According to Blah-ray Blogpost, even though the mono mix on this dvd is in a lossy format, it sounds excellent than the Arrow and Polish Pro-Motion blu-rays where despite them being formatted in a lossless codec, it sounds like they took the audio from the Criterion dvd and applied a lot of strong noise reduction to the point where it sounds like it was underwater. The Criterion DVD sounds like it has minimal, if any, changes at all with the noise intact and the audio sounding clear.

However, there has been no update on whether the blu-ray from the Milestone Films Vol. 3 set sounds better or worse than the other blu-rays or the DVD. So for now, the DVD is the best sounding version of the audio I've heard.

I'm having a bit of trouble trying to sync the dvd mono track to the Blu-ray since it seems to desync as the film progresses. I was wondering if anyone wants to sync the audio to the blu-ray, here's the link to the audio track:
https://drive.google.com/file/d/1sCd2OYj...Jwg0Z/view

Spectrograms show that Kino's mono track is identical or very similar to the German MGM BD, which makes sense, considering MGM likely supplied the audio.

Per Blah-ray, the German BD has additional filtering on top of Criterion BD, and Criterion is the least terrible of the bunch.

Although they both sound bad, I can say there is still an audible difference between the two tracks, especially in ambient sound like the constant rain outside. The rain actually sounds very different, e.g. around 1:04:45 (UHD timestamp), with Kino sounding muffled and lifeless in comparison to Criterion.

Notes on the sync:

• Criterion BD is presented in 23.976 fps, while Kino UHD is 24 fps. I've sped up Criterion's audio to 24 fps.
  
• Criterion opens and closes with the MGM lion. Kino opens with United Artist logo instead. Well, I had to let go of that lion roar
  
• Criterion missed a frame in the middle of a shot (how??), between frames #102216, 102217. This is BD timestamps 1:11:03.259-1:11:03.301, or UHD timestamps 1:10:49.467-1:11:49.509.
  
• After accounting for the above, Criterion and Kino's A/V syncs still diverge very often, sometimes by up to 50-90ms. I've tried to compare them, but I found myself to be fairly tolerant of A/V desyncs. So, for simplicity, I've mostly kept to Criterion's sync seeing how nobody has complained in the >10 years since their BD came out.
  

I've been fairly active over at The Star Wars Trilogy forums and have been involved in some restoration work there.

After seeing several links to this site, I thought I'd give it a try.

It seems like there's alot of good posts here, and much valuable technical information.

Glad to be a part of the community!

Moshrom covered this back in the day: https://blah-ray.blogspot.com/2018/01/ci...-1988.html

According to him, the PCM 2.0 mono on the now-OOP 2013 Arrow UK 25th Anniversary BD was basically flawless. Judging from the cover art, I'd guess the 2017 Arrow US BD is probably identical.

Unfortunately, the same can't be said for Arrow's 2020 UHD. It sports a PCM 1.0 track instead of the previous 2.0 mono, but Arrow didn't simply re-encode the previous master as 1.0. Below is a spectrogram of the UHD's mono (my own disc), compared against 2002 Spanish Lauren Films DVD (I came across it on the Russian site). I've tried to match the spectrogram settings on Blah-ray.



And for reference, here's Moshrom's spectrogram of the 2013 Arrow BD:



It's clear that Arrow applied a low-pass filter to their UHD beginning at ~16kHz. Corroborating this is professorwho's post on BR:

professorwho Wrote:The PCM mono track sounds pretty good. I did detect some minor noise reduction at points, but it’s nothing too severe to my untrained ears, at least.

Leaning... leaning... leaning on the everlasting arms on MGM DVD mono... safe and secure from all the crappy Kino Lorber audios.
This is the most detailed track so far, much better than the Kino UHD mono (which is worse than previous compared); DVD MGM NTSC R1 mono synced to Kino Lorber UHD, pm for the track.
Previous info: https://blah-ray.blogspot.com/search/lab...%281955%29
Input: AC3, 2.0 mono, 192 kbps,
Output: FLAC, 2.0 mono, 48 kHz, 16 bits

P.S: As bonus can also share the commentary available from Criterion to Kino UHD, all in sync; Format: FLAC, 1.0 mono, 48 kHz, 16 bits

Hi everyone, I'm excited to be part of this preservation/restoration community
Have a nice day

Hello guys, I'm excited to be part of this preservation/restoration community )

Here is Blah-ray's excellent coverage: https://blah-ray.blogspot.com/2017/01/dekalog-1989.html

To summarize:

• The Criterion BD is pitched too low, because Criterion had to present the series at 24(23.976)fps for the US market.
  
• The Arrow BD is presented at the original 25fps. However, their audio is inexplicably pitched too high.
  
• The old Polish DVD is perfectly pitched. Plus, it has less filtering.
  

So this is totally something that has sort of fascinated me for a bit and I was wondering if anyone has ever managed to get working?

So apparently, certain audio codecs have the compatibility of including two separate LFE channels.

What's more, there's clearly documentation of it being coded into FFMPEG as part of the added support for 22.2 mixes—as insane as that sounds.

Now whilst I understand that there is the whole argument about whether or not this is something that'd even be worth the effort given low frequencies aren't particularly "directional" so-to-speak, I still at least like the prospect of its existence in a codec since it technically opens the door to properly supporting stuff like Dolby "Baby Boom" stuff which had dual LFEs.

Obviously, the opportunity of preserving something like that is practically nonexistent here, but I'm more curious about whether or not it can even work more than anything else at this point.

So I was spending a fair chunk of my afternoon testing different channel layout combos in FFMPEG to see if I can get it working, but for whatever reason, I can't seem to get it working whatsoever.

Maybe I'm doing something wrong with my code, but for some reason every time I try introducing the LFE2 in AAC, I get an error saying "Channel 3 cannot be referenced twice", so I assume the software's interpreting LFE2 as LFE still or something? I even tried trying to do a test creation of a 22.2 track—since it's the only standard channel layout that includes this mysterious channel—and not even that works.

I don't know. I don't suppose anyone else is willing to give this a shot on their end for experimentality's sake?

A hidden gem of 70s political thriller on conspiracy theories and paranoia. Shot by Gordon Willis, the same cinematographer who shot The Godfather two years prior. Highly recommended.

A shame that Criterion filtered away all the bass: Everything below ~70Hz is gone. Moshrom posted an excellent clip as usual: https://blah-ray.blogspot.com/2021/08/th...-1974.html

I've synced the unfiltered track from AMZN WEB-DL to the Criterion BD. Waveform and spectrogram match what Blah-ray showed. In particular, notice how the low frequencies are absent on the Criterion BD's spectrogram, but full of life on the WEB-DL.

Notes on the sync:

• The WEB-DL looks quite different from the Criterion BD, notably in color timing. I would guess it was encoded from an old DVD master.
  
• Frame wise, there are only two diverging points. The first is WEB-DL contains an extra frame (#129883).
  
• The second is a minor editing error on Criterion's part. During the epilogue, their frame #141582 is a duplicate of the next frame (#141583). The only difference is that it's not color-corrected and has a green tint that's not present from the next frame until the end of the sequence. Easily visualized here: https://slow.pics/c/F49FoV2L. This frame is absent on the WEB-DL. If you're interested, I've described this in more detail on BR.
  
• Again during the epilogue, the dialog is out of sync by > 1 second when the committee announced its findings. This is the case for the Criterion BD, AMZN WEB-DL, and apparently the recent Imprint BD as well. At first I thought the dialog may have been dubbed in post. But after applying a -1300ms delay, it was clear that the dialog matched lip movements. So this must be an error that has gone back for years. I've now fixed it.