Mulholland Drive Cinema DTS

[Turisu]

Criterion obviously felt differently as they "preboosted" their own track for the UHD. Except of course that they did it in such as way as to avoid clipping by reducing peak amplitude and sacrificing dynamic range. I also found this in a review of the R4 DVD:

There is some very noticeable distortion during some of the film’s loud passages, but don’t panic and think your amp’s clipping – it’s recorded that way on the disc, and it appears to be an intentional mix decision.

Suggesting that some "preboosting" may have been applied to this release as well. So it's not an unprecedented idea. The thinking perhaps being that many home theatre set ups are either calibrated or play at a set volume level. Mine isn't calibrated at all but I always use the same volume setting for movies.

For Cinema DTS I always apply the correct LFE gain and if it clips, let it clip. My decider on this was looking at The Dark Knight UHD 5.1 tracks. They're basically the theatrical DTS tracks with some minor differences but the LFE clips in those official releases too. So I figured if it doesn't bother Nolan then it doesn't bother me.

[little-endian]

Except of course that they did it in such as way as to avoid clipping by reducing peak amplitude and sacrificing dynamic range.

Do you mean that they applied some sort of dynamic range compression (and thus boosted other parts while reducing the peaks) or are you implying that an overall level reduction itself already reduces the dynamic range? If one defines it via the total SNR, that is technically true but that theoretical loss should be quite low. As far as I remember, even the TT DR meter hardly gives different results when something is simply amplified or attenuated which is what to be expected (especially with music which only uses relatively small macro dynamic ranges compared to movies, even in case of the yet very dynamic 80s).

For Cinema DTS I always apply the correct LFE gain and if it clips, let it clip.

Here I'd be interested in knowing what is wrong by reducing the other channels instead, if needed. Sure, the total mix will be (a lot) more quiet, but it can turned up in gain at the AVR stage which should only then be a real issue if the SNR becomes noticeably low by doing so.

So, to understand that approach, is it because

- the mix sum would otherwise be perceived as being too quiet (and people hesitant to turn up the gain at their AVRs)?
- the clipping isn't (much) noticeable, acts effectively as some "brute force" dynamic range compression or even adds "desirable distortions"?
- there are real concerns about lost SNR (although I can't imagine that the noise floor would be a problem, especially given that many nowadays sources even come in 24 bit flavours)? Even so, one would still buy that with distortions which could otherwise have been prevented.

Well, as for "hot mixes", Nolan can't be bothered much in general at latest since "Interstellar" and "Tenet" which hardly come to any rest.  . At the end a matter of preference of course, I personally like the "clean" and very dynamic mixes à la "Mission Impossible" or "Air Force One" more.

[Turisu]

Do you mean that they applied some sort of dynamic range compression (and thus boosted other parts while reducing the peaks)

This. The UHD track is basically the Criterion BD track boosted by +3dB with the peaks rolled off so they don't clip. Seems like they figured Lynch's intention was to emphasize the subtle lower sounds in the mix and not to make the bangs louder. Although if that's true I don't know why he couldn't have just had those sounds be louder in the mix in the first place. I believe I read somewhere that it was something to do with limitations imposed by theatrical audio standards but I never looked into this in detail.

Here I'd be interested in knowing what is wrong by reducing the other channels instead, if needed. Sure, the total mix will be (a lot) more quiet, but it can turned up in gain at the AVR stage which should only then be a real issue if the SNR becomes noticeably low by doing so.

There's nothing wrong with this approach. I get it. It's just not the approach I take. I don't like to alter the mix any more than I have to. Some already feel that I alter things too much by slowing down and resampling to achieve sync and DTS-HD compatibility. Also these tracks are intended to be played on a calibrated audio system set to reference so there is an objectively correct volume level which exhibits the track as it is supposed to sound. If they were all adjusted to different volume levels to eliminate LFE clipping then we would lose that entirely.

[Turisu]

The synced tracks are now available. Participating members PM me if interested. There are two:

Mulholland Drive (2001) Criterion BD to Criterion UHD [Unboosted]
Mulholland Drive (2001) Criterion BD to Criterion UHD [3dB Boosted]

The unboosted track needs to be played 3dB louder than normal to be heard as Lynch intended. The boosted track has this gain adjustment baked-in but you may notice some distortion in louder scenes due to clipping.

[BDgeek]

Thanks a lot Turisu!

I’ve been speachlees with your output!
FANTASTIC WORK!!!!

BTW, did you estimate the delay for the SC disc?
I guess after an initial delay adjustment, everything will fall in sync up to the end.

[SpookyDollhouse]

Would love a sync to the SC disc as well, or time for delay so I can remux myself.

[TheLoon]

The recent cap-a-holic comparison showed that the Studio Canal UHD is much superior to the Criterion one, which in comparison, seems to be very much filtered loosing much of the finer grain structure.

Off-topic, but I don't believe this to be the case. The SC image has been artificially sharpened to my eyes, giving the appearance of a finer grain structure.

See this example:
https://caps-a-holic.com/c.php?a=3&x=336...&i=11&go=1

The edge of Rita's red top is razor-sharp on the SC. In my experience, film doesn't resolve at that sort of detail even at 4K. Lots of very strange, elongated line patterns in the grain structure as well - another indication that it's been tampered with. The Criterion looks more authentic to me.

[little-endian]

The SC image has been artificially sharpened to my eyes, giving the appearance of a finer grain structure.

This sentence to me misleadingly suggests that the appearance of the finer grain structure would be the direct result of any sharpening where I don't see how that should happen by the usual "sharpening" by raising the border contrast alone. Rather it could be that SC applied artificial noise afterwards.

Lots of very strange, elongated line patterns in the grain structure as well - another indication that it's been tampered with.

Some counterindication. It should be virtually impossible to arrive at the SC result by only artificially sharpening the Criterion one if you look at the zipper structure of the jacket on the left and on the collar where the second "line" is barely visible on the Criterion.

None of that shall mean that the SC one hasn't been post processed of course but I'd claim that the SC at least also contains a bit more true resolution even if the rest should be artificially baked together.

[TheLoon]

This sentence to me misleadingly suggests that the appearance of the finer grain structure would be the direct result of any sharpening where I don't see how that should happen by the usual "sharpening" by raising the border contrast alone. Rather it could be that SC applied artificial noise afterwards.

Sharpening methods are more varied today than the usual border contrast method though. Some tools go after very fine details that affect grain. But yes, artificial noise may be a possibility too - something very strange indeed about that grain structure in the SC.

Some counterindication. It should be virtually impossible to arrive at the SC result by only artificially sharpening the Criterion one if you look at the zipper structure of the jacket on the left and on the collar where the second "line" is barely visible on the Criterion.

None of that shall mean that the SC one hasn't been post processed of course but I'd claim that the SC at least also contains a bit more true resolution even if the rest should be artificially baked together.

True, reasonable to assume the Criterion has been softened a touch too with that example. Lesser of evils to me though than introducing those weird edges and patterns in the SC.

[SpookyDollhouse]

The recent cap-a-holic comparison showed that the Studio Canal UHD is much superior to the Criterion one, which in comparison, seems to be very much filtered loosing much of the finer grain structure.

Off-topic, but I don't believe this to be the case. The SC image has been artificially sharpened to my eyes, giving the appearance of a finer grain structure.

See this example:
https://caps-a-holic.com/c.php?a=3&x=336...&i=11&go=1

The edge of Rita's red top is razor-sharp on the SC. In my experience, film doesn't resolve at that sort of detail even at 4K. Lots of very strange, elongated line patterns in the grain structure as well - another indication that it's been tampered with. The Criterion looks more authentic to me.

The SC release has not been sharpened at all. It's been better compressed by the guys at Fidelity in Motion. Criterion's compression by Pixelogic/Radius60 is slightly gummier in appearance which is evident on close inspection. SC's is a pin-sharp rendition of grain via a better encode. Nothing about it looks unnatural in appearance. The difference is small but it is there. In fact Criterion's has light evidence of minor filtration but not enough to ruin detail. Interestingly, their Elephant Man Blu-ray has a dose of sharpening applied to it that the regular SC Blu does not have. This is also evident in screenshots.