Just thought of another question that I am considering.
What do you think of the 5D comparing to 35mm film? Is film (Provia?) still better than the 5D when it comes to color/details? Has anyone compare when they both being blown up to 30 inch prints?
The reason I am thinking of this is (as much as I love the look of film), the post production is starting to be costly, especially now I am shooting more and more. I've recently bought a Epson V700 scanner and am not happy with the result. Maybe I got spoiled by drum scans or other high end scans. But part of me also think I can always send my slides back to China (where I am from) and get them drum scan for less that 8 bucks US each (but of course I'll have to wait for them).
In your value opinion, should I stay with the lovely film? Or should I join the digital gangs?
Another one from Chun Ming
Posted by Kalin Wilson on July 13, 2007 at 10:53 PM
LIke some of the answers in your other thread or the thread about MF digital backs mentioned, 5D images can rival MF scans. There is an article in Outdoor Photographer that compares digital to film. Also there are lots of 5D reviews and comparisons online. If you're shooting a lot or plan to, the savings of digital and the flexibility of RAW capture will be worth the initial cost quickly. 5D beats 35mm scans.
Posted by Grant Collier on July 13, 2007 at 11:43 PM
Last year, I did a comparison between shots in Antelope Canyon taken with the 5D and taken with 35mm Velvia100F. I cropped out an equal percentage of each shot and sharpened them the same amount. The above picture is the 5D shot. The below picture is the 35mm film shot. It's not a perfect comparison as these aren't identical scenes. But there's really no need to do a perfect comparison, as I think the 5D obviously produces much greater detail.
Posted by Grant Collier on July 13, 2007 at 11:44 PM
I think a majority of 5D owners who have shot 35mm film in the past, will tell you that 5D image quality is as good as or better than 35mm film (and possibly MF film).
My personal experience... I have shot about 8,000 images with the original Canon Digital Rebel and more recently about 5,000 images with my Canon 30D. My film inventory includes over 15,000 35mm slides, 500 645 transparencies, and 1,000 4x5 transparencies. I have made comparisons using 16" x 24" prints, which I create on my Epson 4000. In comparing prints from 35mm drum scanned film and 30D images, I prefer the image quality of the 30D. In a similar comparison between drum scanned 645 film and the Canon 30D, I found the 2 to be very, very close. But the 4x5 drum scanned film print is clearly superior (IMHO).
Note that I say my personal preference. Print evaluation is subjective - a personal experience. And I won't fault or criticize anyone who comes to a different conclusion than I do. Furthermore, applying mathematical formulas to compare the various formats and determine which has the best image quality is utterly ridiculous! We view prints and form opinions with our eyes and brains, not with microscopes and measuring devices. BTW, my degree is in physics and mathematics, so I know a bit about the technical side of things.
I own the Nikon equivalent of the 5D (D2X). I don't think color is an issue - a raw capture can be made to assume the qualities of just about any film type with little effort. Regading details, ditto what Grant and Scott said. IMO, the 5D, D2X and probably even some lesser bodies will hold up to 35mm easily. I've had drum-scans done of 35mm Provia slides, and those are much closer to the digital images. And while that's nice to know for my existing 35mm library, the cost of drum scanning makes that approach unattractive moving forward (and regardless of cost-savings, I'm not sure I'd want to ship my slides back and forth to China routinely). I can't think of one quality-related (capture or print) issue where 35mm film offers any significant advantage.
The cost savings of digital is widely varied, imo. I think the savings is proportional to how much you shoot. Wanting/needing to have several forms of backup for digital images will add cost. Whereas I only scanned and stored ~5 % of my film shots, I tend to shoot and store more digital shots. I delete the ones that aren't keepers, but I also tend to keep 2 or 3 files for one image for digital blending, and of course, until you can delete the non-keepers, you need to store them somewhere, requiring sufficient storage space (and backup space). For me, digital hasn't saved me much money over film, but it hasn't cost me more either. Upgrading to new bodies regularly is a different issue, though....
For me, the costs savings of digital are significant. Each 35mm film shot costs about 33 cents for film & developing. For each digital shot from the 5D, if I save it on 3 separate drives, it takes up about 39MB of space. You can actually now buy external hard drives for almost as low as 25 cents per gigabyte (and that number will keep dropping). Since each photo takes up 1/25 GB, it costs about 1 cent per photo to store them 3 times. For medium format film, which is more comparable in quality to the 5D, the cost per shot is around $1. And this comes from someone who was rather hesitant to go digital.
Hello all, I'm not really trying to make this a big film versus digital debate, but that is essentially the nature of the post. I'll be up front and say I shoot and scan in film. As noted in Chun's ealier question about MF, I did state digital was a good option, depending upon what he is looking for. My response here is almost more of a question than it is an answer. I'm not getting the lack of sharpness or grain artifacts people are getting when they scan film. I've done drum scans and I was disappointed. My Nikon 9000 ED blows away two separate drum scans I had made in overall quality. I feel I'm getting near pixel for pixel resolution off of a 4000 dpi scan. It looks a little soft when viewed at 100%, but I've downloaded 1DS Mark 2 images off the net and they also aren't tack sharp when viewed at 100%. Nothing againt Grant's photos above, but the 5D image, while looking better than his scanned film, is not super sharp at every pixel. I only note this because I think the image I've posted with the MF image is comparibly sharp.
I got my hands on an 8mp Canon 20D and used my 24-105mm IS L lens and compared it with my 35mm Canon and Pentax 67ii. I know this isn't the same as the comparison requested above, but 12.8 mp equates to only an increase of 26% resolution in each direction over an 8mp camera. Obviously, 6x7 will have about 2x the lateral resolution of 35mm, but many people seem to feel that Canon's full frame digital cameras equate to MF. I know the 8mp Digital Rebel is not a 5D, but the difference in potential resolution is not much greater.
It seems to me if you get a good scan of film, pixels are pixels, and if you can address the grain issue, what you can do with a digitally captured image and what you can do with scanned film. In theory, the quality of the pixels would be better with the scanned film as transparancy scanners get full color data for every pixel and have a larger range of greyscale values than digital, which interpolates color data.
The posted image above has MF viewed at 100% at the left, scanned 35mm uprezed in the middle and 8mp digital uprezed on the right. I then applied the same sharpening techniques to all. I think MF is the clear winner, 35mm still a winner over digital and then digital last. I would guess the 1DS Mark 2 is a better equivalent to 35mm. Am I missing something? Sometimes I feel like I got an advanced prototype scanner from Nikon.
If you really want to delve into this, you can do a search for "film vs. digital, comparison." I think about the only thing you can really conclude is that the 5D will pretty much always be better in terms of resoultion than the 35mm film (night shots excluded). And 4x5 film will always be better than the best DSLR for larger prints. As for closer comparisons like 5D vs. 645, the results can vary widely, as it depends on many factors like the lenses used, the scanner used, the uprez method, the film used, or even the scene you're shooting. And there's a lot of factors besides resolution which can go in favor of film or digital. So it really just comes down to personal preference rather than one being "better" than the other.
If you do want even better resolution with the digital, you can always do stitched images, which can rival 4x5 or even 8x10 depending on how many images you stitch. It will cost a lot less, but it certainly takes more effort to take all the different shots and then stitch them together.
I should add that shooting film can take up big hard drive space as well. 8 bit images of 35mm at 4,000 ppi takes about 54mb of space, which is somewhat of a reduced size due to the mounting. 6x7 images are about 280mb apiece, less alone a scan of LF which will get near a gig. As such, storing scanned film can get expensive as well.
I think I'll borrow a 5D from a friend and try it out. Seems like it will save me quite a bit down the road, both in cash and frustration! Thank you all for your comments again!!
So I am curious, what's the biggest PRINT that you've successfully printed out from the 5D before it gets unsatisfactory? By that, meaning printing out from a single frame. Not stitched.
Chun Ming
Posted by Scott Bacon on July 15, 2007 at 10:04 AM
I've seen some really nice 20" x 30" prints that RichieV has created on his Epson printer from his 5D. Hopefully, he'll see this and chime in. If not, drop him a note. I'm sure he'd be happy to share his opinions...
My opinions on this topic are as follows. But before I pipe in, my basic background is 35mm then went to 6x45 then 6x7 then 4x5. Now I am currently 4x5 and digital with the 5D. I still do almost all of my b/w with the 4x5 and hand process my film with hand crafted and calibrated developers --> drum scan --> and print. I am still not a fan of converting any color images over to B/W over shooting them directly to B/W film. That said: all of the following are my opinions only and only represent my personal experiences with this gear. Also, remember up until 3-4 years ago I had a fully functional LF based darkroom and considered any digital printing total crap and blasphemy.
- my 5D blows away my 6x4.5 Bronica (yes that is drum scanned images)
- my 5D cannot come even remotely close to my drum scanned 4x5's. LF is a different world as most of the LF shooters on the board will attest to. Awesome for some not a good fit for others.
- I have made up to 33x43 from my 5D. (SINGLE FRAME) AND it was a handheld ISO 800, 24-70zoom lens shot. I think the camera was on Program. It really looks awesome. The image required a little noise ninja/gen fractals and then pixel genius work.
- 16x24 images almost always look terrfic from it (the 5D)
- Do not underestimate what others are saying about DOF with a 6x7. It is a tricky world to get around. On my Mamiya 7 while I loved the camera it was always really tough to do anything with near-far DOF.
- Personally the combination of ease of use, RAW cabailities, advanced workflows, and ability to make terrific prints make me like my 5D a lot better than any of the MF gear I have or have had.
- As far as the direct comparison between 6x7 and 5D. Here is my 2cents. Personally I like working with the 5D image better. I scan with Howtek D4000 scanners and know what I am doing with them to get best output possible. I believe that the image is marginally sharper overall and has more raw data to work with. But (and this is really hard for me to say) the images I am getting out of the 5D are better than my Mamiya 7 and 1000% better than the Bronica 6x4.5.
But as almost anybody except for the most technical weenies on this board would agree the final opinion is what does the print look like. If any of you are interested in seeing some examples of big prints from the 5D, 6x7 and 4x5 shoot me an email.
I agree that DSLR has surpassed 35mm and in some cases MF film. I've been scanning and printing images of my own and for clients for many years, from all kinds of film and digital originals, and modern digital captures are superior to film in real-world usage from editing to print.
I've made really nice 20x30 inch prints from 6.3 MP 10D captures and the 8.6 MP from the 30D will easily go to 40x60 inches if properly processed. The upper limit of printed images from the 5D are truly wall-sized.
The only "catch" is that you need to know how to properly expose and process raw captures to provide the best quality. This is quite different from film, where the onus is on the lab and then the scanner operator to work up a decent file. With digital, you need to know what you're doing on the computer to get the best results.
A 10MP digital capture will provide more detail than an 80 MB film scan.
I am a big fan of my 5D captures but need some clarification.
I am trying to figure out how you get more info out of a 10MP digital capture than an 80MB scan? Are you talking about 35mm film or bigger or what? That isn't what i am seeing out of 4x5 or 6x7 drum scans.
RV
Posted by Nat Coalson on July 19, 2007 at 11:29 AM
Rich - yes, I am talking about 35 mm film which doesn't have the inherent resolution of a 10MP capture to begin with, not to mention film grain that is massively evident in even the best scans.
A 4x5 (or some MF) when scanned properly will still undoubtedly still yield more detail than a 10MP capture.
But a 15MP capture starts to approach MF film, and medium format digital backs (20MP and up) are vastly superior even to 4x5.
The problem with comparing film scans to digitally captured megapixels is that they are not the same animal. You need scans with huge file sizes to match the quality of digital captures with smaller file sizes. This is due to the analog nature of film scanners (although digital cameras use analog/digital converters too; it's done at the time of capture.)
The main advantage of digital capture is that is is super-clean compared to film, and on many cameras (when capturing RAW) the dynamic range also matches or surpasses that of film. And digital captures fare MUCH better when any interpolation is necessary, especially upsampling.
Posted by Les Barstow on July 19, 2007 at 11:31 AM
Having just had some 20"x30" prints made up of both, I'd have to go with my D2X digital images over the 35mm Velvia 100F.
I used to drum-scan my slides, but now I use a Minolta DiMAGE 5400; the scanner certainly gives me more data - ~100MB in 8-bit for a 5400dpi scan - but the 36MB worth of data that I get from the D2X (8-bit, not the full 12...) has purer data that's much easier to work with.
I have to say that the 35mm image came out quite nicely after a bunch of work with Noise Ninja, selective edge sharpening, and some extra noise reduction in the sky. But that was a lot of work on a slow computer. One of the two D2X images had definite noise issues (not much light, dark subject matter...); I'd say it came out about the same quality as the 35mm slide after noise reduction and other processing. The other digital shot was, IMHO, much better on the detail than the other two, and it only required 20% of the work effort.
Color-wise, digital images are what you make of them. By default, the D2X acts a lot like Provia - not much punch but very accurate. A bit of tweaking turns it into a "Velvia" style print that looks better on a wall. Using the manufacturer's recommended RAW processing, you get ICC profiled output that's dead-on accurate.
PS - thanks to Nat for the 20x30" prints. Everyone who's seen them thinks they're spectacular. Now let's hope I take images worthwhile enough to draw in customers...
Discussion
16 repliesWhat do you think of the 5D comparing to 35mm film? Is film (Provia?) still better than the 5D when it comes to color/details? Has anyone compare when they both being blown up to 30 inch prints?
The reason I am thinking of this is (as much as I love the look of film), the post production is starting to be costly, especially now I am shooting more and more. I've recently bought a Epson V700 scanner and am not happy with the result. Maybe I got spoiled by drum scans or other high end scans. But part of me also think I can always send my slides back to China (where I am from) and get them drum scan for less that 8 bucks US each (but of course I'll have to wait for them).
In your value opinion, should I stay with the lovely film? Or should I join the digital gangs?
Another one from Chun Ming
My personal experience... I have shot about 8,000 images with the original Canon Digital Rebel and more recently about 5,000 images with my Canon 30D. My film inventory includes over 15,000 35mm slides, 500 645 transparencies, and 1,000 4x5 transparencies. I have made comparisons using 16" x 24" prints, which I create on my Epson 4000. In comparing prints from 35mm drum scanned film and 30D images, I prefer the image quality of the 30D. In a similar comparison between drum scanned 645 film and the Canon 30D, I found the 2 to be very, very close. But the 4x5 drum scanned film print is clearly superior (IMHO).
Note that I say my personal preference. Print evaluation is subjective - a personal experience. And I won't fault or criticize anyone who comes to a different conclusion than I do. Furthermore, applying mathematical formulas to compare the various formats and determine which has the best image quality is utterly ridiculous! We view prints and form opinions with our eyes and brains, not with microscopes and measuring devices. BTW, my degree is in physics and mathematics, so I know a bit about the technical side of things.
Hope this helps...
The cost savings of digital is widely varied, imo. I think the savings is proportional to how much you shoot. Wanting/needing to have several forms of backup for digital images will add cost. Whereas I only scanned and stored ~5 % of my film shots, I tend to shoot and store more digital shots. I delete the ones that aren't keepers, but I also tend to keep 2 or 3 files for one image for digital blending, and of course, until you can delete the non-keepers, you need to store them somewhere, requiring sufficient storage space (and backup space). For me, digital hasn't saved me much money over film, but it hasn't cost me more either. Upgrading to new bodies regularly is a different issue, though....
I got my hands on an 8mp Canon 20D and used my 24-105mm IS L lens and compared it with my 35mm Canon and Pentax 67ii. I know this isn't the same as the comparison requested above, but 12.8 mp equates to only an increase of 26% resolution in each direction over an 8mp camera. Obviously, 6x7 will have about 2x the lateral resolution of 35mm, but many people seem to feel that Canon's full frame digital cameras equate to MF. I know the 8mp Digital Rebel is not a 5D, but the difference in potential resolution is not much greater.
It seems to me if you get a good scan of film, pixels are pixels, and if you can address the grain issue, what you can do with a digitally captured image and what you can do with scanned film. In theory, the quality of the pixels would be better with the scanned film as transparancy scanners get full color data for every pixel and have a larger range of greyscale values than digital, which interpolates color data.
The posted image above has MF viewed at 100% at the left, scanned 35mm uprezed in the middle and 8mp digital uprezed on the right. I then applied the same sharpening techniques to all. I think MF is the clear winner, 35mm still a winner over digital and then digital last. I would guess the 1DS Mark 2 is a better equivalent to 35mm. Am I missing something? Sometimes I feel like I got an advanced prototype scanner from Nikon.
If you do want even better resolution with the digital, you can always do stitched images, which can rival 4x5 or even 8x10 depending on how many images you stitch. It will cost a lot less, but it certainly takes more effort to take all the different shots and then stitch them together.
So I am curious, what's the biggest PRINT that you've successfully printed out from the 5D before it gets unsatisfactory? By that, meaning printing out from a single frame. Not stitched.
Chun Ming
I have barely been home lately.
My opinions on this topic are as follows. But before I pipe in, my basic background is 35mm then went to 6x45 then 6x7 then 4x5. Now I am currently 4x5 and digital with the 5D. I still do almost all of my b/w with the 4x5 and hand process my film with hand crafted and calibrated developers --> drum scan --> and print. I am still not a fan of converting any color images over to B/W over shooting them directly to B/W film. That said: all of the following are my opinions only and only represent my personal experiences with this gear. Also, remember up until 3-4 years ago I had a fully functional LF based darkroom and considered any digital printing total crap and blasphemy.
- my 5D blows away my 6x4.5 Bronica (yes that is drum scanned images)
- my 5D cannot come even remotely close to my drum scanned 4x5's. LF is a different world as most of the LF shooters on the board will attest to. Awesome for some not a good fit for others.
- I have made up to 33x43 from my 5D. (SINGLE FRAME) AND it was a handheld ISO 800, 24-70zoom lens shot. I think the camera was on Program. It really looks awesome. The image required a little noise ninja/gen fractals and then pixel genius work.
- 16x24 images almost always look terrfic from it (the 5D)
- Do not underestimate what others are saying about DOF with a 6x7. It is a tricky world to get around. On my Mamiya 7 while I loved the camera it was always really tough to do anything with near-far DOF.
- Personally the combination of ease of use, RAW cabailities, advanced workflows, and ability to make terrific prints make me like my 5D a lot better than any of the MF gear I have or have had.
- As far as the direct comparison between 6x7 and 5D. Here is my 2cents. Personally I like working with the 5D image better. I scan with Howtek D4000 scanners and know what I am doing with them to get best output possible. I believe that the image is marginally sharper overall and has more raw data to work with. But (and this is really hard for me to say) the images I am getting out of the 5D are better than my Mamiya 7 and 1000% better than the Bronica 6x4.5.
But as almost anybody except for the most technical weenies on this board would agree the final opinion is what does the print look like. If any of you are interested in seeing some examples of big prints from the 5D, 6x7 and 4x5 shoot me an email.
RichieV
PS - I piped in Scott
I've made really nice 20x30 inch prints from 6.3 MP 10D captures and the 8.6 MP from the 30D will easily go to 40x60 inches if properly processed. The upper limit of printed images from the 5D are truly wall-sized.
The only "catch" is that you need to know how to properly expose and process raw captures to provide the best quality. This is quite different from film, where the onus is on the lab and then the scanner operator to work up a decent file. With digital, you need to know what you're doing on the computer to get the best results.
A 10MP digital capture will provide more detail than an 80 MB film scan.
I am a big fan of my 5D captures but need some clarification.
I am trying to figure out how you get more info out of a 10MP digital capture than an 80MB scan? Are you talking about 35mm film or bigger or what? That isn't what i am seeing out of 4x5 or 6x7 drum scans.
RV
A 4x5 (or some MF) when scanned properly will still undoubtedly still yield more detail than a 10MP capture.
But a 15MP capture starts to approach MF film, and medium format digital backs (20MP and up) are vastly superior even to 4x5.
The problem with comparing film scans to digitally captured megapixels is that they are not the same animal. You need scans with huge file sizes to match the quality of digital captures with smaller file sizes. This is due to the analog nature of film scanners (although digital cameras use analog/digital converters too; it's done at the time of capture.)
The main advantage of digital capture is that is is super-clean compared to film, and on many cameras (when capturing RAW) the dynamic range also matches or surpasses that of film. And digital captures fare MUCH better when any interpolation is necessary, especially upsampling.
I used to drum-scan my slides, but now I use a Minolta DiMAGE 5400; the scanner certainly gives me more data - ~100MB in 8-bit for a 5400dpi scan - but the 36MB worth of data that I get from the D2X (8-bit, not the full 12...) has purer data that's much easier to work with.
I have to say that the 35mm image came out quite nicely after a bunch of work with Noise Ninja, selective edge sharpening, and some extra noise reduction in the sky. But that was a lot of work on a slow computer. One of the two D2X images had definite noise issues (not much light, dark subject matter...); I'd say it came out about the same quality as the 35mm slide after noise reduction and other processing. The other digital shot was, IMHO, much better on the detail than the other two, and it only required 20% of the work effort.
Color-wise, digital images are what you make of them. By default, the D2X acts a lot like Provia - not much punch but very accurate. A bit of tweaking turns it into a "Velvia" style print that looks better on a wall. Using the manufacturer's recommended RAW processing, you get ICC profiled output that's dead-on accurate.
PS - thanks to Nat for the 20x30" prints. Everyone who's seen them thinks they're spectacular. Now let's hope I take images worthwhile enough to draw in customers...