[Help] Audio Syncing

[deleted user]

Ah gotcha!

[Falcon]

Sony Vegas Pro is the best for syncing audio. Currently i am using this software to sync Cinema DTS. you can shrink and stretch the audio without any quality loss or distortion in pitch.

For audio conversion from 24 to 24 / 1.001 fps, the difference in pitch will not be heard, for me the best solution is to slowdown the samples, for keep samples intact.
I always keep a Flac 44056 Hz copy for my DTS cinema synchronizations.
Like that, if a better blu ray is released, but in 24 fps (like European blu ray sometimes), I accelerate again in 44100 hz so that it becomes again as before 

Obviously, I resample to 48000 Hz for the final file 

[Falcon]

oups double post sorry, please moderate this message

[alleycat]

Just out of interest is it possible to hear the difference between 48hz and 44.1hz? My hearing is so bad I struggle to hear the difference between uncompressed and AC3 but from what I've read the human ear can't hear the difference between 48hz and 44.1hz? Not an attack on anyone wanting to preserve 44.1hz just curious.

[PDB]

If two encodes come from the same robust source and one is sampled as 44.1 and another 48, there is no human on earth that can tell the difference. No one's hearing is that good despite claims otherwise.

It is possible to tell the difference if something was converted from 44.1 to 48 poorly. There are tell-tale signs that give it away. A good encoder would make it impossible to tell the difference.

People here who want the original 44.1, want to have it as original as possible so people don't have to bother with the resample. Bit-perfect capture doesn't matter if you screw up the encode to 48.

Obviously that is not an option when encoding for BDs.

Personally, I can tell the difference between a lossy DD encode and a PCM/MA/True encode when volume matched. Something like EAC-3/Plus vs. lossless is almost impossible for me to tell even with EAC-3 being lossy.

[deleted user]

In a laboratory setting, (some) humans can apparently hear up to 28 kHz, so that would put the limit at 56 kHz sampling rate (x2 because of nyquist). In real life, probably most can't tell the difference. However just for the sake of preservation I would never downsample. If the original has 48 kHz I would consider it sinful to downsample to 44.1 kHZ.

However there's other factors to keep in mind, for example pre- and post-ringing, which I think is longer at lower sample rates, but again may or may not have real world effects on how something sounds. There's a lot of debate surrounding this I believe and some digital-to-analog converters (DACs) even give you various options on how to deal with this issue, supposedly this affects sound as well. I believe one of the methods is to eliminate pre-ringing but at the cost of a higher post-ringing, but I'm not sure how exactly it works. Something to do with phase and whatnot.

Then there's the quality of the sample rate converter. A poor sample rate converter will introduce distortion that many find very noticeable, but I haven't done tests myself. It's certainly very clearly measurable - see https://src.infinitewave.ca/ . As a rule I always try to use a good resampler like iZotope 64 bit SRC whenever resampling is not avoidable to avoid changing how something sounds.

Another factor is the native operating sample rate of the sound card you're playing something on, tying into the last point. If you're playing a 44.1 kHz file but your soundcard is set to 48 kHz, there will be automatic resampling and most likely that will not use a good resampler. So by doing that you're most likely degrading your audio quality. Whether it's audible, I haven't done tests.

Most of this is probably irrelevant for anything regarding artistic quality, since everyone knows the music/creativity/sound design is infinitely more important than any of these. However since we're kinda in it for restoration/preservation purposes I think it's fair to obsess over it a little bit. Even if it can't be clearly noticed, I always wonder ... with all the advancements in AI and technology ... what will some future AI be able to do with a technically perfectly preserved track versus a track that's "only" good enough for human ears.

[alleycat]

In a laboratory setting, (some) humans can apparently hear up to 28 kHz, so that would put the limit at 56 kHz sampling rate (x2 because of nyquist). In real life, probably most can't tell the difference. However just for the sake of preservation I would never downsample. If the original has 48 kHz I would consider it sinful to downsample to 44.1 kHZ.

However there's other factors to keep in mind, for example pre- and post-ringing, which I think is longer at lower sample rates, but again may or may not have real world effects on how something sounds. There's a lot of debate surrounding this I believe and some digital-to-analog converters (DACs) even give you various options on how to deal with this issue, supposedly this affects sound as well. I believe one of the methods is to eliminate pre-ringing but at the cost of a higher post-ringing, but I'm not sure how exactly it works. Something to do with phase and whatnot.

Then there's the quality of the sample rate converter. A poor sample rate converter will introduce distortion that many find very noticeable, but I haven't done tests myself. It's certainly very clearly measurable - see https://src.infinitewave.ca/ . As a rule I always try to use a good resampler like iZotope 64 bit SRC whenever resampling is not avoidable to avoid changing how something sounds.

Another factor is the native operating sample rate of the sound card you're playing something on, tying into the last point. If you're playing a 44.1 kHz file but your soundcard is set to 48 kHz, there will be automatic resampling and most likely that will not use a good resampler. So by doing that you're most likely degrading your audio quality. Whether it's audible, I haven't done tests.

Most of this is probably irrelevant for anything regarding artistic quality, since everyone knows the music/creativity/sound design is infinitely more important than any of these. However since we're kinda in it for restoration/preservation purposes I think it's fair to obsess over it a little bit. Even if it can't be clearly noticed, I always wonder ... with all the advancements in AI and technology ... what will some future AI be able to do with a technically perfectly preserved track versus a track that's "only" good enough for human ears.

How would you rate resampling with either Audacity or eac3to?

[deleted user]

@alleycat Check that link I posted, it has comparisons to all kinds of resamplers, including Audacity and eac3to I believe. With Audacity there are various versions iirc and eac3to seemed decent from what I remember.

[bronan]

Yea, I've been feeding 44.1 khz files from Oppo through my LG TV, then optical out to my receiver. It turns out my TV has been resampling to 48kHz all along and I didnt realize it (had to hook up my capture equipment to my TV and capture it to be sure). Regardless, if you are collecting these tracks for preservation and syncing for future releases its always best to keep it at the original rate (especially when no one cares about making blurays anymore down the line).

[alleycat]

(especially when no one cares about making blurays anymore down the line).

As long as there is a living breath in my body that will never happen!