Posts: 1,225
Threads: 51
Joined: 2019 Oct
Thanks: 943
Given 654 thank(s) in 384 post(s)
Country: 
(2020-11-13, 01:37 PM)TomArrow Wrote: The other question is if 3dB difference in a stereo panorama are noticeable.
I could hear the difference on headphones pretty clearly, that's how I became aware of it the first time I think. But on normal loudspeakers there is too much crosstalk for me to still hear it, or at least I'd have to train myself to hear it probably. Since some of the left channel reaches your right ear and vice versa, so the differences are further minimized compared to headphone usage.
This, incidentally, is how the Reservoir Dogs thing got noticed: headphones. I don't watch films on headphones but others do so it got spotted that way then I checked it and found it as well.
Posts: 1,108
Threads: 26
Joined: 2015 Jan
Thanks: 679
Given 304 thank(s) in 205 post(s)
I noticed it on a Mad Max 2 Laserdisc capture I was syncing once.
Thought it was a one off.
I remember back in the day that if I was connecting stereo equipment up to mono equipment I was told to use the Left Channel output?
This was pre-internet and probably communicated to me through a user manual of some kind.
Maybe this has something to do with it?
Posts: 471
Threads: 16
Joined: 2019 Oct
Thanks: 813
Given 227 thank(s) in 155 post(s)
Country: 
I did stumble upon this issue about 6 years ago with Sheena LD audio (I know, guilty as charged  ).
It was an analog capture and the stereo track sounded basically one sided (left).
Posts: 2,049
Threads: 56
Joined: 2016 Dec
Thanks: 161
Given 1008 thank(s) in 612 post(s)
I remember a while back I watched a YT video about T2's synth score and one of the things that got mentioned was that with the matrixed Dolby Stereo system parts of the music had to be slightly delayed in order for it to stay in the FL/FR channels and not be sent to the centre channel. Not sure if it's relevant to this discussion but possibly the 'lean' is to widen the soundstage as best as possible given the tech limitations
Posts: 1,225
Threads: 51
Joined: 2019 Oct
Thanks: 943
Given 654 thank(s) in 384 post(s)
Country:
(2020-11-13, 09:29 PM)zoidberg Wrote: I remember a while back I watched a YT video about T2's synth score and one of the things that got mentioned was that with the matrixed Dolby Stereo system parts of the music had to be slightly delayed in order for it to stay in the FL/FR channels and not be sent to the centre channel. Not sure if it's relevant to this discussion but possibly the 'lean' is to widen the soundstage as best as possible given the tech limitations
Could be. FWIW, the (home) decoders including the Dolby SDU4 account for a similar bleeding effect by applying a delay to the surround channel (since that's derived through phase manipulation in real time), you probably know that already. Assume it's the exact same thing there.
I also found that the Reservoir Dogs Dolby Stereo mix, at least on my AVR, has one severe matrix decoding / upmixing issue near the start of the film: when the K-Billy DJ starts talking after the scene around the table in the café where Mr. Pink is being a dick about tipping, it starts in the centre channel (which would make sense because it's a guy talking, so it takes it as dialogue) but then it abruptly slams into L+R mid-sentence because music has started playing (Little Green Bag). It was quite jarring to hear. Could just be my system but my guess would be that it's more to do with Dolby MP Matrix quirks.
deleted user
Unregistered
Thanks:
Given thank(s) in post(s)
@ pipefan413 That makes sense. The system doesn't just cleanly separate the channels, but it also steers the content to speakers. I imagine when, say, both channels are identical, it's clear to the decoder that it's supposed to be center. But when you then overlay stereo music, it thinks "aha, stereo content", and widens the field.
Posts: 1,225
Threads: 51
Joined: 2019 Oct
Thanks: 943
Given 654 thank(s) in 384 post(s)
Country:
(2020-11-13, 09:56 PM)TomArrow Wrote: @pipefan413 That makes sense. The system doesn't just cleanly separate the channels, but it also steers the content to speakers. I imagine when, say, both channels are identical, it's clear to the decoder that it's supposed to be center. But when you then overlay stereo music, it thinks "aha, stereo content", and widens the field.
Aye, exactly.
But it sounds especially terrible and not at all intentional in this particular case, hahah.
Posts: 7,153
Threads: 601
Joined: 2015 Jan
Thanks: 1081
Given 1466 thank(s) in 963 post(s)
Country: 
I thought about the whole thing, and my conclusions are:
- most stereo tracks from 1977 are Dolby Surround encoded
- if L and R channels have the same sound, it will be extracted to C, cancelling it from L and R
- if you want the same sound to be heard from L and R but NOT from C, they must have different levels
- 3dB is high enough to fool the DS encoder but not too much for human hearing
- L level is higher probably because most of humans are right-handed (more explanation in the links in first post), or just taken arbitrarily (50/50 chance), or because right level was... well, right? And left was... left with higher level!
- L is used as mono in stereo output to get higher volume from amps
deleted user
Unregistered
Thanks:
Given thank(s) in post(s)
(2020-11-18, 02:58 PM)spoRv Wrote: I thought about the whole thing, and my conclusions are:
- most stereo tracks from 1977 are Dolby Surround encoded
- if L and R channels have the same sound, it will be extracted to C, cancelling it from L and R
- if you want the same sound to be heard from L and R but NOT from C, they must have different levels
- 3dB is high enough to fool the DS encoder but not too much for human hearing
- L level is higher probably because most of humans are right-handed (more explanation in the links in first post), or just taken arbitrarily (50/50 chance), or because right level was... well, right? And left was... left with higher level!
- L is used as mono in stereo output to get higher volume from amps
Number 3 & 4 are not quite correct. The detection isn't based on volume differences I think but rather on how different the left and right signals are. So, the same identical signal with just a volume difference would and will just result in most going towards center and a bit bleeding into the left. Which is what I think I'm experiencing with these tracks.
Heard about the right-handedness before, it's possible. You'd think they'd use volume meters in a professional setting tho.
Posts: 2,049
Threads: 56
Joined: 2016 Dec
Thanks: 161
Given 1008 thank(s) in 612 post(s)
Matrixed Dolby works on phase, in phase=centre, out of phase=surround. Left channel info must only be in Lt, right channel Rt. Stuff that falls inbetween will be steered accordingly, and attenuations required to ensure correct playback levels after decoding.
Given the nature of the format I think there was a lot of tweaking and little tricks in order to stop the soundstage collapsing vs a discrete mix. I wouldn't be surprised if the dB variance was deliberate
|
