Agfa Copex Rapid vs. 5DmkII (and vs. Efke25)

[Fernando2]

Just a very quick&dirty comparison, nothing scientific.

Time ago I shot a building with my Canon Elan 7 equipped with EF 50/1.4 USM, loaded with Agfa Copex Rapid (rated at 40).
I developed the Copex in Caffenol LC+C (SPUR Modular UR New costs too much per roll, for my taste).

After some time I reshot the same building (from a slightly different point of view) with my Canon 5DmkII, same lens as before.

Then a few days ago I shot it again, this time with Efke 25 (rated at 20 and developed in XTOL stock).

Here's the building, whole frame; note the red rectangle.

Now the crops from the red rectangle.

The two film crops were taken with a microscope.
The 5DmkII crop was upsampled to match the size of the film crops (I know, I know, but there's no "right" way for this kind of comparison).

I'm sure the Copex Rapid would fare even better with its specific developer (SPUR Modular UR New), but it costs too much more than the Caffenol (0.25 eur per roll), so that's it. 😀

Can't wait to pit some 120 gear armed with Copex Rapid against the poor little 5DmkII. 😉


Fernando

 

[tsiklonaut]

Interesting stuff Fernando.

If 35mm film scores like that then 120 will drown it with Copex. Although few of us use ASA 25-50 films in everyday shooting conditions it sure shows us the film is still a very capable medium and just imagine if film R&D would had continued where it stopped somewhere in late 1990s or early 2000s? We'd probably had a regular ASA100-200 color E6 films with similar resolution by now, RMS 4-5 or less.

In any case even IF digital beat the film in resolution those examples clearly show digital just has this utterly sterile and cheated "feel" in it, like something out of tupperware plastics factory with a lot of processing care taken of. I guess it's a taste thing but that's one of the main reasons why I swiched to film myself - the overall "feel" of the image that the digital just quite didn't do it (Foveon X3 is slowly getting there though).

 

[Photo_Smith]

When you do your next test try some randomised detail like distant trees, digital does well with cityscapes and buildings, things that can be defined within grids–take your test into the country to test the resolution rather than the interpolation and smoothing algorithms.

 

[Fernando2]

Sorry, does not make much sense.

I already know that 35mm Copex outperforms 5DmkII by a sizeable margin, I have no use in doing more tests about this.

 

[Photo_Smith]

It makes perfectly good sense.

I didn't say you should re-do the tests, although you did seem to suggest you were going to do another test with 120 if you didn't mea culpa-sorry.
Just ignore the following advice:
When testing any film against digital (In the future) try not to use structures with 90° angles and lines that can be interpolated.
These leave an easy target for the smoothing algorithms in the DSLR image pipeline.

When (if) you (or anyone interested in your post) do that one (or any test) try randomised detail for a better idea of the resolution of the two mediums, because if you did that you'd see an even more conclusive result rather than the one we can see here; where some may actually prefer the digital because of lack of noise.

•Just generic advice so you can really see where the film sings.

 

[Stuart John]

It's interesting, I am wondering what size the print would have to be to see the advantage the film has. I also wonder if film is not better for small optical prints compared to digital. I alway prefer my small 3.5x5 inch traditional B&W prints over digital versions. I wonder if we end up with more detail in small optical prints with digital prints being limited to typically 300PPI.

 

[sjw617]

Sorry, does not make much sense.

I already know that 35mm Copex outperforms 5DmkII by a sizeable margin, I have no use in doing more tests about this.

The pictures above do not support this statement. The 5D MKII is much clearer, much less grain/noise and overall better quality. I do not see how you can say the Copex out performs, let alone by a sizable margin.

 

[Fernando2]

The pictures above do not support this statement. The 5D MKII is much clearer, much less grain/noise and overall better quality. I do not see how you can say the Copex out performs, let alone by a sizable margin.

Well, then I have nothing to say.
The digital shot is so grossly less detailed against the film shot, it's not even funny.
You clearly mix "clean" with "detail-less".

Fernando

 

[Fernando2]

When testing any film against digital (In the future) try not to use structures with 90° angles and lines that can be interpolated.

I still don't agree.
The digital shot has much less details than the film shot, one must be blind not to notice it.
No grill in the conditioner's external unit, no connectors on the unit's pipes, rounded rectangles when actually they are squared (as seen on film), boiler's cabinet key rendered as a black spot with no shape, and so on.
Of course a comparison of this sort is aimed at people who can understand details, otherwise we may have to educate them first; which is not my purpose.

Fernando

 

[Photo_Smith]

I still don't agree.
The digital shot has much less details than the film shot, one must be blind not to notice it.
Fernando

It doesn't matter if you agree or not. It is a fact that DSLRs record (sample) on a grid based sensor.
Just like this:

Buildings and structures that are easily reconstructed by algorithms can guess detail (normally through edges) and smooth the areas in between.

So in your case copex is quite rightly ahead but it seems more marginal than it should, in fact some will object to the random sampling of the film (noise from grain) and prefer the 'made up' detail smoothed by the 5D.
In other words they will object to grain (noise) and will not see the detail in the film shot and prefer the detail-less smooth digital–you can't educate people on this because it's partly personal preference.

The only way to really show them is to remove the easy to interpolate edges and replaced them with random ones.

So to be clear I'm NOT saying the 5D has more detail, what I'm pointing out to you is randomisation of detail is better than ordered detail when trying show films superiority in ultimate resolution terms.
Like below:

So while I agree with you in the case of your test, I'm just pointing out a better method for anyone trying to show resolution, and trying to help you understand why some will prefer the 5D shot, and how to remove that possibility in the future.
Regards
MA

 

[Fernando2]

I work in the field of digital image processing, it's not like I don't know how digital cameras work; for some time I even joined the RawTherapee team to work on some aspects of the demosaicing and sharpening algorythms.
So, I see your point; what I'm saying, is that if someone (not you, evidently) is not able to tell the difference between interpolated blotches and real details, it's not my fault: the presented comparison tells everything it needs, to a knowledgeable observer.

The problem with using trees and other natural details within comparison shots like these, is that they vary over time. Trees grow, leaves fall, greens are mowed etc.
Moreover, wind is a problem when trying to compare shots. Branches move, adding motion blur etc.

Buildings and other artificial structures do not change much over time, which is very convenient; the two shots I've used for the comparison were taken months apart, for example.

Anyway, when I'll pit 120 Copex Rapid against the 5DmkII, I'll try to include trees and non-geometric details as well, to offer a more complete comparison.

Fernando

 

[Photo_Smith]

I didn't mean to infer you didn't know what you are doing; from your prior posts I can see you do.
I just pointing out to you that a great deal of people will prefer sharp edge. clean low resolution just like the 5D shot over the Copex.

With random detail it takes that possibility away as it will be detail vs mush.
It can be different random detail, it doesn't need to be the same tree year after year just as long as the individual comparisons within each test are the same.

So I understand your test and your point; what I'm saying is random detail will make the result so self evident that you won't even need to be remotely 'knowledgeable' and even people like SJW above will see your point and even the Luminous landscape will find it hard to say the 3mp D30 is as good as 6x7 film....

Anyhow happy new year.
regards
Mark Antony

 

[DNG]

Even with the grain of Copex Rapid..
There is more detail in the Compressor units grills, and, the edges of the grills are a bit better defined.
Even though the grain gets in the way of grill edges, the edges seems a tad sharper. While the 5D grill edges seem a tad blurred.

 

[Fernando2]

even the Luminous landscape will find it hard to say the 3mp D30 is as good as 6x7 film....

Hey, I remember that article!!!!! 😀
Oh my God, don't let me start on LL!! :bang:

Happy 2014 to you too, Mark.

Fernando

 

[hlritter]

I was very interested to see your comparison of AGFA Copex Rapid with a full-frame DSLR, Fernando. I have just begun to shoot my first roll of 35mm with this emulsion. In addition, I have been personally doing some of the same sort of comparisons, so some of the frames I shot with this new film are of my standard target for such things, a house across our lake, about 300 m away. I shoot these test frames with a 35mm prime lens, which puts the uprights of a railing on the porch of the house right at the scale which requires a FF sensor of about 24Mpx to just resolve. My (also) quick & dirty tests with Tr-X and Fujicolor 200 gave very inferior results to my Nikon D800, which is why I'm now testing T-Max 100, Copex Rapid, and AGFA RPX 25.

If you're interested, you can see my results with Tri-X, Fujicolor, and several digital cameras at http://gallery.leica-users.org/v/hlritter/Res+Test+Crops/NEX735.jpg.html

One interesting thing I found was that when I took a photo with the NEX-7's zoom set to an actual 35mm FL, so as to get the same image scale on the smaller sensor as on the FF sensors with the 35mm prime, the detail was clearly better than with the D800. If you expand the Sony's APS-C sensor to FF size, keeping the pixel size the same, you have a 54Mpx sensor. This indicates that modern lenses, even inexpensive kit zooms, are capable of significantly better resolution than even the D800's 36Mpx sensor is. What is the optimal Mpx size of a FF sensor for matching to lens resolution, and are the major manufacturers aiming toward reaching this point eventually?

The results with film were so inferior that I'm hoping that using finer emulsions will give better results. Your results with Copex are very impressive, but I must take note of the fact that one part of your images where you said that Copex excelled over the 5DmkII, the grill of the air conditioner, was shot at a much shallower and less favorable angle with the DSLR, so you can't conclude anything from the lack of detail there.

 

[rumblesushi]

Interesting comparison Fernando, thanks for posting.

Everything is subjective, but I do find it hard to believe anyone would prefer the look of the 5d. It just looks so artificial, textureless, and interpolated.

The film looks beautiful, even at this magnification, and holds significantly more detail.

If my roll of Bluefire Police comes out OK, I'll post a comparison with a Sigma DP2m.

 

[Ko.Fe.]

Surprising results, to be honest.
My 6x9 scans with V500 aren't good on pixels level comparing to my 5D RAWs, it seems.
I never tested it for real.


To me DLSR makes sense on pixel level at only one practical case.
Macro for live objects with bigger than life size magnification.
Film for it would be too much hassle to look into bug's eyes 🙂

 

[Spanik]

A bit of the same conclusions and then some different ones.

Yes, the look and feel of a 6x4.5 slide still can't be matched by digital if you project. On the other hand the difficulty of getting that out of a slide is too much compared to using digital. To get that last bit out of film that makes the difference is far too consuming in time and energy to be worth it.

I tried the foveon and while I'm more or less convinced it is the most film-like digital it just isn't worth the hassle. (I just bought a dp-1m and dp-2m) Fine, but even more demanding in exposure and developing than a provia 400x. You only miss the scanning "experience".

But so far I haven't find any digital camera that is as easy to use as any old film camera ever made. With any film camera ever made you get an iso knob, a speed dial, a diaphragm ring and a focus ring (and a decent focusing screen) and with those you can take any picture whatever you want. With any digital camera you first need to read about 100 pages of manual that will learn you absolutely nothing about taking pictures and the user interface will get in the way with every photo you ever try to take.

I always maintained that the first to make a digital camera with the same user interface as a Praktica MTL-3 won't know what hit them as orders will come in faster than they can count. So far there isn't a single camera that has made me think otherwise.

 

[Spanik]

I still don't agree.
The digital shot has much less details than the film shot, one must be blind not to notice it.
No grill in the conditioner's external unit, no connectors on the unit's pipes, rounded rectangles when actually they are squared (as seen on film), boiler's cabinet key rendered as a black spot with no shape, and so on.
Of course a comparison of this sort is aimed at people who can understand details, otherwise we may have to educate them first; which is not my purpose.

Fernando

Hello? Please do not confuse detail than cannot be seen from the different angle between the two shots with lack of detail. The key of the boiler melts into the "window" because of the different angle. There are no connectors on any pipes in any of the shots to be seen at all, the detail in the grid is just as lacking in the efke shot and the detail of the bracket is only visible in the 5D shot, there isn't any in the film shots. In the 5D shot I can see that the side grill is fastened with 2 screws something not visible at all in the 2 film shots. In the efke shot you cannot even make out the trunking containing the piping from the wall! Nor any of the slots in the chimney of the boiler.

That "poor little 5D" is just destroying your film! And I'm a film user myself. These shots are very much telling you how much film is behind digital whatever you want to tell yourself.

 

[VertovSvilova]

I don't quite understand these film versus digital comparisons. They are two different media with two different aesthetics.

Anyway, I've never used Copex Rapid and noticed that Freestyle (one of my local film vendors) says it's microfilm. I might try it. Is it like Tech Pan (which was also a microfilm emulsion)? Does it have similar characteristics in pictorial uses? I was never a huge fan of Tech Pan but I did like what Lewis Baltz (one of the 'New Topographics' artists) did with it in Park City and The New Industrial Parks Near Irvine, CA.

(btw, here's an interesting interview with Baltz from 1993; I took a class from him in grad school when he was a visiting faculty.) http://www.americansuburbx.com/2011/03/interview-interview-with-lewis-baltz.html