I will conduct a controlled test of the same scene with both 8x10 and 80 MPX and I'm sure that 8x10 will hold its own.

The images that follow are:

1. The full scene



2. A comparison of a sharp detail of the scene digitized and sharpened pleasingly for me with the different media (scanner and macro setup)



3. A comparison of an 80 MPX detail shot, sharpened with a 100% crop of a full frame macro shot of a 4x5 chrome.


Shot at 2011-10-04

To conclude, the 4x5 is possibly a little less sharp and detailed than the 80 MPX shot; but 8x10 has 4x times the area, so in my opinion it is safe to say that an expertly executed 8x10 shot with the best lenses and scanned on a high-end scanner will offer more resolution than high-end digital. This is my finding, regardless of all the mathematic approaches or other tests conducted.

Also I find the grain structure and the colors a lot more pleasing. The grain structure allows for enlargements that look nicer to me. Especially the IQSmart produces an exquisit 320 MPX file I would love to print at 1.75m wide. I believe that for ultra large prints in an fine art setting I would always prefer analog over digital. Digital in my eyes makes perfect colors, and all to easily has a clinic look to it. It is too perfect. Of course there are curve tweaks, noise plugins etc. but it still isn't the same. Especially the last linked image above exemplifies this. I love the digital back for its strengths: i.e., speed, sharpness etc. But the look of analog gives it a painterly look which immediately reminds me of that special something I feel when I see art and not a glossy prospectus. Also be aware that I missed focus and only really the signs and the post are sharp on the analog pic and the house is already a bit out of focus...

Kindest regards to you all!

Paul

[yaya]

Paul if you get a chance try to shoot the slide again with the 120mm at f8-f11 as I think you might be able to squeeze a few more pixels that way

One point that was not discussed yet it is that all modern scanners use CCDs that were designed in the 1990's and many of these sensors are no longer being produced.

A. The IQ2 sample shows how good they still are
B. It is safe to say that the sensors of a newer design, such as the one in the Aptus-II 12, are potentially better in terms of DR, sharpness and possibly even colour

Cheers yair

[Fine_Art]

2. A comparison of a sharp detail of the scene digitized and sharpened pleasingly for me with the different media (scanner and macro setup)

So once again, from a separate source, you can see a 4000dpi scan holds more detail than an 80MP back. If you right click, view full size, you will see the 4000dpi scan has detail at the pixel level just like the digital and it is about 50% bigger. So 4x5 has more information than an 80MP back.

There is no point in scaling the digital up to match, it will clearly lose.

Its still easier to use digital but you pay a fortune for the 'fast food'.

[Fine_Art]

Notice noise software cleans the film grain quite well. Here it is applied to the whole thing (all 3) with some sharpening. The sharpening doesnt do anything for the digital; it works on the scan.

[yaya]

So once again, from a separate source, you can see a 4000dpi scan holds more detail than an 80MP back. If you right click, view full size, you will see the 4000dpi scan has detail at the pixel level just like the digital and it is about 50% bigger. So 4x5 has more information than an 80MP back.

There is no point in scaling the digital up to match, it will clearly lose.

Its still easier to use digital but you pay a fortune for the 'fast food'.

In this example, it seems to hold more detail than an 80MP reproduction of the slide (which I believe can benefit from a slightly larger aperture). But you may have missed Paul's comment above: "To conclude, the 4x5 is possibly a little less sharp and detailed than the 80 MPX shot".

The IQ2 is the ideal setup for reproducing the slide but as Paul has just experienced, the speed and ease of using a single shot back for this purpose can make the "fast food" into "fast turnover"...

[Dustbak]

I remember in 1998 the discussions that started about digital never being able to get to the quality of 35mm. Look where we are now...

Anyway, IMO, the people that use 8x10 use it for different reasons than the reasons why people use 80Mpx backs in most cases. At least the 8x10 (or bigger) shooters that I know. Most people I know that use 8x10 would not swap it for digital for whatever reason. I am glad there are still people using this, I still find a beautiful 8x10 having something magical.

[PaulSchneider]

It is exactly as Yair put it: I just tried it out because I had a light table, the camera and the slide around. What I found surprised me: In 5 seconds I can do a 4x5 reproduction, 80 Mpx in size, of a 4x5 that rivals the quality of the IQSmart. The latter is still better, I like the image more. But it a lot more practical to do it that way. Even when printing large, the 80 Mpx repdoduction method looks great. I think it is about 90% as good as the IQSmart scan.

The second thing I found was that if you get a sharp portion of an LF image that 80 Mpx reproduction doesn't look so much less sharp than a compareable digital shot of the same scene. Subjectively I'd say the 80 Mpx possibly has 20-25% more detail? From that I conclude that an 8x10, well-executed, must have higher resolution than an 80 Mpx system.

The problem is to get everything in sharp focus in LF ... this is a lot easier with MF digital ...

[torger]

On the resolution issue, lenses seems to be a rather limiting factor. Even if film grain on 8x10 would allow for fantastic resolution it might not be possible to resolve all that with lenses which has an image circle covering 8x10. So I would not be surprised if an 80 megapixel back with the best "digital" lenses would outresolve 8x10 film due to that you would not find sharp enough lenses with that kind of image circle. I don't know what the case is though.

As far as I can understand, those shooting film today don't do it for the resolution, there are other qualities in the look-and-feel which is interesting. As an engineer I'm still interested in this type of comparisons.

[ced]

Paul nice of you to do this test as I am a firm proponent of using the Hi Res Backs as a repro tool,  speed and quality hold quite well.
I think setting up a permanent bench whereby a good macro lens, transparent glass mask with flash etc. can be left in place will save one even more time if one wants to use the system as a working tool as the setting and focussing take a bit of time to get right. 
You may write of your observations on this point for the sake of interest to others.
As you are comparing scanning and digitising you have to be very careful of using the same kind of sharpening and choice of end points and gradation in general because the 2 systems are very different not to even mention colour correction.
I wonder why the back manufacturers haven't exploited this potential to make a system and bring extra income in marketing a nice setup which will sell quite well I think.  The photographer can also squeeze more out of his investment if he has an archive to digitise and even a service to render.

[BobDavid]

I've been using this techniqe for years with a 39 MP multi-shot back and a macro lens. It is a great way to capture slides and negatives. It would be interesting to compare the results of a 39 MP multi-shot back with an single-shot 80 MP back.

[ced]

Quote: Fine_Art
Notice noise software cleans the film grain quite well. Here it is applied to the whole thing (all 3) with some sharpening.
The sharpening doesnt do anything for the digital; it works on the scan.

This is not a very good comparison as all three images are so different in contrast that the apparent details are affected.

[Fine_Art]

Quote: Fine_Art
Notice noise software cleans the film grain quite well. Here it is applied to the whole thing (all 3) with some sharpening.
The sharpening doesnt do anything for the digital; it works on the scan.

This is not a very good comparison as all three images are so different in contrast that the apparent details are affected.

I think you are right. What I wanted to demonstrate is that the speckling pattern on film can be greatly reduced.

[harlemshooter]

I sincerely think that this forum has a heavy pro-digital bias and that this skews the objectivity of some comments.

Absolutely correct, and most (unlike me) are commercial photographers or equipment dealers who make their dough from it.

[fotometria gr]

I've been using this techniqe for years with a 39 MP multi-shot back and a macro lens. It is a great way to capture slides and negatives. It would be interesting to compare the results of a 39 MP multi-shot back with an single-shot 80 MP back.

It would be even more interested to add a 22mpx 16x image into the above comparison and also.... to compare 4x5 with the largest possible 120/220 film (say 6x8 to keep close proportions?) on the same LF camera and then perhaps even compare the LF 6x8 image with the same image from a Fuji GX680 or similar with other MF. Would it surprise many if the later would surpass anything from any LF and still compare well enough with the 80mpx DB?  Regards, Theodoros. www.fotometria.gr

[BobDavid]

Actually, I just posted an article that uses pictures originally shot in 1976 with a second-hand Calumet 4 X 5 view camera and a 90mm Schneider Super Angulon lens. I photographed the negatives (Plus-X) using a 39 MP multi-shot camera and processed the files in Photoshop. The digital prints look much better than the original silver prints. This might have to do with my Photoshop and digital printing skills being better than my darkroom skills were. Here's the article. If you don't want to read it, scroll down the page and look at the last two pictures. http://goo.gl/pF1JU

[fotometria gr]

I remember in 1998 the discussions that started about digital never being able to get to the quality of 35mm. Look where we are now...

Anyway, IMO, the people that use 8x10 use it for different reasons than the reasons why people use 80Mpx backs in most cases. At least the 8x10 (or bigger) shooters that I know. Most people I know that use 8x10 would not swap it for digital for whatever reason. I am glad there are still people using this, I still find a beautiful 8x10 having something magical.

+++1. I really don't see why the OP comparison is between a LF camera when in film and the use of a MF camera when in digital. I believe that comparisons shouldn't exist for the matter of comparison but for useful conclusions, I mean it would be more useful if somebody would shoot the Leaf on an older Mamiya that can also do film (afd, afd2 or 3) and then compare apples with apples, its apples with oranges now isn't it? OTOH the comparison can be done with both on a view camera, but a view camera is used in most cases when the movements are needed not for extreme resolution. At least I don't know anybody that bought a high res DB to primarily use it on LF! I know people that got a multi-shot back to shoot in both LF and MF but why would somebody use an 80 or 60 mpx back on a view camera? Regards, Theodoros. www.fotometria.gr
  P.S. I have compared my 22mpx DB with (very well scanned) negative film on my Contax and results are comparable, only that film will print larger, has insignificantly less resolution at equal size prints, better highlight DR and ....much slower workflow and much more hassle. 

[fotometria gr]

Actually, I just posted an article that uses pictures originally shot in 1976 with a second-hand Calumet 4 X 5 view camera and a 90mm Schneider Super Angulon lens. I photographed the negatives (Plus-X) using a 39 MP multi-shot camera and processed the files in Photoshop. The digital prints look much better than the original silver prints. This might have to do with my Photoshop and digital printing skills being better than my darkroom skills were. Here's the article. If you don't want to read it, scroll down the page and look at the last two pictures. http://goo.gl/pF1JU

Actually I did the same but with MF film in micro-step (16x) mode with my 22mpx back and it did no difference than scanning it (perhaps scanning was better). It must have to do with yours being sheet film, they are really a bagger to scan..., perhaps we are all better if we .....shoot it (bang bang) instead!  Regards, Theodoros. www.fotometria.gr