Digital vs Film Comparison (Blow-up)

[bobwysiwyg]

Came across this on another site. http://fwd.five.tv/videos/challenge-blow-up-part-3. It depicts an interesting comparison, but raises a question for me about my lack of knowledge about image sensors and their role in final product quality.

The Nikon model used in this comparison had a 12 MP sensor whose dimensions were very similar to 35mm. Yet Nikon has a model that costs much less that also has a 12 MP sensor, but I couldn't find the dimensions of the sensor for that one.

Since both have the same number of pixels, but assuming one was physically smaller, what role would that play in the quality of the final image and any subsequent enlargements?

 

[sanking]

There is a short piece here comparing image quality versus sensor size. Kind of old but the comparison is still valid.

http://photo.net/equipment/digital/sensorsize/

My Canon G9 has a sensor 7.6 X 5.7 mm in size. I have to say that I am pretty amazed at the quality this little camera is capable of in a print sizek up to about 11X15".

Sandy King

Came across this on another site. http://fwd.five.tv/videos/challenge-blow-up-part-3. It depicts an interesting comparison, but raises a question for me about my lack of knowledge about image sensors and their role in final product quality.

The Nikon model used in this comparison had a 12 MP sensor whose dimensions were very similar to 35mm. Yet Nikon has a model that costs much less that also has a 12 MP sensor, but I couldn't find the dimensions of the sensor for that one.

Since both have the same number of pixels, but assuming one was physically smaller, what role would that play in the quality of the final image and any subsequent enlargements?

 

[bobwysiwyg]

Very informative, thanks. It even made sense.

 

[Joe Lipka]

Sandy - The G9 is a remarkable "little camera" with some amazing capabilities. The Canon G9 is used to record the LensWork darkroom tour and interview sessions in the LensWork extended publication. So not only does it make nice still images it's a heck of a video recorder too.

 

[sanking]

It is quite a remarkable camera. And the ability to shoot RAW, as well as bracket automatically exposures for HDR photography should also be mentioned as very positive features.

Sandy King

Sandy - The G9 is a remarkable "little camera" with some amazing capabilities. The Canon G9 is used to record the LensWork darkroom tour and interview sessions in the LensWork extended publication. So not only does it make nice still images it's a heck of a video recorder too.

 

[Loris Medici]

HDR bracketing

...bracket automatically exposures for HDR photography...

Yes that's a neat feature. They added this feature to A700's last firmware update (3 frames, -2, 0, +2 or 0, -2, +2). It's very welcomed! In normal shooting I never feel the need of HDR, since A700 does very good in selectively lightening the shadows with its Dynamic Range Optimization feature (it does wonders!), but I plan to do some HDR photography "exploiting" the technique...

Regards,
Loris.

 

[sanking]

Loris,

Attached is a shot I made with the Canon G9 last winter in Oaxaca, Mexico. The contrast range of this scene was very great and it could not have been made in digital without use of HDR.

Sandy

Yes that's a neat feature. They added this feature to A700's last firmware update (3 frames, -2, 0, +2 or 0, -2, +2). It's very welcomed! In normal shooting I never feel the need of HDR, since A700 does very good in selectively lightening the shadows with its Dynamic Range Optimization feature (it does wonders!), but I plan to do some HDR photography "exploiting" the technique...

Regards,
Loris.

 

[Loris Medici]

Sandy hi,

Yes probably so, but that's not a "normal shooting" situation for sure Maybe I could do it (if I were there) with the DRO feature in A700 by overexposing the highlights by 1 stop (and recover in RAW processor) and using the DRO adjustment at maximum (Level 5) but that would cause pretty much noise / processing artifacts in the shadows. What was the SBR of the scene? (A700 does something like 9 stops, something more if you don't mind clipping in one channel, see: http://www.dpreview.com/reviews/sonydslra700/page19.asp for DRO and dynamic range info...)

Regards,
Loris.

 

[sanking]

Loris,

The SBR of the scene was way more than 8 stops. I had a 5X7 camera with me at the time and made a B&W negative of the same scene. From my notes I exposed and developed for an SBR of 13. It may have been higher than that but I just never record anything larger than SBR 13.


Sandy

Sandy hi,

Yes probably so, but that's not a "normal shooting" situation for sure Maybe I could do it (if I were there) with the DRO feature in A700 by overexposing the highlights by 1 stop (and recover in RAW processor) and using the DRO adjustment at maximum (Level 5) but that would cause pretty much noise / processing artifacts in the shadows. What was the SBR of the scene? (A700 does something like 9 stops, something more if you don't mind clipping in one channel, see: http://www.dpreview.com/reviews/sonydslra700/page19.asp for DRO and dynamic range info...)

Regards,
Loris.

 

[Erik Ehrling]

Came across this on another site. http://fwd.five.tv/videos/challenge-blow-up-part-3.

This comparison is flawed in several ways. First of all it uses ISO 400 film, which everybody knows never have been very good in 35mm format. (If they used, say the new Ektar 100 my guess is that film would come out with more detail than digital from the D700). Also, they don't say how they scanned the negative - and scanning makes the comparison unfair anyway. A true comparison would show an analogue enlargement from film vs digital print from digital.

Regards,
Erik Ehrling (Sweden)

 

[Loris Medici]

Vaughn,

The SBR of the scene was way more than 8 stops. I had a 5X7 camera with me at the time and made a B&W negative of the same scene. From my notes I exposed and developed for an SBR of 13. It may have been higher than that but I just never record anything larger than SBR 13.

Vaughn? You're multitasking between here and APUG I presume

SBR 13 is pretty serious indeed. But I'm pretty sure I could have fetched something similar (but not in the same quality department) w/o HDR w/DRO at that spot -> it's amazing how much detail you can recover from the shadows when working digital; but with the cost of quality (higher noise/signal ratio)... If you handle A700 or A900 later, definitely evaluate the Advanced DRO feature. (I often find myself shooting at manual DRO Level 2 or 3...)

Regards,
Loris.

 

[sanking]

Loris,

Yes, I was reading a post by Vaughn when I posted. I corrected it as soon as I realized the mistake.

Anyway, I really know you fellows apart. But the thing is, maybe there is some subconscious thing going on here because Vaughn is really tall also!!

Sandy

Vaughn? You're multitasking between here and APUG I presume

SBR 13 is pretty serious indeed. But I'm pretty sure I could have fetched something similar (but not in the same quality department) w/o HDR w/DRO at that spot -> it's amazing how much detail you can recover from the shadows when working digital; but with the cost of quality (higher noise/signal ratio)... If you handle A700 or A900 later, definitely evaluate the Advanced DRO feature. (I often find myself shooting at manual DRO Level 2 or 3...)

Regards,
Loris.

 

[R Shaffer]

I would agree with Erik that not knowing how the scans were made and what other digital manipulations may have been made to the images makes this more of a show than science.

The sensor size for Nikons are DX for the smaller ( less $$ ) sensor and FX for the full frame sensor. Only the Nikon D700, D3 & D3x have the full frame sensors. I don't shoot Cannon, but the 5D & 1D Mark II & III also have full frame sensors.

If the SLR camera your looking at has a 'crop factor' then you know that it does not have a full frame sensor. Since they don't need to cram the sensors on the full frame sensors so tightly the noise ( digital grain ) is significantly reduced. Megapixels can be more sales pitch than really adding image quality.

I won't jump on the third-rail of photography digital vs film other than to say they are both fine tools for capturing images. Your millage may vary.

 

[sanking]

Loris,

BTW, if you go to this site you will see the carbon photograph I made from my visit to the old mosque in Edirne.

http://www.alternativephotography.com/artists/sandy_king.html

There are a few other image files there from my visit to Turkey, but I still have some decent ones that have not been seen anywhere.

Sandy

 

[Loris Medici]

Loris,

BTW, if you go to this site you will see the carbon photograph I made from my visit to the old mosque in Edirne.

http://www.alternativephotography.com/artists/sandy_king.html

There are a few other image files there from my visit to Turkey, but I still have some decent ones that have not been seen anywhere.

Sandy

Yes, I remember that image. Beatiful images / prints Sandy... Everytime I see your carbons I'm more sure I won't be returning to silver gelatin B&W printing again. Is that good or bad? I can't say...

The stairs image is very familiar (Nuzhetiye Mosque, Tophane district?), I have a similar one made with a 4x5" pinhole camera. Your viewpoint for that image is something like 50 feets on the right of the spot I did the pop pd stairs image I gave to you, right?

Best regards,
Loris.

 

[donbga]

This comparison is flawed in several ways. First of all it uses ISO 400 film, which everybody knows never have been very good in 35mm format. (If they used, say the new Ektar 100 my guess is that film would come out with more detail than digital from the D700). Also, they don't say how they scanned the negative - and scanning makes the comparison unfair anyway. A true comparison would show an analogue enlargement from film vs digital print from digital.

Regards,
Erik Ehrling (Sweden)

I agree with most of what you say. But I think the D700 is a really good D camera.

However I wish they would have used Ektar 100 as you suggest to compare with though I do think there are some good 35 mm 400 color negative films.

Perhaps if we write to them they will reveal more information about how the film was scanned and plead for them to repeat the test.

Don Bryant

 

[sanking]

Loris,

The stairs image was made somewhere in the old Greek district. I ccould not pinpoint the location but it is not so far away from an old, and very large Greek school we walked by.

BTW, there was an article in the Smithsonian magazine on Haigia Sophia specifically, and indirectly on how the old Byzantine part of Istanbul is rapidly disappearing.

Sandy

The stairs image is very familiar (Nuzhetiye Mosque, Tophane district?), I have a similar one made with a 4x5" pinhole camera. Your viewpoint for that image is something like 50 feets on the right of the spot I did the pop pd stairs image I gave to you, right?

Best regards,
Loris.

 

[Loris Medici]

Loris,

The stairs image was made somewhere in the old Greek district. I ccould not pinpoint the location but it is not so far away from an old, and very large Greek school we walked by.

BTW, there was an article in the Smithsonian magazine on Haigia Sophia specifically, and indirectly on how the old Byzantine part of Istanbul is rapidly disappearing.

Sandy

Ok, then it's not near to my spot (the old Greek school is pretty far from Tophane, it's in the Balat district). Sometimes architectural details look alike too much...

That's so true for any kind of "old" Istanbul (I mean including the Turkish / Ottoman / whatnot historical parts)... Unfortunately Turkish people aren't concerned about preserving history. I got so envious while I was in Europe; they are preserving / restoring and giving much importance to their history... (Spain / Italy - I'm sure other parts are the same!)

Anyway, this eventually is stealing the thread(!); so, let's continue privately if you like.

Best regards,
Loris.

 

[sanking]

I have had the Canon 5d, with 24-104 IS lens, for close to two weeks now and have done a fair amount of comparisons with the Fuji GA645Zi, withVario 55-90 lens. What I have basically tried to find out if I would be able to get more quality from a full sensor DSLR like the 5d than from my Fuji GA645. It is a pretty close call, as I thought it might be.

First, a few basic observations. My conclusions are based on shooting resolution targets and on field comparisons. For the tests I used an ASA of 100 with the 5d, and for the film camera I used Portra 160 VC, at EI of 100.

1. Size wise the Fuji GA645 Zi wins hands down, but the Canon 5d with the 24-105 lens has more versatility.
2. If you scan with a consumer flatbed the Canon 5d wins easily. In fact, the Canon 5d ends even with a scan with a very good dedicated film scanner when the resolution is 2540 spi. When scanning with resolutions of 5080 spi the film is better.
3. For color the Canon 5d wins easily again, in major part because scanning in color with my old scanner takes too long at 5080 spi. And at 2540 spi, as noted, the 5d wins easily.


I am going to upload a few files.

1. Epson4800–Crop Crop of a scan with the Epson 4990 at 4800 spi
2. Leaf2540-Crop – Crop of a scan made with the Leafscan 45 at 2540 spi.
3. Leaf5080-Crop – Crop of a scan made with the Leafscan 45 at 5080 spi.
4. 5d-Crop –Crop of the image file of the target shot with the 5d.

If you look at these crops, which have been adjusted to show all at about the same size, you will agree that the one with the Epson 4990 is pretty awful, the Leaf at 2540 is much better, the Leaf at 5080 has the best resolution, and the 5d crop is slightly better than the Leat at 2540, and very clean. BTW, the Leaf at 5080 spi is pulling all of the detail out of the film. I looked at the negative with a microscope and basically 5080 spi gets all of the detail.

Just for the record, the resolution of the Canon 5d at focal length of 35mm was 52 lp/mm, not bad for a zoon lens since 61 lp/mm is the theoretical maximum form this 12.8 mp camera.

I will put up some files of a real scene later.

Sandy King

 

[sanking]

OK, Part 2.

First, I will show a picture of the scene. I took shots of this scene with the 5d and with the Fuji GA645 Zi, adjusting the zoom of the 5d to match the horizonal field of the 5d. 35mm on the 5d is about 55mm on ther GA645 Zi. The Portra 645 negative was scanned in B&W at 5080 spi. Then I adjusted both files to give 360 dpi at size of 18X24"

I am attaching also two tight crops of the church on the left, one from the 5d file, the other from the GA645 file. There is slightly more detail in the GA 645 crop, but the 5d crop is much cleaner. I was a bit surprised that the 5d was as close in terms of detail to the GA 645 Zi as it is, but when you look closely at the two files you see that some of this is just illusion. For example, look closely at the twin windows just to the upper right of the dark car. On the crop from the GA 645 Zi you will see that the curtains are depicted as they should be, with horizontal lines. Now look at the same windows in the 5d file. They have assumed some type of strange geometric design that looks nice, but is very unreal. I can see several other features of this type in the crops.

Sandy King

 

[donbga]

OK, Part 2.

First, I will show a picture of the scene. I took shots of this scene with the 5d and with the Fuji GA645 Zi, adjusting the zoom of the 5d to match the horizonal field of the 5d. 35mm on the 5d is about 55mm on ther GA645 Zi. The Portra 645 negative was scanned in B&W at 5080 spi. Then I adjusted both files to give 360 dpi at size of 18X24"

I am attaching also two tight crops of the church on the left, one from the 5d file, the other from the GA645 file. There is slightly more detail in the GA 645 crop, but the 5d crop is much cleaner. I was a bit surprised that the 5d was as close in terms of detail to the GA 645 Zi as it is, but when you look closely at the two files you see that some of this is just illusion. For example, look closely at the twin windows just to the upper right of the dark car. On the crop from the GA 645 Zi you will see that the curtains are depicted as they should be, with horizontal lines. Now look at the same windows in the 5d file. They have assumed some type of strange geometric design that looks nice, but is very unreal. I can see several other features of this type in the crops.

Sandy King

Sandy.

What kind of post processing did you do for the 5D files?

Don

 

[sanking]

Don,

Pretty much the same as I did for the scanned files. I gave them a small amount of capture sharpening, and a small amount of creative sharpening with high pass sharpening.

I also adjusted the RAW files when I opened them with Photoshop. And I adjusted the image contrast with levels and curves.

Sandy

Sandy.

What kind of post processing did you do for the 5D files?

Don

 

[onnect17]

Sandy,
Do you have any comparison using the same lens and a finer grain film (non c-41)
Thanks,
Armando

 

[sanking]

Armando,

I have some comparisons of the resolution target with Fuji Acros but I have not scanned the film yet. Will add it to the comparison when I finish the scan.

The actual real scene with the church was shot only with Portra 160 and the Canon 5d. I may be able to get out and make another comparison of the 5d with the GA645Zi using Acros in the film camera Monday or Tuesday. I had planned to ship the 5d back to its owner but he does not need it for a while so I might hang on to it next week and do a few more tests with B&W film.

Sandy



Sandy

Sandy,
Do you have any comparison using the same lens and a finer grain film (non c-41)
Thanks,
Armando

 

[Loris Medici]

Sandy, after testing full frame 12.x Mp Canon 5d Mk I do you think a 24.x Mp full frame camera is going to be adequate for your needs?

Regards,
Loris.