@ DoomBot - When it comes to Heavy Metal, send it to me whenever you get a chance (no rush) and I'll see what can be done with it. No promises, besides the fact that I will try
@ ALL - ok now... back to topic
It's been a long time since I've seen this film, so I'm actually glad that the syncing on this was needed because it reminded me of this flick and the fact that I should get a copy of it. I was kinda confused about the audio though...
What was sent to me for the BD reference audio, was a 5.1 DTS-HDMA track but the US BDs (from 2009 and the 2013 "steel-book" edition) only contain DolbyTrueHD... hmmm... After doing some more research, it turns out that the 5.1 DTS-HDMA audio track is contained on the 2012 "Special Collector's Edition" Bluray from JAPAN(!)
I'm sure that some of you are already aware of this, but I wanted to mention it because:
- the JPN 5.1 DTS-HDMA track sounds different than the 5.1 DolbyTrueHD track (and I will include comparison observations)
- the JPN BD has a different run-time than the US BDs (both the 2009 and 2013 US BDs have the same run-time)
- wanted to inform those that weren't aware of this already
So, before I get to the LD audio, let me get the comparison between the US DolbyTrueHD and JPN DTS-HDMA tracks out of the way first...
The two audio tracks are from, what appears to be, the same master (they sync perfectly to each other, aside from the fact that the US BDs start 2100ms LATER) BUT, the DolbyTrueHD audio track is very "flat" and lifeless. The dialog in both sounds decent, but things in the background sound VERY muffled on the DolbyTureHD audio track. I also noticed a lack of directionality (almost like the audio was upmixed from a MONO source) of the sfx in some of the scenes (mainly when there are lots of gunshots occurring) The overall volume of the JPN DTS-HDMA track is a LOT more consistent than the US BDs. Even taking into consideration that Michael Mann's films usually have inconsistent volume levels (The Last of the Mohicans and Miami Vice come to mind here) the US BDs are excessively inconsistent compared to the JPN BD. If I were to recommend an official release (based on AQ) I would have to go with the JPN BD hands-down. In terms of video quality (for those that are interested) the JPN BD is ALSO superior, with a LOT LESS compression artifacts AND more consistent color palette, is encoded using AVC, unlike the US BDs that still use the older VC1 codec
Now that I have established that the JPN DTS-HDMA track is superior amongst the BD releases, let's look at it compared to the LD PCM audio track...
The first thing I noticed when I downmixed the JPN DTS-HDMA audio to a stereo track, is that the waveform was nearly identical to the LD audio track which I have only seen once before. This did make the syncing work a LOT easier.
When comparing the way the two sound on my laptop and my HDTV, I was not able to hear nearly ANY difference between the two. Listening on pro-headphones, I immediately noticed that the LD audio has a "richer," "fuller" sound, but the difference was not as noticeable as some other LD tracks I have compared to their BD counterparts. HOWEVER, once I started listening to the two on my YAMAHA surround sound system, the difference became a LOT more noticeable (although still not as much as some others I have compared)
The LD audio track has better fidelity and a much wider dynamic range. Dialog has a much crisper sound as well. The BD audio appears to have been filtered a little bit for noise, in this case unnecessarily, which results in some of the background sfx being a little muffled at times (although you will NEVER notice it unless you do a side-by-side comparison) On a couple of occasions, the panning of the sfx (specifically, the "echoes" of gunshots) seems to be reversed on the BD audio (where what should be on the right is on the left and vice-versa; this is true with the TrueHD and HDMA tracks) The gunshot sfx have a "punch" to them on the LD audio, that the BD audio lacks completely. (almost like listening to something with "bass boost" on versus it being off)
Now, onto the syncing "changelog" notes I took and extra comments:
Whenever I sync one audio track to another (have been doing this for years, with more audio tracks synced that I can count) my general "rule of thumb" is a +/-10ms tolerance before I start to make ANY changes. This is a "safe" tolerance, because it's technically 1/4th of a SINGLE frame when dealing with a 23.976fps video source. Any difference larger than that is where I start making edits. (There have been times where a 20ms correction "falls" on a different frame after the edit and can be noticeable [at shot/scene changes] which is why I decided to go with +/-10ms) Although, if I see the audio I'm syncing fall behind by 20-25ms when there is MUSIC playing, only to "auto-align" itself back with reference track within 30-60 seconds, I usually leave it "as-is" because cuts made during music can be very difficult to make without it being obvious. (yes, it can still technically be done, but involves a lot more time and playing some "games" and YMMV based on project)
When I make notes for myself about the edits I've made, I count the amount of cuts that are LESS than a frame (10-35ms) and make a note of it WITHOUT writing down a timecode (seems pointless) IF there is an edit of a SINGLE FRAME or more (so +/- 40ms) I write down the timecode and duration of changes applied. So, here are those notes:
When syncing the "Side 1" LD, after aligning the beginning to the BD (had to remove the WB home video intro from the LD) I only had to make six minor re-cuts to keep things in-sync. (they were between 10-25ms)
At the side change, from "Side 1" to "Side 2" there was audio MISSING from 56m 01.469sec to 56m 02.878sec (total of 1.409seconds) on the LD which I HAD to replace with my downmixed DTS-HDMA track because otherwise there would have been a VERY obvious audio drop-out. The splice is seamless and there is no noticeable change in quality (due to the fact that this is at a scene change)
When syncing the "Side 2" LD, there were seven minor re-cuts to keep things in-sync (they were between 10-25ms) and 3 "major" cuts, plus one "honorable" mention:
- 1h 03min 44.5sec is where I had to completely re-cut the entire sequence. This is where the missing "...detritus..." line occurs. I cut the LD audio just before the "extra" dialog is spoken and then I cut the audio right after. I then aligned the latter part until the dialog re-syncs back with the BD, while crossfading the music into itself from before and after the cut. This gave me a decent sounding, seamless transition between the two parts, while keeping the music properly in-tact throughout (although it sounds slightly different on the BD for those couple of seconds) Then, I exported just the missing line of dialog (with a half second extra before and after the line is spoken to make for easier filtering) and loaded that into DiamondCut Forensics8. I tinkered with it for a while, until I managed to remove nearly all of the background music during this line of dialog, while still keeping the fidelity and crispness of the dialog. It was frustrating because, at first, it kept giving me a "crystalized" sort of sound until I realized that I might want to try a filter that gives me more precise control over the alterations in the frequency spectrum. I won't bore anyone with the technical details here, LOL... After I achieved the result I was looking for, I had to figure out exactly where to insert the line back in. This was VERY tricky as well. I had to add it in after the shot changes to showing her back (otherwise you would hear her talking while her lips are not moving, acceptable with Kung-Fu Dubs, but not here) BUT before the next line of dialog which could NOT be moved to make things easier, because then there would be a lip-sync issue as well. This gave me a VERY narrow "playing field." The first time I inserted the line, it sounded like the next line after it was extremely rush and came out very un-natural. (Funny because the line where the BD audio resumes was actually the one that sounded like it was spliced in) So I got a copy of the JPN BD, threw it up on a timeline in Sony Vegas and figured out the exact timecode of when the shot changes and inserted the line there. The result is decent, seamless, but the line after the "...detritus..." still seems like it's slightly too soon but nothing can be done about it. (the way she speaks in that scene, there is between 0.800-2.500seconds of "silence" between her lines of dialog, but the silence between the "...detritus..." line and the next is only .380 seconds)
- HONORABLE MENTION: 1h 18min 35.6sec is where the second instance of the BD audio being censored/altered occurs. The BD is missing Pacino's sinister laugh, followed by the line "Ferocious, aren't I?" No visual lip-sync issues here and nothing had to be altered by me
- 1h 35min 06sec is where I had to remove 41ms (one frame) of the LD audio (at scene change) to keep things in-sync. Seamless
- 1h 38min 04sec is where I had to insert 41ms (one frame) of "extra" audio (at scene change) via dupping to keep things in sync. Seamless
The side change from "Side 2" to "Side 3" of the LD audio occurs at 1h 53min 24sec and is seamless.
When syncing the "Side 3" LD, there were three minor re-cuts (between 10-25ms) and one "major" re-cut.
At 2h 12min 07sec, there was 55ms that had to be inserted via dupping, to keep things in sync. Seamless
About one minute before the credits are over, the LD audio ends up ahead of the BD audio by about 35ms but I left this "as-is" because it's during the credits and doesn't matter.
When MUXING the synced LD audio to the BD:
For the 2012 JPN BD, this will be a "direct" mux because the audio was synced to this release.
For the 2009 & 2013 US BDs, there will need to be a DELAY of 2100ms applied when muxing(!)
Naturally, you guys will know where to get and it will be available later tonight
