Difference 65mm/70mm to Epic?

[Joseph Ward]

What is the difference for 65mm/70mm film camera to Epic? Will it convert better to imax? Thanks. :unsure:

[Jeff Kilgroe]

In terms of resolution, EPIC would be a good fit. There's some 4K DI going on in the IMAX world right now. But EPIC won't have the same optical characteristics of 70mm cameras. Other aspects such as dynamic range, we can only speculate on because we don't know anything about the new sensor and its abilities.

[Joseph Ward]

thank you!:)

[Tom Lowe]

65mm is still likely to wax Epic. But how many people are shooting 65mm? Next to none.

[David Mullen ASC]

Technically you'd want to double the RED sensor size and pixels from 4K to 8K to be in the large film format ballpark... though IMAX work is now done in all sorts of file sizes, from 2K D.I.'s like for "Charlie and the Chocolate Factory" and the last "Harry Potter" movie being blown-up to IMAX, not to mention HD-shot projects like "Superman Returns" and "Speed Racer". But ideally, you'd be shooting well over 4K for true IMAX-type photography on an IMAX screen.

A 15-perf horizontal 65mm IMAX negative is the equivalent of three 5-perf 65mm frames, and a 5-perf 65mm frame is twice as wide and one-perf taller than a 4-perf 35mm cine frame. So if you would normally scan a 4x3 4-perf tall 35mm frame at 4K x 3K, that's close to 3x 3K (for 12-perf, three 4-perf frames end to end) to get close to a 15-perf long IMAX frame, or 9K, to match the same scanning resolution.

So my feeling is that a true digital replacement for an IMAX camera would have to be in the 6K to 8K range, though people will make due with less obviously.

[Jorge Díaz-Amador]

ARRI estimates the equivalent pixel size to match the resolution of a 65mm camera (with modern optics and film) at 8746 x 3855 pixels.

The aperture size of 65mm is 52.5 x 23 mm which is double the width of Super 35, hence the DoF is shallow, even at medium stops.

All this being said, 65mm film acquisition (at least 5-perf) is all but dead (IMAX is 15 perf, and it's a niche market). So the best way to get large format quality is with a very high end digital cinema camera.

If the Epic enlarges the captured area to 24 mm wide or bigger, the DoF will be slightly less than RED One. With the full S35mm aperture, shooting at f2 will give the same DoF as 65mm at f4.

I think it's possible that with full S35 size capture, very high resolution fast lenses (like the Zeiss Master Primes) and the RED Epic, you should be able to get a 4K image that when properly projected will closely approximate the quality of a 70mm 4th generation contact-printed release print from a 65mm original.

[Jorge Díaz-Amador]

Now that I think about it, 4K digital projection might be a good way to show 65mm originated movies. One day there might be more 4K projectors than 70mm projectors.

What would be needed is an 8K scanner for 65mm (does this already exist?) From the 8K scan you could make a really good 4K master.

I've only seen 70mm projection of 65mm movies twice (2001 and Lawrence). In both cases it was AMAZING. As a filmmaker you owe it to yourself to see this quality at least once, to know what is possible, and what we are shooting for in the future.

[laguun]

IMAX is going digital for 100 new screens.
http://phx.corporate-ir.net/phoenix.zhtml?c=118725&p=irol-newsArticle&ID=1085167&highlight=
http://phx.corporate-ir.net/phoenix.zhtml?c=118725&p=irol-newsArticle&ID=1117681&highlight=
Imax DMR is usually done at ~3K.

Digital 8K is done and working, had the joy to see it working in 2006 from the NHK.
One has to be closer than ~1 meter to the ~10 meter screen to get any idea if there would be any pixels.
http://upload.wikimedia.org/wikipedia/de/9/98/Uhdfuji0000.jpg

The difference between 4K and 2K is almost impossible to see (with an human eye) if you dont have
a) a seat in very very first rows of the cinema
a) a/b split of the exact same image

Whenn IMAX or 65mm is projected via 4K, it does look good. As sharp as the 4K digital.
In 2007, there was a nice shootout of 1080/2K, 35mm2K/4K, 65mm/imax and digital 4K at the D-Cinema summit of IBC07.
35mm scanned (Spider 3, kitchen scene, street and alley scene) at 4K looked grainy in comparision. 65 mm quality was slightly below 4K. Only IMAX and digital 4K (Dalsa was used mainly) really showed the full potential of the Sony SRX.

However - one has to be *really* close to the screen to see any difference. As with 24bit96khz audio, digital exhibition is beginning to surpass the human eyes.

In todays (and yesterdays) 35mm mechanical projection cinema, no matter if with DI or without, one doesnt even get 1000 lpm - so the discussion is pretty academic ayhow.

From the 400-500 cinemas which are going digital monthly, >90% chose 2K, among other also for these reasons.

[Tom Lowe]

ARRI estimates the equivalent pixel size to match the resolution of a 65mm camera (with modern optics and film) at 8746 x 3855 pixels.

That seems maybe a little bit high. Film looses a lot of luster in the scanning process. The Canon 1Ds Mark III is roughly the same size frame as Vista Vision, I think. And it's certainly the same frame size as 35mm stills. It's resolution is 21MP (5616 x 3744). I don't know if a chemical 35mm still camera could match that resolution and image fidelity, after being scanned? Something makes me think, no.

Now, standard 65mm is a bit wider than 35mm still frames, right? But you have to figure that even with the wider negative, after you scan the 65mm moving film, it's not going to beat a digital sensor that is shooting 6000x 3500 or whatever. Technically it might hold a little bit more info in its chemical state, but if your end goal is a digital copy (which everything will be at some point), then you really would not need a sensor much beyond the 1Ds Mark III's resolution to match 65mm film.

[David Mullen ASC]

Just keep straight that IMAX is the equivalent of three 5-perf 65mm frames combined. So you want to separate 5-perf 65mm Super Panavision / Todd-AO photography from 15-perf 65mm IMAX photography in these discussions.

[Jorge Díaz-Amador]

I've had stills shot with my Contax G2 in 24 x 36 mm film format (nearly the same as VistaVision) scanned on an Imacon scanner, producing a file that's roughly 5155 x 7732 pixels, or 39.9 MegaPixels!

There are details in the image that are about two pixels wide. I'll post a sample here if you guys don't believe me (someone PM me and tell me how to post images properly here).

According to ARRI's "4K+ Systems: Theory Basics for Motion Picture Imaging", audience members with good (20/20) vision can see 4K resolution viewing a 25 meter [82 ft] wide screen at up to 20 meters [65 ft] distance.

I can DEFINITELY hear the difference between 16 bit 44.1 kHz audio and 24 bit 96 kHz audio (and I'm 40 and can't hear above 16 kHz). And remember how Sony told everyone that 44.1 kHz was "perfect sound forever"? [and 1080i compressed HD was perfect, blah blah]

I've heard that some audio manufacturer did a test with sampling rates, and people with good hearing could hear differences until about 500 KHz, then they could hear no improvement above that.

So 96 kHz audio does not surpass the human hear, and 4K digital cinema projection does not surpass the human eye.

5K capture and 4K projection. As Ridley Scott would say "It's a great start".

[Tom Lowe]

Hmmm... I don't know. I'm very skeptical of scanning technology. The file size might be 39MP, but that in no way means that your 39MP film scan could hold a candle to a Hasselblad 39MP shot, for example.

If the scanner is so great, why not cut out the middle man (chemical film) and just shoot digitally? That's basically what DSLRs are.

[David Mullen ASC]

If the scanner is so great, why not cut out the middle man (chemical film) and just shoot digitally? That's basically what DSLRs are.

Scanners don't have to work in real time... and most DSLR's don't have to shoot 24 times a second, if not more. This is why single-sensor digital cine cameras like the RED, Phantom, Dalsa, etc. are not as easy to build as some people think. There are heat issues with firing a sensor that often, and data management issues with processing and recording that much data.

[Tom Lowe]

Scanners don't have to work in real time... and most DSLR's don't have to shoot 24 times a second, if not more. This is why single-sensor digital cine cameras like the RED, Phantom, Dalsa, etc. are not as easy to build as some people think. There are heat issues with firing a sensor that often, and data management issues with processing and recording that much data.

Yeah. My DSLR's sensor get really hot when I shoot timelapse at night with like 60 or 90 second exposures - for 8 hours at a time, with only 7 second intervals to "cool off." It's funny, because when I shoot during the winter, and it's 30 degrees out at night, the sensor does fine (no overheating). But when I shot last summer at Lake Powell, when it was 85 degrees at night, my sensor got really hot and super noisy.

David, do you agree with my point that a 35mm (still frame) scan at 39mp will not even come close to a Hasselblad 39mp shot, in terms of a digital final? (Of course, the Hasselblad is a larger format.)

Every time I have seen even top-quality neg scans from chemical flim, I have never been impressed with the resolution, color, etc. Scans just look awful, to me. I would very much like to see a comparison, for example, between a 21mp RAW shot on the new Canon 1Ds Mark III vs a 21mp scan of a 35mm still shot on a regular Canon EOS chemical camera. My guess is, that displayed digitally, the 1Ds will blow the chemical scan out of the water.

[Robin Balas]

Yeah. My DSLR's sensor get really hot when I shoot timelapse at night with like 60 or 90 second exposures - for 8 hours at a time, with only 7 second intervals to "cool off." It's funny, because when I shoot during the winter, and it's 30 degrees out at night, the sensor does fine (no overheating). But when I shot last summer at Lake Powell, when it was 85 degrees at night, my sensor got really hot and super noisy.

David, do you agree with my point that a 35mm (still frame) scan at 39mp will not even come close to a Hasselblad 39mp shot, in terms of a digital final? (Of course, the Hasselblad is a larger format.)

Every time I have seen even top-quality neg scans from chemical flim, I have never been impressed with the resolution, color, etc. Scans just look awful, to me. I would very much like to see a comparison, for example, between a 21mp RAW shot on the new Canon 1Ds Mark III vs a 21mp scan of a 35mm still shot on a regular Canon EOS chemical camera. My guess is, that displayed digitally, the 1Ds will blow the chemical scan out of the water.

This is a way old topic in the stills photo world and such comparisons were carried out years back with 1Ds (1.edition) and scanned 645 film and the Canon was slightly better or almost there depending on the specifics of the tests, but no doubt as to being better than 35mm film scans even on provia slide film.
Today the Hassy 39MP, PhaseOne 39MP, Leaf Aptus 31MP and Sinar 31MP offerings is bettering 4x5 largeformat film and comparisons towards 5x7 and 8x10 is beeing made and with the multishot 39MP back it seems to be better than 8x10, but that is a whole different story.

Scanned MP vs. RAW capture MP is definitely NOT equal. 8000ppi scans (max I can test) of 35mm positive stills film which is way better than the movie stocks, show no benefit over 4-5000ppi scans (depending on the stock), and it was done on my drum scanner and on a colleague's Imacon scanner.
The gain above 3000ppi is mostly in grain shape, not real info - but I tend to prefer the higher scan ppi as the prints on paper look more natural when the grain is visible and important for the expression.
3000ppi of a 35mm stills is a 12MP scan file. A 11MP dSLR (1Ds) is way better than the scans or resolution and delivers cleaner results.
One example to look at is: http://luminous-landscape.com/reviews/cameras/1ds/1ds-field.shtml
and remember this is the 1.edition of the 1Ds camera and years back, the new 1DsIII is a lot better.

Claiming IMAX is equivalent to a 8-9K digital cinema camera is ridiculous, claiming it needs a 8K scan for DI is at best being conservative and playing it safe, but scanned MP is definitely not equal to raw capture MP, thats a fact known for years in the stills photo business.
MHO.

[Tom Lowe]

Robin, thanks for the info. That's pretty much what I figured.

Do you beieve this claim that "ARRI estimates the equivalent pixel size to match the resolution of a 65mm camera (with modern optics and film) at 8746 x 3855 pixels"?

To me, that number sounds far too high. It's much higher than the 1Ds Mark III's resolution, but I seriously doubt that a scanned 65mm movie neg is going to blow the 1Ds Mark III out of the water, like those number would seem to suggest. I mean, that might be the max resolution that 65mm cine film can resolve to in a scan, but I bet after scanning the chemical film, a 6000x3500 RAW digital frame (using the same optics) would easily match it, or beat it, in terms of clean resolution.

[Radoslav Karapetkov]

There is something about film... dunno what it is...

Compares to the effect that some mushr..... umm nevermind.

EDIT:

Dunno if these were just CC-ed in the right way, but.... there something about them:

http://www.foveon.com/gallery.php

[David Mullen ASC]

Just keep in mind that movies are made up of moving images, so an individual film frame, while showing some degree of graininess, tends to look cleaner as frames are flashed in succession.

Also, tests clearly show that the 35mm motion picture frame needs to be scanned at 2K to 4K, 4K preferable, to capture all the information on the frame. You can see that here:

http://www.etconsult.com/papers/Technical%20Issues%20in%20Cinema%20Resolution.pdf

So if 24mm-wide frame of a full aperture 4-perf 35mm film is best scanned at 4096 pixels, then basic math says that a piece of film that is 70.1mm wide (IMAX) needs to be scanned at 11,964 pixels across just to equal the same resolution as the 4,096 pixel scan of the 4-perf 35mm piece of film where the frame is 24mm wide. Otherwise, a 6K scan of an IMAX frame is equivalent to a 2K scan of 35mm in simple real estate.

As for 8K to 9K RAW being ridiculous for a digital version of an IMAX camera, personally I believe it has to be double that of a 4K RAW camera to be a significant improvement for throwing onto a 5-story tall screen. Of course, people will make do with less, as they currently are, but that doesn't mean they should. IMAX has to be signicantly better than what people currently see in 35mm or even 5-perf 65mm for it to be a special experience.

5-perf 65mm is twice as wide as 4-perf 35mm, so if 4-perf 35mm should be scanned at 4K, it's no big stretch to think that some piece of film that is physical twice as wide should be scanned at 8K.

Just remember that there is a difference between ideal scanning resolutions and picture resolution, which shouldn't even be measured in terms of pixels. The number of pixels in an image is not an indication of sharpness. So ARRI is being a bit confusing by mixing the two issues up. This will continue as long as people keep thinking that pixel resolution and sharpness are the same thing.

[Tom Lowe]

I agree that nothing shot digitally can even come close to being legitimate on an IMAX screen right now. That neg is so massive that it will be years before a digital IMAX is possible. The only digital camera I think that might be close to doing IMAX work, as I have said many times before, would be a timelapse IMAX shot on something like a 39mp 'Blad.

As for 65mm, my guess is that singles frames from the 1Ds Mark III could compete with single cine frames of 65mm. I could be wrong, but I bet after scanning the 65mm will lose some luster. In other words, I don't think digital is all that far away from being able to compete with 65mm. It all comes back to dynamic range, really. Technology has proven over and over and over again that the pixel count and resolution will continue to increase year to year. The real question out there is increasing dynamic range.

[Radoslav Karapetkov]

And full film-like color and random pixels, please. :)

I wonder how this random grains thing could be achieved.

Isn't it possible in post?

Cause just adding digital grain doesn't do it really. Not the same effect....

The image-carrying pixels themselves have to look randomized somehow...

Hmm...

I think that the day digital will really replace film is the day when it starts to physically acquire the image the way film does...

Three-layer sensors, 4K+ full resolution, 15+ stops DR, random pixels... you know?... all the good things. :)

[David Mullen ASC]

True 15-perf 65mm IMAX photography shot on slow film in daylight is amazing to see on the big screen, but there have been a lot of smaller formats used for IMAX release, including 3D HD like for "Ghosts of the Abyss", not to mention all the 35mm and HD movies that have gone through the IMAX DMR process (de-graining and re-sharpening). I'm sure 4K RAW will be used for some IMAX project, no doubt, and it may look OK. At least it won't have to be degrained, and that will help retain some sharpness.

As much as everyone says that 35mm is more like 2K than 4K, every HD movie I see on the big screen (technically 1.9K, nearly 2K) looks slightly soft unless it is digitally projected. Last one I saw was "Speed Racer", before that "21", "The Other Boleyn Girl", and "Forbidden Kingdom", the last looking the best because I saw it digitally projected.

To really be crude and inaccurate (using pixels as a measure of resolution, which I said was not a good idea), I'd say that good 35mm anamorphic photography is more like 4K, good Super-35/1.85 is more like 3K, and HD is naturally like 2K. And your more mediocre work in 35mm is more like 2K.

As for the advantages of grain, the biggest advantage is something that can't be simulated in post, that the capture medium changes the position of its "photosite" (the silver halide grain) on every frame, reducing the chance of any aliasing artifacts. Sensors with fixed patterns of photosites, on the other hand, need anti-aliasing filters to get rid of the problem, thus losing resolution. Adding grain in post just adds a graininess to the image, which is fine if you want that.

Another benefit to some moderate grain on a theater screen is that if the grain looks sharply focused, it gives the impression of a sharper image, which is why some scenes shot on 50 ASA film, nearly grainless, can seem blurrier than scenes shot on grainier stock. Digital has the same "problem", the grainless image forces the eye to look at the true sharpness of detail in the frame and judge that. This is one reason, I believe, that still frame comparisons almost always show that the 35mm movie frame has less detail than some 4K digital frames, yet the impression when the movie is projected is that the digital image seems softer than the film image.

[Tom Lowe]

I never paid all that much attention to grain as a moviegoer, until I started getting 1080p bluray movies and watching them pixel-for-pixel on a large LCD. Now I see grain all over the place!... and to be honest, I don't really care for it. It's utterly random, and not that pretty to look at in detail. Some may like it, out of nostalgia, or just for some other aesthetic reason... I dunno.

At 4K, displayed on 4K LCDs pixel for pixel, I imagine that grain could become quite distracting. The farther from the screen, though, the less you notice it.

[Jorge Díaz-Amador]

I'm enjoying this thread. OK as promised, here is my proof that a 40 MegaPixel scan from a 24 x 36 still film image is not a waste of pixels:

http://www.cinematechnic.com/images/photography/Big_Ben_400h_tagged.jpg
This is the full frame image 7732 x 5155 pixels 39.9 MP (reduced to 400 pixels v)

http://www.cinematechnic.com/images/photography/Big_Ben_CU_sample.jpg
And here is an enlargement 1:1 pixel ratio, from the center of the image.

And here's the crazy part: I shot that picture hand-held.

I believe the film I used was Kodak Ektachrome 160 D (which is also, or was also, available in 400' rolls for cine use). Pretty sure I used the Zeiss G-Planar 45mm f2, probably at f5.6.

Enjoy!

[Tom Lowe]

Nice photo!

Okay, here is a sample 1Ds Mark III image.

http://img337.imageshack.us/img337/586/1dsmarkiiisamplepn6.jpg

Here is the same image after I used the simple photoshop bicubic upsample to 7732 x 5155. After up-resing, I did a 100% crop, and this was the result:

http://img79.imageshack.us/img79/929/1dsmarkiiisamplecropyo0.jpg

To me, this looks much cleaner than the chemical film scan.

Take a look at how clean the 1Ds Mark III is even at night (large photo!)....

http://web.canon.jp/imaging/eosd/eos1dsm3/downloads/nightscene.jpg

[Jorge Díaz-Amador]

Hey Tom,

Thanks for the comparison photos. You won't hear me putting down the 1Ds Mk.III (I am accepting donations ;-) but I paid about $700 for the Contax G2 / 45mm Planar combo. How much for the 1Ds?

[Tom Lowe]

Hey Tom,

Thanks for the comparison photos. You won't hear me putting down the 1Ds Mk.III (I am accepting donations ;-) but I paid about $700 for the Contax G2 / 45mm Planar combo. How much for the 1Ds?

Enough to make me cry when I think about trying to buy it. :waaa:

[David Mullen ASC]

To me, this looks much cleaner than the chemical film scan.


Sure, it's less grainy, because it's not film, but it doesn't look sharper or more detailed.

[Tom Lowe]

With some sharpening...

http://img172.imageshack.us/img172/662/1dsmarkiiisamplecropshaet7.jpg

Causes some aliasing, though.

[Robin Balas]

There is something about film... dunno what it is...

Compares to the effect that some mushr..... umm nevermind.

EDIT:

Dunno if these were just CC-ed in the right way, but.... there something about them:

http://www.foveon.com/gallery.php

;)
you do realize they are from the Foveon X3 sensor and not from film?

[Craig W. Bickerstaff]

With some sharpening...

http://img172.imageshack.us/img172/662/1dsmarkiiisamplecropshaet7.jpg

Causes some aliasing, though.

Now it looks like Mini DV

[Radoslav Karapetkov]

;)
you do realize they are from the Foveon X3 sensor and not from film?


Yeah, yeah, I miswrote the post.

I know it's not film, the point is that something about the color of foveonX3 images makes my film-accustomed heart feel at ease.

[Robin Balas]

Just keep in mind that movies are made up of moving images, so an individual film frame, while showing some degree of graininess, tends to look cleaner as frames are flashed in succession.

Also, tests clearly show that the 35mm motion picture frame needs to be scanned at 2K to 4K, 4K preferable, to capture all the information on the frame. You can see that here:

http://www.etconsult.com/papers/Technical%20Issues%20in%20Cinema%20Resolution.pdf

So if 24mm-wide frame of a full aperture 4-perf 35mm film is best scanned at 4096 pixels, then basic math says that a piece of film that is 70.1mm wide (IMAX) needs to be scanned at 11,964 pixels across just to equal the same resolution as the 4,096 pixel scan of the 4-perf 35mm piece of film where the frame is 24mm wide. Otherwise, a 6K scan of an IMAX frame is equivalent to a 2K scan of 35mm in simple real estate.

As for 8K to 9K RAW being ridiculous for a digital version of an IMAX camera, personally I believe it has to be double that of a 4K RAW camera to be a significant improvement for throwing onto a 5-story tall screen. Of course, people will make do with less, as they currently are, but that doesn't mean they should. IMAX has to be signicantly better than what people currently see in 35mm or even 5-perf 65mm for it to be a special experience.

5-perf 65mm is twice as wide as 4-perf 35mm, so if 4-perf 35mm should be scanned at 4K, it's no big stretch to think that some piece of film that is physical twice as wide should be scanned at 8K.

Just remember that there is a difference between ideal scanning resolutions and picture resolution, which shouldn't even be measured in terms of pixels. The number of pixels in an image is not an indication of sharpness. So ARRI is being a bit confusing by mixing the two issues up. This will continue as long as people keep thinking that pixel resolution and sharpness are the same thing.

This is a little bit of a deja vue from some years back, but when I wrote what I did I based it on RW experience and we need to include the lenses and film stock in that in that, and for stills photo there is no lenses as sharp as 35mm lenses (pr.mm). So I agree if you can find a lens which renders a large image circle with the same amount of sharpness pr. mm as the smaller one you can scale up everything linearly. But thats in theory, in real life there is no such lens available at least to stills photography, and I would be surprised to see it for Imax over S35 lenses, hence the lessened criteria for the larger film formats. Look at The Hassy X5 film scanner (best there is) which can scan 35mm at 8000ppi but is limited to 3200ppi for medium format and 2040ppi for large format film, this is totally in line with the real world needs of top-end stills photo and as much as one needs, as there really isn't much more useful information there due to real world limitations in lenses, film base and actual detail the viewer expects to see.

I find it a bit strange that motion film needs higher scan resolution than stills film as the film stock normally is less detailed and more grainy than good stills stock, but not always. However it could be more important to render the grain structures in film and/or avoid grain aliasing as it is a moving image with different grain structures from frame to frame. Anyhow this is not an issue with digital capture and pixels scales elegantly past scanned film without needing the same res no matter what your PDF claims as every pixels is full color and every grain is a monochrome structure which needs to rasterize a color to make it a tone.
I am open to new facts as film is differnet than static stills, but I have seen so much rw evidence that limitations in optics and also thickened film bases in the larger formats impacts real world performance to such an extent that doubling the physical size of each frame gains you a lot less than linear scaling suggests. BUT there is the issue of the moving images and grain stuctures which may be very important for film and it could be more important than I see ,and I have yet to see a good scanned Imax frame up close 1:1, as they aren't really commonly laying around in my part of the world, so... :whistling:
But you are probably right that 2x 4K RAW is optimal for IMAX size images, but it might not be neccesary to present better than a lot of the non-optimal Imax productions I've seen. But as a goal, definatly! But to avoid lens issues the chip also have to be larger physically as well as having a high pixelcount.
MHO.

[Peter Majtan]

I have seen the test done by ARRI couple of years back that analyzed how far is it reasonable to go while scanning 35mm film. The ARRI-LASER is capable of 6K scans from 35mm and I had some done for our restoration project. The thing was that even at 4K You start seeing more of the actual film grain and material structure, rather then any more useful image information. Sure - You can scan it at 500K if You wanted, but it will not give You any more image information.

Personally I liked the best the 3K scans (native format for ARRI-SCAN). Anything more then that and the balance shifts away from the actual image captured. You could create the same effect by up-scaling the 3K image to 6K and then apply digitally film grain. It wouldn't be much different from the 6K scan. This is almost like the chroma sub-sampling issue with video.

I am going to hiding now...

[Dj Joofa]

I believe, that still frame comparisons almost always show that the 35mm movie frame has less detail than some 4K digital frames, yet the impression when the movie is projected is that the digital image seems softer than the film image.

Digital image appearing soft may have to do with what compression was used. If wavelet-type JPEG2000-based compression was used then it is not surprising that the image looked soft. A different compression may have a different result. For e.g., JPEG2000 and JPEG kind of trade sharpness with blockiness. Hence, if JPEG were used, then the image may have appeared blocky but sharper in relative comparison at least to the JPEG2000 image.