Comparison of home digital workflows?

[Michael Heald]

Hello! I'm curious. Currently, I shoot with a Tachihara 4x5 with a 90mm Superangulon and 210mm Sironar N using TMax 400 and developed with a unicolor drum with TMax RS 1:7 at roomt temperature. I can push/pull about 1 1/3 stops with this set-up, and my nominal tested speed is ISO 500.

I scan the negatives with an Epson 4990 and print 8x10 with MIS inks on an Epson R220 with Kirkland paper.

How would these prints compare to the same scene if it were captured with one of the new 12 MP prosumer DSLRs that have come out using a good lens of equivelent focal length to what I shoot 4x5 with?

Would the difference in quality be obvious at this size, or would it become apparent with magnification, say a 4x loupe? Best regards.

Michael A. Heald

[rippo]

hey michael.

quality is a slippery concept! and, because this is the LF forum, of course the LF camera is going to be superior in every way!

but seriously...

since you've mentioned a loupe, i assume you're defining 'quality' as sharpness and resolution. which is only a small part of the equation.

if i scan my 4x5 neg at 1200 dpi, i get a 28.8 mp image according to my calculations (6000 pixels x 4800 pixels). 1200 dpi is pretty leisurely for my scanner...i could go to 4800 dpi, although i don't know if yields anything useful over 2400 dpi (epson 4990).

so, case closed. 4x5 is still better.

well...maybe not. there are so many variables! how finely your lens can resolve, how much detail your film can record, effects of grain (or noise, on the digital cameras). at some point, with enough pixels and/or film, your lens becomes the defining factor in resolution, and your recording medium ceases to record additional useful data as it gets bigger.

how flat is your film? how clean is your sensor? can you focus your 4x5 as well as the auto-focus can on the dSLR? all of these things can affect image quality. but, done properly, your 4x5 will still smoke a prosumer dSLR.

i shoot a Fuji dSLR, so the megapixel count is sort of vague with the crazy sensor it has. if i need a big print, i'll shoot LF, no question. but even if i had a hasselblad with a digital back, i'd still choose to shoot LF for some things, for these reasons:

• film has a wider dynamic range.
• i can do tilts, shifts and twirls.
• i can't think in black and white when i know i've got color capability in the camera. (applies to all formats of film)
• i can make homemade lenses and easily test and attach them to an LF camera. harder to do on a dSLR, and riskier.
• LF makes me take my time with pictures. dSLR is great when i'm shooting family and children's portraits for work, but i don't always want to shoot that way.

erm...but to answer your question, 4x5 is still better.

[ljb0904]

So a Canon 5D will give you an image that is 4368 x 2912, which at 300dpi gives you a print that is 9.6"x14.5". That is the best print you can make. Anything bigger, and you have to upsample the image. If you get an optically true 1200dpi scan from 4"x5" you get 4800x6000 which at 300dpi is 16x20".

So the answer to your question is how big do you want to print?

[eric black]

scans and megapixels aside, to me perspective control is very noticeable which eliminates the DSLR from my workflow- too limiting in the control of the final image.

[jimmayor007]

I have tried a few different cameras, like these (http://rss.api.ebay.com/ws/rssapi?FeedName=SearchResults&siteId=0&language=en-US&output=RSS20&sacqy=&catref=C5&sacur=0&sorefinesearch=1&fsop=1&fsoo=1&copagenum=1&from=R14&saobfmts=exsif&coaction=compare&_trksid=m37&sacqyop=ge&saslc=0&floc=1&sabfmts=0&saprclo=&coentrypage=search&saprchi=&saaff=afcj&ftrv=1&ftrt=1&fcl=3&frpp=50&afcj=2212781&nojspr=y&satitle=large+format&afmp=&sacat=625&saslop=1&fss=0), and found what I think is good, others do not. try a few out first before you commit your cash :)

[rippo]

The limiting factor is the scanner (weak link in system is the Epson). Up to 11x14 I would think they are the same, I wouldn't bother with 4x5 film costs, weight. The Epsons are only good to 3-4x enlargement based on what many of us have agreed with. Once you start printing bigger (with pro class scans), there is no comparison. You can print 40 inch with very good quality. After that the 8x10 format has its place. Of course quality differences vary between people. A lot of amateurs have no idea what real quality is, so some will argue till the cows come home that the Epsons are good for 10-28x enlargement. I say try them in the stores, bring home the disc and test it yourself.

so that i might move up from the ranks of amateurs, can you point me to where this has been discussed and/or tested? thanks.

[Donald Miller]

scans and megapixels aside, to me perspective control is very noticeable which eliminates the DSLR from my workflow- too limiting in the control of the final image.

You can correct perspective in several different programs...PS being only one of them. So the LF superiority in this area is non existant.

[Donald Miller]

Hello! I'm curious. Currently, I shoot with a Tachihara 4x5 with a 90mm Superangulon and 210mm Sironar N using TMax 400 and developed with a unicolor drum with TMax RS 1:7 at roomt temperature. I can push/pull about 1 1/3 stops with this set-up, and my nominal tested speed is ISO 500.

I scan the negatives with an Epson 4990 and print 8x10 with MIS inks on an Epson R220 with Kirkland paper.

How would these prints compare to the same scene if it were captured with one of the new 12 MP prosumer DSLRs that have come out using a good lens of equivelent focal length to what I shoot 4x5 with?

Would the difference in quality be obvious at this size, or would it become apparent with magnification, say a 4x loupe? Best regards.

Michael A. Heald


I shoot with a 5D in addition to 5X7. I have made what to me are very nice prints up to 13X19 with the digital. One needs to understand that a digital print, other than with the Harman glossy paper is going to be different...to me a digital print on Hannemuhle matt paper has a better "feel" than a FB glossy print...but they are certainly different.

It depends on personal tastes.

[Michael Heald]

Hello! Thank you for the comments.

The reason I asked this question is that I'm finding that about 80% of the images I shoot with my 4x5 B&W film, I don't use movements (except, perhaps, to tweak the front to adjust the plane of focus for my landscape shots). I don't plan on enlarging these past 8x10. Most of these are my neighborhood project - what in my immediate neighborhood is beautiful that I might have otherwise overlooked because I walk past it everyday.

I find that I take a lot of these shots to see if the final shot matches my visualization. Proofing in a way. The ones that do, I'll mat and hang on the wall (about 5 at a time), and rotate the images periodically as I have new ones that I like.

For this type of shooting, 4x5 seems overkill. I could "proof" my visualization with a 12 mp DSLR, printing to 8x10 (if the R220 prints are equal to what I can get with 4990 scanned 4x5, since I like to wall hang and rotate the ones that I am pleased with as I get new shots I am pleased with), and shoot 4x5 when I have scenes that I think might warrant more enlarging. However, if 12mp DSLR images printed with R220/MIS inks is inferior, I wouldn't bother

Of course, trips designed to take photos are different creatures. In that case, I know that the images should be designed for further enlarging.

Thank you for your thoughts. Best regards.

Michael A. Heald

[ljb0904]

"""You can correct perspective in several different programs...PS being only one of them. So the LF superiority in this area is non existant."""

Donald, I don't mean to be rude, but you know that statement is not true. I'm sure you know that it's better to do as much perspective control before image capture as possible. LF allows you to do that. The best way to correct perspective is with movements, period. Perspective algorithms are simply a way of getting around not having a tilt/shift lens.

[Donald Miller]

"""You can correct perspective in several different programs...PS being only one of them. So the LF superiority in this area is non existant."""

Donald, I don't mean to be rude, but you know that statement is not true. I'm sure you know that it's better to do as much perspective control before image capture as possible. LF allows you to do that. The best way to correct perspective is with movements, period. Perspective algorithms are simply a way of getting around not having a tilt/shift lens.

I respectfully disagree with you. I have shot LF for over twenty years and until recently thought that was the only format to allow perspective control. I have found that is no longer true. My prints (LF and digital) are indistinguishable so far as this matter.

[Brian Ellis]

I don't know anything about the new prosumer 12 mp cameras and the lenses that go with them. But with the Canon 5D (12.8 mps) and Canon L lenses, and ignoring camera movements, you won't see a significant difference with an 8x10 print. You likely wouldn't see a significant difference even with a drum scan at that print size. When using my digital camera the only LF movement I occasionally miss is the use of tilt to change the plane of focus but not very often. The effects of tilt for other purposes (e.g. keeping parallel lines parallel) as well as the other camera movements can be pretty well duplicated in Photoshop. The only thing I really miss about LF photography when I'm using my digital camera is the pleasure of the process. If you don't enjoy the process then at your print sizes there probably isn't a good reason to be using 4x5 if you can afford a good quality DSLR and lenses.

[Michael Heald]

I thought LF lenses had lower depth of field because of their longer focal length to give an equivalent image? That is, a 150mm lens for 4x5 has a shallower depth of field than a 50mm lens on 35mm? Best regards.

Michael A. Heald

[D. Bryant]

You can correct perspective in several different programs...PS being only one of them. So the LF superiority in this area is non existant.

Sorry Donald, that's not exactly correct. Perspective correction in Photoshop requires the corrected image to be cropped. Perspective correction with a LF camera obviously does not, thus is the superior method, IMO.

Don Bryant

[Marko]

Sorry Donald, that's not exactly correct. Perspective correction in Photoshop requires the corrected image to be cropped. Perspective correction with a LF camera obviously does not, thus is the superior method, IMO.

Don Bryant

On the other hand, a view camera (doesn't really have to be LF either) requires a tripod and much more time on location. With a DSLR, you can easily shoot handheld and a competent photographer will know in advance that perspective correction will require cropping so he will shoot wider.

A camera that stays at home/in the car/in the bag takes no pictures, so the superior method is the one that is available at the moment and which can be used under the circumstances.

[Frank Petronio]

It is superior to apply your perspective controls in camera, but in practice, with a good scan and Photoshop skills, you can adjust perspective digitally just fine. Especially at normal print sizes. And even the best architectural and commercial photos are often digitally tweaked that last 1-2% amount of adjustment because you can't really get it "perfect" on a ground glass like you can with a grid overlay on your digital file. The smart way is to do the gross corrections in camera and get as close as possible, but you know you're going to have to do something digitally anyway, so don't have a cow over doing it.

As for the DSLR versus 4x5 at modest print sizes, while the pixel counts maybe comparable, and you can sharpen the heck out of the digital files so they look great, you still have a greater sense of 3-D depth and a difference in the tonal response, esepcially to the highlights, with film. Most of my prints are in the 8x10/11x14 range and I don't really care about the ultimate resolution that much. And I am skilled at dslr raw conversions. But I can distinguish my 4x5 film shots from my digitals onscreen, and even more so on paper.

Right now I use 4x5 Tri-X, consistently normal lab processed, scanned on a lowly Epson 4990 and printed on an Epson 2200 using the Harrington Quadtone RIP with Ilford Smooth Pearl Paper. Works great at moderate sizes, this is the best workflow yet, the only improvements would be going bigger - Scitex scans and bigger inkjets.

[Donald Miller]

Sorry Donald, that's not exactly correct. Perspective correction in Photoshop requires the corrected image to be cropped. Perspective correction with a LF camera obviously does not, thus is the superior method, IMO.

Don Bryant

If you will look back to what I said earlier, Don, I said that there were several programs that would correct perspective in post processing. Shiftn is another such program and it does not require any cropping of the image. Now I am not going to get into a digital vs film p***ing match. But I do stand by what I said earlier.

I don't shoot 4X5 film any longer since I really could not justify that as a film choice for me. I shoot 5X7 film (when I use film) instead.

[D. Bryant]

It is superior to apply your perspective controls in camera, but in practice, with a good scan and Photoshop skills, you can adjust perspective digitally just fine.


This can be done in the wet darkroom if one is using an enlarger, though not as easily.



As for the DSLR versus 4x5 at modest print sizes, while the pixel counts maybe comparable, and you can sharpen the heck out of the digital files so they look great, you still have a greater sense of 3-D depth and a difference in the tonal response, esepcially to the highlights, with film. .... But I can distinguish my 4x5 film shots from my digitals onscreen, and even more so on paper.


Exactly my experience.

Don Bryant

[Marko]

This can be done in the wet darkroom if one is using an enlarger, though not as easily.

Most tasks can be accomplished using a variety of methods, but the topic of this thread does not include traditional darkroom.

The OP stated very clearly that he scans his film. The question was about comparing various digital workflows. Specifically, comparing digital processing of scanned film vs. post-processing of straight digital capture.

Please, let's not have yet another topic degenerate into traditional vs. digital.

[D. Bryant]

On the other hand, a view camera (doesn't really have to be LF either) requires a tripod and much more time on location. With a DSLR, you can easily shoot handheld and a competent photographer will know in advance that perspective correction will require cropping so he will shoot wider.

A camera that stays at home/in the car/in the bag takes no pictures, so the superior method is the one that is available at the moment and which can be used under the circumstances.

I always prefer shooting on a tripod regardless of the camera except for point and shoot bodies.

Shooting wider isn't always a viable answer and will often introduce other problems to the composition. Also digital transformations to correct perspective can cause other distortions that may not be pleasing.

Any competent architectural photographer will prefer a view camera over a camera that doesn't support adjustments.

I use my DSLR differently than my view camera and to suggest that because one is more portable than the other means that it is a more versatile camera doesn't make sense to me. They are different tools for different jobs and different cameras requiring a different mindset and approach. If portability is the main concern pickup your Leica (digital or film) and leave the SLRs at home.

But we are drifting off topic from the OPs question.


Don Bryant

[D. Bryant]

If you will look back to what I said earlier, Don, I said that there were several programs that would correct perspective in post processing. Shiftn is another such program and it does not require any cropping of the image. Now I am not going to get into a digital vs film p***ing match. But I do stand by what I said earlier.

I don't shoot 4X5 film any longer since I really could not justify that as a film choice for me. I shoot 5X7 film (when I use film) instead.

My comment was not intended to be a digital vs film pissing match. My photographic workflow is highly integrated with digital methods but experience has shown me that perspective correction in PS vs in camera doesn't always provide equivalent results although it is very handy. So I still stand by what I said earlier, I've been using view cameras since about 1972, and I've been using PS for about 10 years or so, so I base my answer on my experiences.

Don Bryant

[sanking]

If you will look back to what I said earlier, Don, I said that there were several programs that would correct perspective in post processing. Shiftn is another such program and it does not require any cropping of the image. Now I am not going to get into a digital vs film p***ing match. But I do stand by what I said earlier.

I don't shoot 4X5 film any longer since I really could not justify that as a film choice for me. I shoot 5X7 film (when I use film) instead.

This discussion has absolutely nothing to do with digital versus film capture. I have been using perspective control in Photoshop with scanned LF and MF negatives for at least 5 years and am very familiar with the pros and cons versus in-camera controls. I agree absolutely with what Don Bryant and Frank Petronio have said, i.e. the best place to do perspective controls, *if you can*, is in the camera. Or, do as much as you can in the camera and then complete the job in Photoshop. The fact of the matter is that there are limitations to what you can do both in the camera and in Photoshop. If you can combine the two, so much the better.

If your camera has no perspective controls you will need to carefully plan the composition because correcting in Photoshop is going to pull parts of the scene out of the frame. I have not used Shiftn. Perhaps it resolves this problem? Another issues is that if the controls used are very extensive there will be a loss of resolution in the the image.

On the other hand, for prints up to 8X10 in size where few perspective controls are going to be used (as per the OP) the DSLR would probably give results just as good as 4X5, so in this case I would agree with him that a view camera is overkill for the job.


Sandy King

[Donald Miller]

Sandy,

Apparently you failed to read the original posters questions. The questions, as I understood them, was about a comparison of two different methods. I stated my experience with the two methods. Apparently by virtue of some of the comments that have subsequently been tendered, I do not have the priviledge of stating a personal opinion while others do. Do you have an idea why that is the case? I did not say that digital was preferable to LF. I clearly stated that for my purposes and consistent with my standards what my experience has been. It seems that some think that any disagreement with their opinion is a flagrant statement about their manhood. That in my estimation is a very immature and unreasonable position to take.

I did not say that in camera corrections were inferior, superior or a wash...I think that determination depends on a lot of things.

I have taken the liberty of posting a couple of images that illustrate my results with perspective control using the software that I mentioned. At the level of the original poster's stated output size, I find that the results are acceptable to me. Simply put that is all I meant to say and think that I said.

[Michael Heald]

Frank, you stated that you feel that 4x5 when printed to 8x10 after scanning on an Epson 4990 are superior to a DSLR at 12 MP. You stated that you felt they had more depth.

Are you using color or B&W? Would this make a difference on how you find the prints for each work flow?

By depth, it seems to me that you find you keep more deetail in the highlights with scanned 4x5 compared to DSLR. Is this a fair way to summarize what you've stated? If so, would this be a way of saying that you find the dynamic range of the 4x5 scanned negatives compared to DSLR 12 MP is noticeably different when printed to 8x10? Would the kind of printer, then, be important? I'm wondering if I would see the difference on my R220 while I would be able to see it on a better printer. Best regards.

Michael A. Heald

[sanking]

Sandy,

Apparently you failed to read the original posters questions. The questions, as I understood them, was about a comparison of two different methods. I stated my experience with the two methods. Apparently by virtue of some of the comments that have subsequently been tendered, I do not have the priviledge of stating a personal opinion while others do. Do you have an idea why that is the case? I did not say that digital was preferable to LF. I clearly stated that for my purposes and consistent with my standards what my experience has been. It seems that some think that any disagreement with their opinion is a flagrant statement about their manhood. That in my estimation is a very immature and unreasonable position to take.

I did not say that in camera corrections were inferior, superior or a wash...I think that determination depends on a lot of things.

I have taken the liberty of posting a couple of images that illustrate my results with perspective control using the software that I mentioned. At the level of the original poster's stated output size, I find that the results are acceptable to me. Simply put that is all I meant to say and think that I said.


Donald,

Please make some effort to not deliberately insult. I know exactly what the original poster asked and I addressed his question with the comment that for his project the view camera is overkill and that he would be just as well off with the DSLR for prints up to 8X10 size.

I also know exactly what you said. You said that for perspective control the view camera is not superior to digital controls. You wrote, “You can correct perspective in several different programs...PS being only one of them. So the LF superiority in this area is non existant.”

So that is y our opinion. My opinion is that the use of perspective controls with a view camera is much superior to digital perspective control, though digital control is clearly a wonderful tool.

Maybe I am wrong in this. It would be nice to hear the perspective (sorry for the pun) of someone like Kirk Gittings who does a lot of architectural work professionally and uses both DSLR and view camera.

Sandy King

[Kirk Gittings]

I've just been traveling, picking up a Creo flatbed scanner in AZ actually. From Flagstaff to Phoenix and all the way back to Albuquerque in driving rains. I have never seen anything quite like it.

I still think, to avoid interpolation issues, that it is important to do perspective correction in the camera whether it is LF or DSLR for the highest quality work. Small fine tuning corrections are fine, but when you get above stretching the top of the frame more than about 1/4 the frame width you are doing some serious interpolation. Which is why after doing some testing I invested in 24, 35, 45, 90mm shift lenses for the DSLR. IMO to get a usable image out of a 12MP DSLR you have to work very very carefully, cropping and interpolating as little as absolutely little as possible.

[sanking]

I've just been traveling, picking up a Creo flatbed scanner in AZ actually. From Flagstaff to Phoenix and all the way back to Albuquerque in driving rains. I have never seen anything quite like it.

I still think, to avoid interpolation issues, that it is important to do perspective correction in the camera whether it is LF or DSLR for the highest quality work. Small fine tuning corrections are fine, but when you get above stretching the top of the frame more than about 1/4 the frame width you are doing some serious interpolation. Which is why after doing some testing I invested in 24, 35, 45, 90mm shift lenses for the DSLR. IMO to get a usable image out of a 12MP DSLR you have to work very very carefully, cropping and interpolating as little as absolutely little as possible.

Congratulations on the Creo flatbed. Which one did you get? Though all of them provide a huge leap in improvement over the consumer flatbeds.

Sandy

[Kirk Gittings]

I got a steal on a complete, mint condition, Eversmart Jazz +, which has enough resolution for me since I don't print above 16x20 usually. These will not run on anything past OX9 right?

[sanking]

I got a steal on a complete, mint condition, Eversmart Jazz +, which has enough resolution for me since I don't print above 16x20 usually. These will not run on anything past OX9 right?

Kirk,

What driver software did you get with the Jazz +? Hopefully it will allow high bit scans. If not I can suggest a good work around procedure that is almost as good.

For questions about EverSmart scanners try this site on Yahoo. http://tech.groups.yahoo.com/group/ScanHi-End/

One or two people there have great knowledge about EverSmart scanners.

Sandy

[Frank Petronio]

Congrats Kirk! But back to the topic for a second... ;-)

Michael:
In general and under the same circumstances the dslr images are going to favor a very linear response to light, and the highlights will tend to blow out or not be as "separated' as you will get from film, especially B&W and color neg film. On the upside, you tend to get more shadow detail from the dslr under the same circumstances.

Of course you can work around this by underexposing, combining exposures, tweaking the curves in your RAW processing, etc. All is not lost with your dslr, but you do tend to find yourself overall underexposing sometimes in order to hold highlight details.

For a really good fast example of this, go outside and shoot matching, "nice" exposures of passing white puffy clouds with your dSLR and 4x5 w negative film. Put the brightest spot other than the sun on Zone 9 or so. When you compare the 4x5 scan to the daw dslr file, you'll see the clouds have incredible subtle range of micro tones with infinite steps... and the digital file will be either blown out or muddy...

As for the perspective issue, geez wacking 25% of the width is pretty radical, I agree with Kirk, if you have to do that much then it might show some decline quality even at moderate sizes.

One thing that stuck me was when I scanned some of my really old professional, widely published architectural shots from the roaring 80s... things that were acceptably straight and lined up {"in camera") back then really were a few degrees off once I overlaid a Photoshop grid on the image. Being anal retentive I fixed them, but I probably wouldn't have ever worried about it if I never looked for the "error".

[Brian Ellis]

"All is not lost with your dslr, but you do tend to find yourself overall underexposing sometimes in order to hold highlight details."

I never underexpose to hold highlight details. I first try a graduated ND filter. If that doesn't work I make two exposures, one for the highlights the other for the midtones and shadows, and merge them. If fail to do either and later discover that I should have then I make duplicate "exposures" in Photoshop, one for the highlights, one for the midtones and shadows, and merge them. Is there a reason why it's preferable to underexpose to hold highlight details rather than doing any or all of these three things? Please note - this is a question, not an argument. Also, it isn't a film vs digital question. My solutions to excessive contrast work equally well (or poorly) regardless of whether one starts with a scan or a photograph from a digital camera.

[Frank Petronio]

Well it's not underexposure unless you have some higher authority telling you it's underexposed ;-)

And while you can do multiple exposures and ND grads and all kinds of other tricks, I'm just talking straight shooting -- you try to avoid blowing the highlights with a dslr even more than with film...

[audioexcels]

You can correct perspective in several different programs...PS being only one of them. So the LF superiority in this area is non existant.

So you can shift/tilt/rise with a DSLR file?

I can make images like this with a Canon 5D or better?

http://www.ebonycamera.com/gal.contrib/freemanp/index.html

[Marko]

So you can shift/tilt/rise with a DSLR file?

I can make images like this with a Canon 5D or better?

http://www.ebonycamera.com/gal.contrib/freemanp/index.html

Those are all fine examples of controlling plane of focus with a view camera, but the statement you responded to was about controlling perspective.

The way I understand it, those are two very easily differentiated, distinct concepts. ;)

[Marko]

I can make images like this with a Canon 5D or better?

If you can do it with a Canon 1Ds III, you can do it with Digital Rebel just as well, save for a few very specific tasks.

But only you know whether you can really do it, it depends on you, not the camera.

[PViapiano]

You can correct perspective in several different programs...PS being only one of them. So the LF superiority in this area is non existant.

I've done perspective control in PS and it works to a certain extent, IMO, however you lose a portion of your image depending on how much you are correcting, and you must be very careful about pincushioning, bloating, and other distortions that inevitably rear their heads.

Regardless of what anyone says, it's NOT the same as or as elegant as doing it right the first time and using the correct tool for the job, which in this case is a LF camera or other hardware-based solution.

Note: I just finished reading all the posts re this above, and see that Kirk, Frank, Sandy agree. I'm not bashing the digital method, just expressing the caveat to look out for other distortions that may arise. My LF architecture shots are superior to my PS-adjusted ones because of thisphenomenon. Yes, small adjustments can easily be made without detection but larger ones can be tricky. I'd be interested in checking out the Shiftn program...

Here's a link to one of my PS-adjusted images:

http://home.pacbell.net/viapiano/gehry/Web_Pano_2.html

It's a PanoTool created image with PS-adjusted perspective. It's not technically accurate but I like it dramatically...

Here's a link to a "pure" LF architectural image:

http://www.flickr.com/photos/viapiano/1602479072/