UHD Review A Few Words About A few words about...™ - Once Upon a Time in the West -- in 4k UHD (1 Viewer)

[OliverK]

The Italian Eagle Pictures 4kult 4K UHD release has not yet been officially announced. There is the potential for encoding the master differently, by using a 100 GB disc. Bill Hunt feels that a bigger bucket for the data would benefit the picture quality, for those who project, or who sit very close to the screen.

There is a Reddit post detailing how other Eagle Pictures 4K UHDs have fared:



Note: regarding "superior encoding" for an identically graded master - I definitely think that this is more a term bandied about by people who sit and put everything through a computer and print out graphic readings of the data file, than having anything to do with what is actually visible on a large screen, in motion. I have purchased several European 4K UHDs, which people have touted as having "superior encoding" - Citizen Kane, Mulholland Drive, Videodrome, Carrie, Wings Of Desire, The Three Colors Trilogy, to name a few, and saw no arresting visual differences, compared with their Criterion counterparts, on a projected image in motion. There might be some very subtle differences in motion artifacts, or how black levels or grain is managed. But these are practically invisible in motion. And certainly not worth all the fuss generated. After all, the real upgrade lies in the rescanning and regrading of the original elements.

Here is a case where one encode heavily and clumsily affects grain structure from the same master:

Invasion of the Body Snatchers • UK Ultra HD Blu-ray vs. US Ultra HD Blu-ray

caps-a-holic.com

Invasion of the Body Snatchers • UK Ultra HD Blu-ray vs. US Ultra HD Blu-ray

caps-a-holic.com

This is a difference that I certainly would call noticeable and it is also much bigger than the difference I have seen for Mulholland Drive that indeed is much smaller and probably not worth the upgrade for most people.

Regarding OUATITW it should not have that much potential to look different judging by what has been said about the Kaleidescape version. From what I read it does not look that different from the UHD disc despite its ca. 50% higher bitrate and therefore I am not expecting huge improvements from a BD100 release unless Eagle Pictures somehow does not use the same master.

 

[tenia]

Note: regarding "superior encoding" for an identically graded master - I definitely think that this is more a term bandied about by people who sit and put everything through a computer and print out graphic readings of the data file, than having anything to do with what is actually visible on a large screen, in motion. I have purchased several European 4K UHDs, which people have touted as having "superior encoding" - Citizen Kane, Mulholland Drive, Videodrome, Carrie, Wings Of Desire, The Three Colors Trilogy, to name a few, and saw no arresting visual differences, compared with their Criterion counterparts, on a projected image in motion. There might be some very subtle differences in motion artifacts, or how black levels or grain is managed. But these are practically invisible in motion. And certainly not worth all the fuss generated. After all, the real upgrade lies in the rescanning and regrading of the original elements.

There are several different aspects related to that.

First and foremost : a good home video release is supposed to transfer as transparently as possible a restoration for home video, aiming for people unable to access the re-release theatrically to be able to enjoy the new work as closely as possible as it was done. Obviously, it's limited to the home video format used for that (BD or UHD), but within the possibilities of each format, it's quite clear that not all releases are equal. As such, if a label can't get right the one thing it's absolutely supposed to do (properly offer a home video conversion of the restoration), is it any good ? It's arguable, hence the fuss. It's also something that is entirely in the hands of the video label, so they have 0 excuse when it comes to not providing what is known to be achievable.

Then, there's how sensitive you are to it. I've seen rubbish BD encodes, they're perfectly perceptible on my 50" Plasma at 2m30 seating distance, I can't imagine them not being so on a larger screen. You'd get cyclic compression making the grain be fine on one frame and rubbishly blocky on the next 3 frames only to pop back up again on the 5th frame, which generates a nasty flicker that's straining for the eye (a recurring issue in France). Or you'll get a handling of the grain, that turns fine grain of a 4K restoration from a 35mm OCN into LEGO-y chunks (a long-term issue on plenty of Criterion discs). Or you'll get plain awful compression all over (like on Criterion's My Own Private Idaho). UHD will add things like chroma issues, as seen on Paramount's Days of Thunder or Potemkine's Red or Saving Private Ryan, on top of the blocky messes above. If tomorrow, someone was to release Walkabout on UHD and encode it through FiM or Hiventy, I have 0 doubt it'll top this mess :
https://caps-a-holic.com/c.php?go=1&a=0&d1=18599&d2=18598&s1=219304&s2=219279&i=15&l=0

And then, there's the question of how you'd score down such issues. I don't think they're all as bad as each other, I mean : this is awful and should never happen :

Three Colors: Blue • US Ultra HD Blu-ray vs. FR Ultra HD Blu-ray

caps-a-holic.com

Three Colors: White • US Ultra HD Blu-ray vs. FR Ultra HD Blu-ray

caps-a-holic.com

This ?
https://caps-a-holic.com/c.php?go=1&a=0&d1=16962&d2=16781&s1=185843&s2=182372&i=5&l=0
It's better on the Warner disc, no doubt (plus, the grain has slightly be filtered on the Criterion disc), but it's unlikely to make a major difference in motion. Still, there is a gradation between both presentations. The same goes for Wings of Desire, though it's more visible :
https://caps-a-holic.com/c.php?go=1&a=0&d1=18179&d2=17953&s1=209221&s2=204765&i=15&l=0

So not all "superior encodes" are that superior, because (fortunately) not all sub-optimal encodes aren't absolutely garbage, but they're all the more appreciated (and perceptible) when someone else drops the ball.

Finally, not all equipments are equal to each other : my plasma tends to dither slightly so attenuate the issues related to blockiness, while VP-ing might tend to do the same.

This being written, and in part because of all the above, there'll tend to be bigger things to look for before looking at an encode, though it pains to see a gorgeous presentation being dumbed down by rubbish compression.

 

[OliverK]

For those who are interested they now have caps of the Paramount UK UHD and BD on caps-a-holic.com:

UK UHD vs previous US BD:

Once Upon a Time in the West • UK 55th Anniversary Edition Ultra HD Blu-ray vs. US Blu-ray

caps-a-holic.com

Once Upon a Time in the West • UK 55th Anniversary Edition Ultra HD Blu-ray vs. US Blu-ray

caps-a-holic.com

Once Upon a Time in the West • UK 55th Anniversary Edition Ultra HD Blu-ray vs. US Blu-ray

caps-a-holic.com

Once Upon a Time in the West • UK 55th Anniversary Edition Ultra HD Blu-ray vs. US Blu-ray

caps-a-holic.com

Looks like the process that they used to obliterate the grain structure was also rather unkind to other parts of the picture like smoke, wires, wrinkles and so on.

So again this looks pretty bad to me but something else is quite interesting:

The Blu-ray certainly seems to look more like film than the UHD version, the difference is quite significant and it also points to the master retaining more textures than Paramount allows us to see on the UHD release:

UK UHD vs UK BD

Once Upon a Time in the West • UK 55th Anniversary Edition Ultra HD Blu-ray vs. UK 55th Anniversary Edition Blu-ray

caps-a-holic.com

Once Upon a Time in the West • UK 55th Anniversary Edition Ultra HD Blu-ray vs. UK 55th Anniversary Edition Blu-ray

caps-a-holic.com

So there is a chance that the Italian UHD release will actually look better than the UHD releases we have gotten so far from Paramount.

For all members who plan to buy the Paramount releases I would strongly suggest to check out the included Blu-ray as clearly it has a less processed look to it.

 

[titch]

Here is a case where one encode heavily and clumsily affects grain structure from the same master:

Invasion of the Body Snatchers • UK Ultra HD Blu-ray vs. US Ultra HD Blu-ray

caps-a-holic.com

Invasion of the Body Snatchers • UK Ultra HD Blu-ray vs. US Ultra HD Blu-ray

caps-a-holic.com

This is a difference that I certainly would call noticeable and it is also much bigger than the difference I have seen for Mulholland Drive that indeed is much smaller and probably not worth the upgrade for most people.

Yes, but do you have both titles yourself? I do. I have both the Arrow Limited Edition 4K UHD box set AND the Kino Lorber 4K UHD and guess what? Projected at 130 inches, the difference is much more subtle, than the screen caps on the computer would suggest. Arrow Video always uses Fidelity In Motion - and they are absolutely the best - but from the same master, the encoding differences aren't actually that visible on screen, in motion.

And don't forget: premiere film prints never looked like this either. I like visible and well-resolved grain, but Robert Harris prefers a more velvety look, akin to what projected prints looked like.

 

[Robert Harris]

Yes, but do you have both titles yourself? I do. I have both the Arrow Limited Edition 4K UHD box set AND the Kino Lorber 4K UHD and guess what? Projected at 130 inches, the difference is much more subtle, than the screen caps on the computer would suggest. Arrow Video always uses Fidelity In Motion - and they are absolutely the best - but from the same master, the encoding differences aren't actually that visible on screen, in motion.

And don't forget: premiere film prints never looked like this either. I like visible and well-resolved grain, but Robert Harris prefers a more velvety look, akin to what projected prints*** looked like.

*** Projected via an absolutely optimal system.

 

[OliverK]

Yes, but do you have both titles yourself? I do. I have both the Arrow Limited Edition 4K UHD box set AND the Kino Lorber 4K UHD and guess what? Projected at 130 inches, the difference is much more subtle, than the screen caps on the computer would suggest. Arrow Video always uses Fidelity In Motion - and they are absolutely the best - but from the same master, the encoding differences aren't actually that visible on screen, in motion.

And don't forget: premiere film prints never looked like this either. I like visible and well-resolved grain, but Robert Harris prefers a more velvety look, akin to what projected prints looked like.

No, not a direct comparison, I only saw the Arrow version and it looked quite beautiful even though we had to make sure that sharpening in the signal chain is off everywhere. I have enough other titles from Kino where I am rather unhappy with how textures are handled so no need for more of them if I can help it.

You are right about premiere prints looking different but erasing and/or manipulating the film grain like it is often done is now supposed to get us closer to that look? What about the other high frequency detail, do we need to reduce that, too or should we not within reason strive to reproduce what is on the negative without adding too much sharpening? That will help a lot with grain visibility and result in a more filmic image without the need to reduce film grain after the master has already been produced.

 

[OliverK]

I have purchased several European 4K UHDs, which people have touted as having "superior encoding" - Citizen Kane, Mulholland Drive, Videodrome, Carrie, Wings Of Desire, The Three Colors Trilogy, to name a few, and saw no arresting visual differences, compared with their Criterion counterparts, on a projected image in motion. There might be some very subtle differences in motion artifacts, or how black levels or grain is managed. But these are practically invisible in motion. And certainly not worth all the fuss generated. After all, the real upgrade lies in the rescanning and regrading of the original elements.

Why did you get so many of these movies twice - did you HAVE to have them knowing they were better or did you expect more of a difference each time?

I am only familiar with Citzen Kane and Mulholland Drive and the better encodes at the time did not seem to warrant another purchase, the difference did not seem big enough.

 

[Robert Harris]

No, not a direct comparison, I only saw the Arrow version and it looked quite beautiful even though we had to make sure that sharpening in the signal chain is off everywhere. I have enough other titles from Kino where I am rather unhappy with how textures are handled so no need for more of them if I can help it.

You are right about premiere prints looking different but erasing and/or manipulating the film grain like it is often done is now supposed to get us closer to that look? What about the other high frequency detail, do we need to reduce that, too or should we not within reason strive to reproduce what is on the negative without adding too much sharpening? That will help a lot with grain visibility and result in a more filmic image without the need to reduce film grain after the master has already been produced.

The difference between an OCN scan or an IP scan is extremely minimal. So much so that a modern Eastman print derived from a dupe neg, derived from a quality IP, derived from the OCN should have hardly perceptible differences in projection. The difference should be softer edges on the halide crystals.

One can achieve the same appearance either scanning an IP or very slightly softening an image. This has nothing to do with de-graining. The only problem in scanning an IP, which I know is preferred by some filmmakers, is financial, as there is no 4k+ scan of the OCN as a digital archival element.

The most important attribute of QUALITY digital work, is that it go unnoticed.

As an example, I submit MFL. There was no need to soften the image as the original 8k scans had been brought down to 4k, and sharpening which we did to equalize grain structure between OCN and masters was limited to +7%.

The entire final 4k image had the appearance of an original 70mm print.

 

[Dave H]

Here are some screencap comparisons for Once Upon A Time in the West - UHD BD vs remastered BD.

Interesting how the remastered BD shows film grain that is not present on the UHD BD.

Once Upon a Time in the West • UK 55th Anniversary Edition Ultra HD Blu-ray vs. UK 55th Anniversary Edition Blu-ray

caps-a-holic.com

 

[Dave H]

The Italian Eagle Pictures 4kult 4K UHD release has not yet been officially announced. There is the potential for encoding the master differently, by using a 100 GB disc. Bill Hunt feels that a bigger bucket for the data would benefit the picture quality, for those who project, or who sit very close to the screen.

There is a Reddit post detailing how other Eagle Pictures 4K UHDs have fared:



Note: regarding "superior encoding" for an identically graded master - I definitely think that this is more a term bandied about by people who sit and put everything through a computer and print out graphic readings of the data file, than having anything to do with what is actually visible on a large screen, in motion. I have purchased several European 4K UHDs, which people have touted as having "superior encoding" - Citizen Kane, Mulholland Drive, Videodrome, Carrie, Wings Of Desire, The Three Colors Trilogy, to name a few, and saw no arresting visual differences, compared with their Criterion counterparts, on a projected image in motion. There might be some very subtle differences in motion artifacts, or how black levels or grain is managed. But these are practically invisible in motion. And certainly not worth all the fuss generated. After all, the real upgrade lies in the rescanning and regrading of the original elements.

It hasn't been officially announced, but it is coming.

What projector are you using? I've seen stark differences with some films projected vs others in regards to encoding. Grain rendition, detail, and motion artifacts. On some discs, Dolby Vision encodes are superior vs the HDR10 layer. Encoding is not just about bitrate, but how it's done.

 

[tenia]

The difference between an OCN scan or an IP scan is extremely minimal.

I'm surprised to read that, as from my experience, even just on BD, the difference between a presentation sourced from a 35mm OCN and one sourced from a 35mm IP is visible, which is how one can quite accurately guess, when the info isn't publicly available, what it's sourced from.

One can achieve the same appearance either scanning an IP or very slightly softening an image. This has nothing to do with de-graining.

How the softening would be achieved ? Low-pass filtering the picture ?

Interesting how the remastered BD shows film grain that is not present on the UHD BD.

Once Upon a Time in the West • UK 55th Anniversary Edition Ultra HD Blu-ray vs. UK 55th Anniversary Edition Blu-ray

caps-a-holic.com

I actually wonder how much of this gap is closed by the DV layer. From what it seems : not much, since I haven't seen many feedbacks pointing out this specific point.

On some discs, Dolby Vision encodes are superior vs the HDR10 layer. Encoding is not just about bitrate, but how it's done.

Absolutely. Obviously, there is 0 reason to give oneself extra challenges and having to cram stuff in a limited disc space, but you can target a certain AVB and yield visibly different results. I've even seen BDs having visibly different compression quality at the same instant bitrate.

As for DV encodes, there are reasons why people like FiM don't bother with FEL and why the "DV encodes superior to the HDR10 layer" are from discs with FEL DV layer (and, most of the time, are from authoring houses that, let's say, don't have the best pedigree). In any case, to sum it up, it's mostly (always ?) because the HDR10 layer is sub-optimally compressed, and DV FEV is used as firefighter. It's not even always enough to salvage the whole thing, because that's how bad those doing this can be even in DV, but all this to say it's not so much that DV is, by design, doing so, but because of how it's used by certain authoring houses.

 

[Robert Harris]

Here are some screencap comparisons for Once Upon A Time in the West - UHD BD vs remastered BD.

Interesting how the remastered BD shows film grain that is not present on the UHD BD.

Once Upon a Time in the West • UK 55th Anniversary Edition Ultra HD Blu-ray vs. UK 55th Anniversary Edition Blu-ray

caps-a-holic.com

How does one know that it’s film grain?

 

[Dave H]

How does one know that it’s film grain?

That's a point. Although I wonder why it would it be existent on the BD and not the UHD?

 

[Robert Harris]

That's a point. Although I wonder why it would it be existent on the BD and not the UHD?

Different finalizing from the same base master.

 

[tenia]

How does one know that it’s film grain?

What else might it be ?
I mean : if it's not, what's that on the BD that probably shouldn't be there (digitally-generated grain ? digital noise ?), and if it is, why isn't it on the UHD ?

 

[Robert Harris]

What else might it be ?
I mean : if it's not, what's that on the BD that probably shouldn't be there (digitally-generated grain ? digital noise ?), and if it is, why isn't it on the UHD ?

Interesting questions for each of those involved at every facility at every level of the work-flow.

 

[tenia]

Of course, but one can still do educated guesses.

An easy answer is "film grain seems here to be better reproduced on the BD than on the UHD", but since you seemed to be questioning that, I supposed you have alternate answers. Hence my question, especially because the other answers might just be different issues.

 

[Kyle_D]

I actually wonder how much of this gap is closed by the DV layer. From what it seems : not much, since I haven't seen many feedbacks pointing out this specific point.

I plan on doing a visual comparison next weekend after my FEL-capable player arrives. A BDInfo comparison of the data rates for the base layer and FEL suggests the FEL may be doing a lot of heavy lifting here. When the base layer bitrate drops, the FEL rate goes up and often hits 20+ mbps.
None of that FEL info is showing up in screenshots, nor will it be decoded in any projection set up.

 

[Robert Harris]

I'm surprised to read that, as from my experience, even just on BD, the difference between a presentation sourced from a 35mm OCN and one sourced from a 35mm IP is visible, which is how one can quite accurately guess, when the info isn't publicly available, what it's sourced from.


How the softening would be achieved ? Low-pass filtering the picture ?


I actually wonder how much of this gap is closed by the DV layer. From what it seems : not much, since I haven't seen many feedbacks pointing out this specific point.


Absolutely. Obviously, there is 0 reason to give oneself extra challenges and having to cram stuff in a limited disc space, but you can target a certain AVB and yield visibly different results. I've even seen BDs having visibly different compression quality at the same instant bitrate.

As for DV encodes, there are reasons why people like FiM don't bother with FEL and why the "DV encodes superior to the HDR10 layer" are from discs with FEL DV layer (and, most of the time, are from authoring houses that, let's say, don't have the best pedigree). In any case, to sum it up, it's mostly (always ?) because the HDR10 layer is sub-optimally compressed, and DV FEV is used as firefighter. It's not even always enough to salvage the whole thing, because that's how bad those doing this can be even in DV, but all this to say it's not so much that DV is, by design, doing so, but because of how it's used by certain authoring houses.

Sorry. Should have noted difference OCN vs IP is minimal at 2k. Was rushing to dress for my Bull Run Civil War re-enactment.

 

[tenia]

I plan on doing a visual comparison next weekend after my FEL-capable player arrives. A BDInfo comparison of the data rates for the base layer and FEL suggests the FEL may be doing a lot of heavy lifting here. When the base layer bitrate drops, the FEL rate goes up and often hits 20+ mbps.
None of that FEL info is showing up in screenshots, nor will it be decoded in any projection set up.

That's in part why I'm asking. It seems that often, big FEL layers are doing that, compensating for a sub-par HDR10 layer encode. Looking at how big the FEL layer is here (especially compared to the ridiculously small HDR10 layer), and that caps are most of the time HDR10 only, it might only be part of the minute story. It won't change that the restoration is grain managed, but it might explain the gap between BD and UHD on the caps.

Sorry. Should have noted difference OCN vs IP is minimal at 2k.

You mean minimal difference between 2k scans of those ? Or 2k presentations based on presentations sourced from those, regardless of the scan resolution ?
Again, in my experience, it's also possible to tell when a BD is sourced from a 2K resto of a 35mm IP or a 2K resto of a 35mm OCN, because the difference is visible (for instance, on Fredi Murer's L'âme soeur, scanned and restored in 2K from an IP, which could be guessed by looking at the BD).