2021-01-10, 11:39 PM
(This post was last modified: 2021-01-11, 12:08 PM by pipefan413.)
LDDB lists both LaserDisc releases of this film as having Dolby Surround audio tracks. Similarly, IMDb's tech specs claim it has "Dolby" sound. But it just so happens that I own and have captured both... and I'm inclined to disagree.
At best, Dolby may have provided noise reduction for the recording of the soundtrack, but it sure as hell sounds and looks like mono to me, no Dolby Stereo/Surround. Even that's dubious because there's a huge amount of analogue noise in all versions I've checked (which is almost all of them). The NTSC DVDs (US and AU) do have 5.1, but so do lots of films that originally had mono tracks, and in this case the 5.1 sounds pretty crap and is clearly sourced from the same audio that's on the US LaserDisc, which sounds like 35 mm optical tracks (possibly even 16 mm) that haven't been cleaned up much at all.
The older Japanese LD is analogue-only, which is a bit of a bummer, though it has a better video transfer than the US one. Although the later US LD has digital PCM audio, it sounds pretty awful (I'll get to exactly how in a minute), so it might theoretically be that the JP LD is the better source between them but I'm not convinced of that yet because it's pretty atrocious too. I had to fix 3 particularly problematic moments in the US PCM, and it's very possible that there may be some more I haven't yet noticed:
1. A frame was missing in the middle of a scene with lots of repetitive background noise including electronic bleepy bloopy noises that made it very difficult to hide (in the source PCM it just jumps abruptly, they didn't attempt to fix it). Conversely, in the 5.1 audio from the US DVD, there's another moment where they *did* attempt to cover missing frames but made a hash of it: you hear the noise of a door opening twice in quick succession because of a botched loop in a stupid place. It was very difficult to handle this moment transparently, I'll give them that, but I feel like if an amateur plonker like me can do it then it should have been doable on the actual official DVD release...
2. After the side change there was a massive pop in the waveforms on both channels that I tried my best to correct with iZotope RX, but it's still not perfect so you'll probably notice if you're listening for it. Compared to how it sounded beforehand though it's *significantly* better.
3. I just noticed another huge pop much earlier in the film during a piece of music, which I've also tried to remove to the best of my ability but it's still there (just significantly less unpleasant on the ears).
I started off just resyncing based on the analogue audio track captured with the LD video, but found this to be a little off. I then switched to targeting the DVD's 5.1 track sync, but realised that this too was pretty noticeably off in several places, so I shunted the audio back a frame from there and synced it all visually using AviSynth and VirtualDub, occasionally exporting as a 32-bit float and fixing more delicate problems in RX then importing that back into the AviSynth script. This approach gets really messy fast which is irritating but it means I'm not compromising due to the limitations of any single piece of software (or, one might legitimately argue, my skill level with it) and I'm generally leaving myself fairly specific notes so I can retrace my steps easily enough. In other words, this was a bit of a pain in the arse and it still sounds pretty bad because the source was poor to start with, but there are no good sources for this film AFAIK. I know of only a handful of disc releases (there are VHS releases that I haven't yet investigated, but I don't really expect them to be especially different/better):
1. JP NTSC LD, analogue audio only (packaging claims it's Dolby Surround). Video is widescreen but still cropped at sides somewhat, apart from end credits.
2. US NTSC LD, digital PCM audio (packaging calls it "stereo surround sound"... but interestingly there is no mention of Dolby and no Dolby Surround logo). Video is 4:3 pan & scan, so heavily cropped at the sides, except the end credits where it's in the original ratio because the credits wouldn't work at all if it were cropped this much (or at all, really).
3. 2003 US NTSC DVD, 5.1 Dolby Digital audio and original anamorphic widescreen ratio. Buuuut... the audio is pretty bad here too.
4. 2014 DE (or possibly Austrian?) PAL DVD containing a cut version of the film, which may have been a bootleg of some sort. Don't have a copy, tempted to check but isn't exactly cheap.
5. 2017 AU DVD (supposedly also NTSC), which looks like it's more or less just a straight reissue of the old US one (but I'd like to check if I could find a copy for less than £100!)
The thing is, almost everything lists the soundtrack as being stereophonic, including the end credits:
![[Image: rtncapinv144858.png]](https://i.postimg.cc/D0Dg8wjh/rtncapinv144858.png)
... and yet absolutely nothing I'm actually hearing or seeing as I check the reality of the situation is backing that up. It very much seems like every home release has been created from a single optical mono track from a print, which was in bad shape, even the 5.1 from the NTSC DVDs. What I think might have somehow happened is that because the film was shown "in selected theatres" with Dolby Stereo, but it absolutely tanked, they probably didn't ship many prints with the Dolby Stereo 4:2 matrixed surround mix on them but instead shipped a bunch of mono prints without the Dolby stuff because it was cheaper. Then, when it came to making a home video transfer, one of those mono prints was used as the source, but it was incorrectly assumed to be Dolby Stereo because of the information they had at the time (e.g. the end credits) so that's what went on the packaging, and this was propogated in every further home release.
To test this theory, I did my own matrix decode using the Dolby Pro Logic II software decoder from Cyberlink PowerDVD, and sure enough I got 6 tracks but what was in them seemed to support my hypothesis. The surround channels at the back contained nothing but music from L+R, dialogue and other sounds bleeding from C, and undesirable analogue noise. The 5.1 track off the US DVD isn't anything much more sophisticated, it just seems to be the exact same master thrown through a matrix decoder, which has indeed spat out 6 channels, but they're much like the ones I got from my own test: the surround channels don't seem to contain any surround information, just messed up portions leaking from the fronts. It isn't just the surrounds, either: the left and right channels are full of leaking dialogue that should be in the centre channel if it were actually Dolby Stereo, and the LFE channel is an atrocity in its own way, being about 95% complete silence and 5% horrible woolly residual noise that clearly wasn't designed to represent "low frequency effects" (it's just awful, awful noise, like somebody standing in a hurricane holding a microphone).
On top of the 2:6 upmixing problems described above, the left and right channels on the US LD pretty much just seem to be exactly the same except that they have different sonic characteristics, which is to say that the *content* is identical (i.e. it's monophonic) but the frequency response and noise levels are not. The right channel seems to be the overall winner to my ears, mostly because the left channel has this horrible pulsating quality that was also apparent on the Japanese LD... it's quite hard to describe but it kinda sounds like they've put a really crappy noise gate on the entire thing in a vain attempt to attenuate tape hiss and electrical interference, because every time there is an actual recorded sound, there's this surge of analogue noise along with it but it disappears again when the recording goes quiet. The right channel has no such issue here. This in and of itself would seem really bizarre, but there is actually some precedent for it in my own collection: the US CAV LD release of THE SWORD IN THE STONE claims to have a "restored stereo soundtrack" but it clearly doesn't, instead having one version of the mono (which sounds a bit dodgy) on the left channel, and another version of the mono (which sounds *absolutely atrocious*) on the right channel, both on the analogue and PCM tracks. I think that might be exactly what's going on here, except that it's the other way around: the right channel here sounds ever so slightly less awful than the left, due to it not having that weird pulsating noise the whole way through. If anything, you have almost the inverse effect, which is standard for a monophonic soundtrack recorded without Dolby noise reduction being employed: when the recorded sound kicks in, the tape hiss is knocked down a peg because some of it being sort of pushed out of the way (though not very much of it, since there's very little high frequency activity in Academy mono that would knock it out of the spectrogram). And yeah: the 5.1 track from the US DVD sounds like this as well.
My usual test to check for mono doesn't apply here, by the way, because the left and right channels have completely different EQ. If they were at least properly balanced (in terms of both gain and frequency response), then inverting the phase of one side and then mixing down to 1 channel would cause phase cancellation to occur on almost the entire recording, but if the gain and/or EQ is significantly different then the phase cancellation doesn't work and you're just left with a horrible racket, which is exactly what happens here. This doesn't make me think that it's stereo, just that it's extremely poorly presented mono, as I'm suggesting based on all the other evidence I've just talked about.
But wait, I said "almost" all the info claims that the film is stereophonic. Why? Because the 2014 German DVD allegedly contains an English 1.0 mono soundtrack! However, it probably won't be very useful, given that it's heavily cut (not to mention low-bitrate Dolby AC-3 with PAL speedup).
What I might do is consider the right channel from the US LD the best of a bad bunch and take that as a mono track on its own, excluding that horrid pulsating left channel from both LDs. It's re-synchronised as 2 channels anyway so I can very easily leave it as lopsided 2ch, split them into two separate mono tracks, or discard one of them altogether (most likely the left channel).
At best, Dolby may have provided noise reduction for the recording of the soundtrack, but it sure as hell sounds and looks like mono to me, no Dolby Stereo/Surround. Even that's dubious because there's a huge amount of analogue noise in all versions I've checked (which is almost all of them). The NTSC DVDs (US and AU) do have 5.1, but so do lots of films that originally had mono tracks, and in this case the 5.1 sounds pretty crap and is clearly sourced from the same audio that's on the US LaserDisc, which sounds like 35 mm optical tracks (possibly even 16 mm) that haven't been cleaned up much at all.
The older Japanese LD is analogue-only, which is a bit of a bummer, though it has a better video transfer than the US one. Although the later US LD has digital PCM audio, it sounds pretty awful (I'll get to exactly how in a minute), so it might theoretically be that the JP LD is the better source between them but I'm not convinced of that yet because it's pretty atrocious too. I had to fix 3 particularly problematic moments in the US PCM, and it's very possible that there may be some more I haven't yet noticed:
1. A frame was missing in the middle of a scene with lots of repetitive background noise including electronic bleepy bloopy noises that made it very difficult to hide (in the source PCM it just jumps abruptly, they didn't attempt to fix it). Conversely, in the 5.1 audio from the US DVD, there's another moment where they *did* attempt to cover missing frames but made a hash of it: you hear the noise of a door opening twice in quick succession because of a botched loop in a stupid place. It was very difficult to handle this moment transparently, I'll give them that, but I feel like if an amateur plonker like me can do it then it should have been doable on the actual official DVD release...
2. After the side change there was a massive pop in the waveforms on both channels that I tried my best to correct with iZotope RX, but it's still not perfect so you'll probably notice if you're listening for it. Compared to how it sounded beforehand though it's *significantly* better.
3. I just noticed another huge pop much earlier in the film during a piece of music, which I've also tried to remove to the best of my ability but it's still there (just significantly less unpleasant on the ears).
I started off just resyncing based on the analogue audio track captured with the LD video, but found this to be a little off. I then switched to targeting the DVD's 5.1 track sync, but realised that this too was pretty noticeably off in several places, so I shunted the audio back a frame from there and synced it all visually using AviSynth and VirtualDub, occasionally exporting as a 32-bit float and fixing more delicate problems in RX then importing that back into the AviSynth script. This approach gets really messy fast which is irritating but it means I'm not compromising due to the limitations of any single piece of software (or, one might legitimately argue, my skill level with it) and I'm generally leaving myself fairly specific notes so I can retrace my steps easily enough. In other words, this was a bit of a pain in the arse and it still sounds pretty bad because the source was poor to start with, but there are no good sources for this film AFAIK. I know of only a handful of disc releases (there are VHS releases that I haven't yet investigated, but I don't really expect them to be especially different/better):
1. JP NTSC LD, analogue audio only (packaging claims it's Dolby Surround). Video is widescreen but still cropped at sides somewhat, apart from end credits.
2. US NTSC LD, digital PCM audio (packaging calls it "stereo surround sound"... but interestingly there is no mention of Dolby and no Dolby Surround logo). Video is 4:3 pan & scan, so heavily cropped at the sides, except the end credits where it's in the original ratio because the credits wouldn't work at all if it were cropped this much (or at all, really).
3. 2003 US NTSC DVD, 5.1 Dolby Digital audio and original anamorphic widescreen ratio. Buuuut... the audio is pretty bad here too.
4. 2014 DE (or possibly Austrian?) PAL DVD containing a cut version of the film, which may have been a bootleg of some sort. Don't have a copy, tempted to check but isn't exactly cheap.
5. 2017 AU DVD (supposedly also NTSC), which looks like it's more or less just a straight reissue of the old US one (but I'd like to check if I could find a copy for less than £100!)
The thing is, almost everything lists the soundtrack as being stereophonic, including the end credits:
... and yet absolutely nothing I'm actually hearing or seeing as I check the reality of the situation is backing that up. It very much seems like every home release has been created from a single optical mono track from a print, which was in bad shape, even the 5.1 from the NTSC DVDs. What I think might have somehow happened is that because the film was shown "in selected theatres" with Dolby Stereo, but it absolutely tanked, they probably didn't ship many prints with the Dolby Stereo 4:2 matrixed surround mix on them but instead shipped a bunch of mono prints without the Dolby stuff because it was cheaper. Then, when it came to making a home video transfer, one of those mono prints was used as the source, but it was incorrectly assumed to be Dolby Stereo because of the information they had at the time (e.g. the end credits) so that's what went on the packaging, and this was propogated in every further home release.
To test this theory, I did my own matrix decode using the Dolby Pro Logic II software decoder from Cyberlink PowerDVD, and sure enough I got 6 tracks but what was in them seemed to support my hypothesis. The surround channels at the back contained nothing but music from L+R, dialogue and other sounds bleeding from C, and undesirable analogue noise. The 5.1 track off the US DVD isn't anything much more sophisticated, it just seems to be the exact same master thrown through a matrix decoder, which has indeed spat out 6 channels, but they're much like the ones I got from my own test: the surround channels don't seem to contain any surround information, just messed up portions leaking from the fronts. It isn't just the surrounds, either: the left and right channels are full of leaking dialogue that should be in the centre channel if it were actually Dolby Stereo, and the LFE channel is an atrocity in its own way, being about 95% complete silence and 5% horrible woolly residual noise that clearly wasn't designed to represent "low frequency effects" (it's just awful, awful noise, like somebody standing in a hurricane holding a microphone).
On top of the 2:6 upmixing problems described above, the left and right channels on the US LD pretty much just seem to be exactly the same except that they have different sonic characteristics, which is to say that the *content* is identical (i.e. it's monophonic) but the frequency response and noise levels are not. The right channel seems to be the overall winner to my ears, mostly because the left channel has this horrible pulsating quality that was also apparent on the Japanese LD... it's quite hard to describe but it kinda sounds like they've put a really crappy noise gate on the entire thing in a vain attempt to attenuate tape hiss and electrical interference, because every time there is an actual recorded sound, there's this surge of analogue noise along with it but it disappears again when the recording goes quiet. The right channel has no such issue here. This in and of itself would seem really bizarre, but there is actually some precedent for it in my own collection: the US CAV LD release of THE SWORD IN THE STONE claims to have a "restored stereo soundtrack" but it clearly doesn't, instead having one version of the mono (which sounds a bit dodgy) on the left channel, and another version of the mono (which sounds *absolutely atrocious*) on the right channel, both on the analogue and PCM tracks. I think that might be exactly what's going on here, except that it's the other way around: the right channel here sounds ever so slightly less awful than the left, due to it not having that weird pulsating noise the whole way through. If anything, you have almost the inverse effect, which is standard for a monophonic soundtrack recorded without Dolby noise reduction being employed: when the recorded sound kicks in, the tape hiss is knocked down a peg because some of it being sort of pushed out of the way (though not very much of it, since there's very little high frequency activity in Academy mono that would knock it out of the spectrogram). And yeah: the 5.1 track from the US DVD sounds like this as well.
My usual test to check for mono doesn't apply here, by the way, because the left and right channels have completely different EQ. If they were at least properly balanced (in terms of both gain and frequency response), then inverting the phase of one side and then mixing down to 1 channel would cause phase cancellation to occur on almost the entire recording, but if the gain and/or EQ is significantly different then the phase cancellation doesn't work and you're just left with a horrible racket, which is exactly what happens here. This doesn't make me think that it's stereo, just that it's extremely poorly presented mono, as I'm suggesting based on all the other evidence I've just talked about.
But wait, I said "almost" all the info claims that the film is stereophonic. Why? Because the 2014 German DVD allegedly contains an English 1.0 mono soundtrack! However, it probably won't be very useful, given that it's heavily cut (not to mention low-bitrate Dolby AC-3 with PAL speedup).
What I might do is consider the right channel from the US LD the best of a bad bunch and take that as a mono track on its own, excluding that horrid pulsating left channel from both LDs. It's re-synchronised as 2 channels anyway so I can very easily leave it as lopsided 2ch, split them into two separate mono tracks, or discard one of them altogether (most likely the left channel).
