Wasn’t normal 35mm film about the equivalent of somewhere between 4k and 8k depending on the film stock?
Plus, the projector optics will always limit the sharpness of the picture. No lense is ideal, and even ideal lenses would have fundamental limitations due to diffraction.
As far as lens optics, we’re really splitting hairs here. 70mm through a quality lens in an imax theater is going to look absolutely fantastic and stunning. Digital is just more convenient and at some point it will catch up and surpass film.
My point was more like that even IMAX film doesn’t quite get to 18k equivalent, more like 12 to 16k. Honestly, anything above 4k (for normal widescreen content) even on big screens is barely noticeable if noticeable at all. THX recommends that the screen should cover 40° of your FOV; IMAX is what, 70°, so 8k for it is already good enough. Extra resolution is not useful if human eye can’t tell the difference; it just gets to the meaningless bragging rights territory like 192 kHz audio and DAC-s with 140 dB+ S/N ratio. Contrast, black levels, shadow details, color accuracy are IMO more important than raw resolution at which modern 8k cameras are good enough and 16k digital cameras will be more than plenty.
Yup that’s why people can go back and rescan old film movies to make them into 4k now that we have better cameras, but you can’t do that with movies that were recorded with digital
Yeah, we’ll have this brief digital gap from the era when film was going out of fashion and 4k and higher resolution digital cameras weren’t a thing yet. But now that even average youtubers are shooting 4k with cheap(ish) DSRL-s, we generally don’t have to worry about the content having “not good enough quality for the future”.
The bigger problem IMO is the ephemeral and profit-driven nature of modern content distribution. Once the studio decides a film/series is not making enough money and pulls it from streaming, it’s gone. IIRC, DRM of DCP is also remotely managed so even if a cinema physically has the drive with the movie, they can’t play it when the studio pulls the plug–this was not the case with film.
Yeah all that is a huge problem, I remember Microsoft pulled the game Scott pilgrim from the Xbox 360 so if you didn’t buy it beforehand you couldn’t get it anymore until they did some legal stuff to get the game back in the store.
I still think film today is a great tool for getting high resolution photography at a cheap entry cost, a full sized digital sensor camera can be pretty pricey where as a 35mm film camera can be had pretty easy, then once you go to medium format it’s gets more expensive and then I’m not even sure there is large format digital cameras
I don’t think there’s any reason we couldn’t make a store 18k video.
And we could make screen at much higher resolutions that that at imax size, or even quite a bit smaller, though I suspect it would be absurdly expensive.
As I said I’m sure we could make screens that could do that. They would be absurdly expensive and heavy and stupid, but it could be done. Not worth it though.
And it looks like at least 16k cameras have been made.
The screens aren’t the problem. It’s often the hardware driving it. The current top generation of gaming gpus struggles at 8k. There’s very little chance of being able to render and play 16/18k
I can’t speak for video, but for audio production that isn’t true. Audio signals can be perfectly reproduced, up to some frequency determined by the sample rate and up to some noise floor determined by the bit depth, digitally. Set that frequency well beyond that of human hearings and set that noise floor beyond what tape can do or what other factors determine, and you get perfect reproduction.
I don’t know if perfect reproduction necessarily sounds better. It’s probably subjective, but projects I’ve worked on that were tracked with tape have a quality that you can’t get from digital. I’m not talking about tape hiss or anything like that. There’s a roundness to the sound.
Some people do like the distortion that analog audio provides, that’s true. But it is because of something that wasn’t in the original audio. It’s an artistic choice.
Yeah, but they’re likely digitally editing it all now, so it loses that in the middle of the process. Can’t really see why it would make sense to print a digital file back onto film.
Despite being what a lot of analog fetishists focus on, a purely analog editing chain is not necessarily what causes the effects that they actually like about analog media. Most of it comes from the playback media itself.
In the case of vinyl records, there’s mastering techniques that have to be used to make sure the needle stays in the groove that some people feel sound better. There’s minute harmonic distortions that people prefer over the perfectly clear and clean sound of digital.
In the case of film payback, you have actual image frames being flashed on the screen. You have shutter dwell where no image is shown at all while the next frame is being advanced. You’d never consciously notice the flicker, but it affects the way you perceive the image and the motion on the screen. Film can have very high contrast and there may also be minor differences in color profile due to being displayed by pigments rather than being created digitally.
Most digital media these days is technically far superior to it’s analog counterpart. It just happens that it’s the limitations and artifacts of the analog media that some people find pleasing.
Yeah, I like vinyl too. But digital to analog conversion is always imperfect. I don’t see that being too fitting for an IMAX cinema, where the aim is just biggest and best, no niche. Aesthetic imperfections are more fitting for arthouse and such I think.
However, I read into it some more now and it’s quite interesting. In the case of Oppenheimer, they actually do manage to avoid the digital conversion for much of the film!
For movies shot on film, all of the film negatives are scanned to digital files so that they can be edited using AVID, and the process continues as before. The finished movie can then be “printed” back on to physical film using a laser scanner, which is how most film prints are made these days. However, some filmmakers like Christopher Nolan refuse to use this method, because it doesn’t allow you to take full advantage of the resolution of IMAX film. So in Nolan’s case, once the movie is finished, an Edit Decision List (EDL) is created, which contains a text list of all the edit points in the film, and which physical pieces of film negative those correspond to. Then, a person called a Negative Cutter actually physically cuts together and assembles the film negative to create the movie in the analog realm. It’s a very specialized profession - there are only one or two people in Hollywood that still do it!
That’s pretty cool. Of course Christopher Nolan would insist on editing on film. I bet Tarantino and a few others would as well. In the case of laser printing vs optical printing, he’s probably right, especially at IMAX scale. And you’re absolutely right that he’s choosing it because it’s the highest possible fidelity, whereas someone like Tarantino chooses film for sentimental, kitsch, and artistic reasons. Apparently digital IMAX is still only about 4k at best.
In the audio world though, analog DACs really are perfect reproduction many times beyond what is perceptible in sound. Compared to cutting to vinyl from tape, or even transferring from tape to tape, digital is orders of magnitude higher fidelity to the source signal. There’s lots of reasons to love analog audio, but higher measurable fidelity is not among them.
Idk. One benefit is that if preserved, in the future it might allow digital captures of higher resolution. I say might because maybe we already reached the max level of detail you could extract from these type of analog films I do not know.
Depends on the film itself being used. You can get a pretty insane level of detail on the professional-grade stuff. It’s how we’re able to do 4K remasters of stuff from decades ago. Digital still has a fair bit to go before it can fully outshine analogue film.
We have such a thing as an 8K camera. It’s kinda hard to beat an 8K camera. Especially given film has a very low frame rate, so you’re trading one kind of quality for another.
8K is basically more pixels than your ever going to need unless you like staring at a tiny portion of the screen. 24 FPS (films normal FPS fyi) on the other hand is very noticeably different from 60 FPS, nevermind the 120+ FPS some cameras can do.
The reason we have these cameras is partly so we can crop and still get a usable 4K or 6K image.
As for compression we have true lossless and visually lossless formats that are used on professional video cameras. This includes Red Raw.
Did I mention that this kit is cheap enough that Linus Tech Tips - a YouTube channel - can afford to have at least one.
IMAX 1570 film is equivalent to somewhere between 12K and 18K resolution, though as you say that’s only going to matter on a really big display or if you’re creating content for future VR headsets. As for frame rate, more than 24 FPS doesn’t really seem popular in cinema - I remember that Hobbit film getting quite a negative reaction when it tried 48 FPS, and then there’s the Soap Opera Effect. The colour characteristics of different kinds of film also appeal to some artistic filmmakers.
None of those points really justify the significantly more expensive and complicated workflow that comes with analogue film, mind. I do wish digital hadn’t taken over so soon because a whole bunch of media is now stuck with 480p digital as the best it’s realistically ever going to look barring AI upscaling, but I can’t say I really blame the producers for making that decision because digital is far less of a pain in the ass.
Nope not even display size can save this, as you need to sit some distance to see the whole thing, unless you only want to look at a portion. The human eye has a finite effective resolution.
VR it could maybe make a difference, even then 8K is good compared to most modern VR devices.
VR possibly needs to get as high as 16K to be truly perfect for the human eye, so it could be useful there. But 16K headsets are a long way off, and refresh rate also needs to improve since the human eye can perceive up to 2,000 frames per second, so I feel like refresh rate will be the priority over pixel density by the time we get even 8K headsets.
While some of what you said is questionable, I do wish more movies were shot in 60+ FPS. Some people clams to prefer 24 FPS, but I honestly think that’s just because we’re so used to it.
What have I said is actually questionable? 8K is already beyond the limit of what a human can see at any practical viewing distance. Super high end film might beat it in equivalent resolution, but at that point it’s purely academic because you’re eyes can’t see that much detail.
Unless you want to blow up a small portion of the image to poster size… then I am sure IMAX film is better. Just not sure why you would want that.
Meanwhile in framerate you can see a fairly large difference between the two technologies.
Film is still higher quality than TVs these days. There’s a reason it’s easy to remaster and rereleasr classic content shot on film than more recent content shot digitally.
To store digitally you would need a compression algorithm. Pretty much all video compression algorithms are lossy, which means you automatically lose detail.
Storing an uncompressed video isn’t feasible as each frame could be hundreds of megabytes (or more) in size. This is due to resolution + color info + audio channels.
Even uncompressed, they would be large, but not unfeasible. Even assuming about 12 MB per frame (reasonable for lossless 4k) that gives us about 1TB per hour. Using lossless video compression would push that smaller. That’s very large for consumers, but not for a film studio. I’m certain a few terrabytes Iof storage are way cheaper than that much film.
So more like 20TBs. And again, that’s lossleslly compressing the individual images, but not the video. The video is still uncompressed. Lossless video compression would significantly reduce that.
It is a huge file, but it’s just as tractable as that film.
An uncompressed 4K frame with 16-bit color is about 50 MB. An uncompressed 18K frame with 16-bit color is just over 1 GB.
I don’t disagree with you that lossless 18K video storage is trivially easy—digital storage is shockingly cheap these days—but I’m curious where you’re getting those numbers from. Compressing an hour of 18K video from 87 TB to 20 TB seems like a remarkable feat.
I was using lossless compressed image sizes becuase they were relatively easy to find. So those 4K 16-bit frames were more like 12 MB instead of 50. That’s where the compression comes from. Lossless image compression details were much easier to find than losses video compression details, and I could test them myself easily. The 12 MB will depend on the original image, as some compress much more readily than others, but it’s reasonable.
I’m sure I’m wrong, but it’s hard to imagine this being better quality than what we can do digitally these days.
You are in fact wrong lol. Actual film has a resolution equivalent of something like 18K.
Wasn’t normal 35mm film about the equivalent of somewhere between 4k and 8k depending on the film stock?
Plus, the projector optics will always limit the sharpness of the picture. No lense is ideal, and even ideal lenses would have fundamental limitations due to diffraction.
Something like that.
As far as lens optics, we’re really splitting hairs here. 70mm through a quality lens in an imax theater is going to look absolutely fantastic and stunning. Digital is just more convenient and at some point it will catch up and surpass film.
My point was more like that even IMAX film doesn’t quite get to 18k equivalent, more like 12 to 16k. Honestly, anything above 4k (for normal widescreen content) even on big screens is barely noticeable if noticeable at all. THX recommends that the screen should cover 40° of your FOV; IMAX is what, 70°, so 8k for it is already good enough. Extra resolution is not useful if human eye can’t tell the difference; it just gets to the meaningless bragging rights territory like 192 kHz audio and DAC-s with 140 dB+ S/N ratio. Contrast, black levels, shadow details, color accuracy are IMO more important than raw resolution at which modern 8k cameras are good enough and 16k digital cameras will be more than plenty.
The extra resolution isn’t completely useless from an editing standpoint.
If you’re working with 16k footage and a 4K deliverable and the shot isn’t quite right you can crop up to 75% of the image with no loss in quality.
This kind of thing would be mostly useful for documentaries, especially nature, or sports where you can’t control the action.
Yup that’s why people can go back and rescan old film movies to make them into 4k now that we have better cameras, but you can’t do that with movies that were recorded with digital
Yeah, we’ll have this brief digital gap from the era when film was going out of fashion and 4k and higher resolution digital cameras weren’t a thing yet. But now that even average youtubers are shooting 4k with cheap(ish) DSRL-s, we generally don’t have to worry about the content having “not good enough quality for the future”.
The bigger problem IMO is the ephemeral and profit-driven nature of modern content distribution. Once the studio decides a film/series is not making enough money and pulls it from streaming, it’s gone. IIRC, DRM of DCP is also remotely managed so even if a cinema physically has the drive with the movie, they can’t play it when the studio pulls the plug–this was not the case with film.
Yeah all that is a huge problem, I remember Microsoft pulled the game Scott pilgrim from the Xbox 360 so if you didn’t buy it beforehand you couldn’t get it anymore until they did some legal stuff to get the game back in the store.
I still think film today is a great tool for getting high resolution photography at a cheap entry cost, a full sized digital sensor camera can be pretty pricey where as a 35mm film camera can be had pretty easy, then once you go to medium format it’s gets more expensive and then I’m not even sure there is large format digital cameras
Speaking of projectors, don’t they max at something like 100 nits?
The JVC DLA-NX5 I have the pleasure to have set up for demos at work is 1800 lm, or 525 nits. Plenty bright, HDR looks amazing on it.
I don’t think there’s any reason we couldn’t make a store 18k video.
And we could make screen at much higher resolutions that that at imax size, or even quite a bit smaller, though I suspect it would be absurdly expensive.
It is technically possible but you’re right it would be absurdly expensive.
Storing it isn’t the problem, you’ll still need to be able to record and project at that resolution.
As I said I’m sure we could make screens that could do that. They would be absurdly expensive and heavy and stupid, but it could be done. Not worth it though.
And it looks like at least 16k cameras have been made.
https://youtu.be/oIhCyPaDP6g
The screens aren’t the problem. It’s often the hardware driving it. The current top generation of gaming gpus struggles at 8k. There’s very little chance of being able to render and play 16/18k
Rendering video and rendering games are pretty different. Video is generally easier especially once it’s mixed down.
Not enough to justify using such a proprietary technology
Not like the off the shelf stuff you can get to store and show 18K.
deleted by creator
Resolution and color reproduction is still unmatched. Plus there are a lot of things happening in the analog domain that our eyes notice as beautiful.
Same thing is true for analog vs digital music production btw
I can’t speak for video, but for audio production that isn’t true. Audio signals can be perfectly reproduced, up to some frequency determined by the sample rate and up to some noise floor determined by the bit depth, digitally. Set that frequency well beyond that of human hearings and set that noise floor beyond what tape can do or what other factors determine, and you get perfect reproduction.
See here. https://youtu.be/UqiBJbREUgU
deleted by creator
I don’t know if perfect reproduction necessarily sounds better. It’s probably subjective, but projects I’ve worked on that were tracked with tape have a quality that you can’t get from digital. I’m not talking about tape hiss or anything like that. There’s a roundness to the sound.
True! Analog can distort the audio in a way some people like.
But, it is a distortion. It’s not there in the original audio. Sometimes, that is desired though.
Valves and vinyl still sound better
Some people do like the distortion that analog audio provides, that’s true. But it is because of something that wasn’t in the original audio. It’s an artistic choice.
Yeah, but they’re likely digitally editing it all now, so it loses that in the middle of the process. Can’t really see why it would make sense to print a digital file back onto film.
EDIT: I did some reading, some movies have a solution for this!
Despite being what a lot of analog fetishists focus on, a purely analog editing chain is not necessarily what causes the effects that they actually like about analog media. Most of it comes from the playback media itself. In the case of vinyl records, there’s mastering techniques that have to be used to make sure the needle stays in the groove that some people feel sound better. There’s minute harmonic distortions that people prefer over the perfectly clear and clean sound of digital. In the case of film payback, you have actual image frames being flashed on the screen. You have shutter dwell where no image is shown at all while the next frame is being advanced. You’d never consciously notice the flicker, but it affects the way you perceive the image and the motion on the screen. Film can have very high contrast and there may also be minor differences in color profile due to being displayed by pigments rather than being created digitally. Most digital media these days is technically far superior to it’s analog counterpart. It just happens that it’s the limitations and artifacts of the analog media that some people find pleasing.
Yeah, I like vinyl too. But digital to analog conversion is always imperfect. I don’t see that being too fitting for an IMAX cinema, where the aim is just biggest and best, no niche. Aesthetic imperfections are more fitting for arthouse and such I think.
However, I read into it some more now and it’s quite interesting. In the case of Oppenheimer, they actually do manage to avoid the digital conversion for much of the film!
That’s pretty cool. Of course Christopher Nolan would insist on editing on film. I bet Tarantino and a few others would as well. In the case of laser printing vs optical printing, he’s probably right, especially at IMAX scale. And you’re absolutely right that he’s choosing it because it’s the highest possible fidelity, whereas someone like Tarantino chooses film for sentimental, kitsch, and artistic reasons. Apparently digital IMAX is still only about 4k at best. In the audio world though, analog DACs really are perfect reproduction many times beyond what is perceptible in sound. Compared to cutting to vinyl from tape, or even transferring from tape to tape, digital is orders of magnitude higher fidelity to the source signal. There’s lots of reasons to love analog audio, but higher measurable fidelity is not among them.
Idk. One benefit is that if preserved, in the future it might allow digital captures of higher resolution. I say might because maybe we already reached the max level of detail you could extract from these type of analog films I do not know.
Depends on the film itself being used. You can get a pretty insane level of detail on the professional-grade stuff. It’s how we’re able to do 4K remasters of stuff from decades ago. Digital still has a fair bit to go before it can fully outshine analogue film.
We have such a thing as an 8K camera. It’s kinda hard to beat an 8K camera. Especially given film has a very low frame rate, so you’re trading one kind of quality for another.
8K is basically more pixels than your ever going to need unless you like staring at a tiny portion of the screen. 24 FPS (films normal FPS fyi) on the other hand is very noticeably different from 60 FPS, nevermind the 120+ FPS some cameras can do.
The reason we have these cameras is partly so we can crop and still get a usable 4K or 6K image.
As for compression we have true lossless and visually lossless formats that are used on professional video cameras. This includes Red Raw.
Did I mention that this kit is cheap enough that Linus Tech Tips - a YouTube channel - can afford to have at least one.
IMAX 1570 film is equivalent to somewhere between 12K and 18K resolution, though as you say that’s only going to matter on a really big display or if you’re creating content for future VR headsets. As for frame rate, more than 24 FPS doesn’t really seem popular in cinema - I remember that Hobbit film getting quite a negative reaction when it tried 48 FPS, and then there’s the Soap Opera Effect. The colour characteristics of different kinds of film also appeal to some artistic filmmakers.
None of those points really justify the significantly more expensive and complicated workflow that comes with analogue film, mind. I do wish digital hadn’t taken over so soon because a whole bunch of media is now stuck with 480p digital as the best it’s realistically ever going to look barring AI upscaling, but I can’t say I really blame the producers for making that decision because digital is far less of a pain in the ass.
Nope not even display size can save this, as you need to sit some distance to see the whole thing, unless you only want to look at a portion. The human eye has a finite effective resolution.
VR it could maybe make a difference, even then 8K is good compared to most modern VR devices.
VR possibly needs to get as high as 16K to be truly perfect for the human eye, so it could be useful there. But 16K headsets are a long way off, and refresh rate also needs to improve since the human eye can perceive up to 2,000 frames per second, so I feel like refresh rate will be the priority over pixel density by the time we get even 8K headsets.
While some of what you said is questionable, I do wish more movies were shot in 60+ FPS. Some people clams to prefer 24 FPS, but I honestly think that’s just because we’re so used to it.
What have I said is actually questionable? 8K is already beyond the limit of what a human can see at any practical viewing distance. Super high end film might beat it in equivalent resolution, but at that point it’s purely academic because you’re eyes can’t see that much detail.
Unless you want to blow up a small portion of the image to poster size… then I am sure IMAX film is better. Just not sure why you would want that.
Meanwhile in framerate you can see a fairly large difference between the two technologies.
Film is still higher quality than TVs these days. There’s a reason it’s easy to remaster and rereleasr classic content shot on film than more recent content shot digitally.
deleted by creator
To store digitally you would need a compression algorithm. Pretty much all video compression algorithms are lossy, which means you automatically lose detail.
Storing an uncompressed video isn’t feasible as each frame could be hundreds of megabytes (or more) in size. This is due to resolution + color info + audio channels.
Many lossless video compression algorithms exist.
https://en.m.wikipedia.org/wiki/List_of_codecs#Lossless_video_compression
Even uncompressed, they would be large, but not unfeasible. Even assuming about 12 MB per frame (reasonable for lossless 4k) that gives us about 1TB per hour. Using lossless video compression would push that smaller. That’s very large for consumers, but not for a film studio. I’m certain a few terrabytes Iof storage are way cheaper than that much film.
4K isn’t nearly enough resolution to compare to IMAX. Plus, I assume your calculations are for 8bit color. To hang with film. Would need to be 12bit
The color was in fact 16 bit.
If you want 18k, multiple 1TB by (18/4)^2 ish
So more like 20TBs. And again, that’s lossleslly compressing the individual images, but not the video. The video is still uncompressed. Lossless video compression would significantly reduce that.
It is a huge file, but it’s just as tractable as that film.
An uncompressed 4K frame with 16-bit color is about 50 MB. An uncompressed 18K frame with 16-bit color is just over 1 GB.
I don’t disagree with you that lossless 18K video storage is trivially easy—digital storage is shockingly cheap these days—but I’m curious where you’re getting those numbers from. Compressing an hour of 18K video from 87 TB to 20 TB seems like a remarkable feat.
I was using lossless compressed image sizes becuase they were relatively easy to find. So those 4K 16-bit frames were more like 12 MB instead of 50. That’s where the compression comes from. Lossless image compression details were much easier to find than losses video compression details, and I could test them myself easily. The 12 MB will depend on the original image, as some compress much more readily than others, but it’s reasonable.