Posts: 7,151
Threads: 601
Joined: 2015 Jan
Thanks: 1081
Given 1497 thank(s) in 969 post(s)
Country:
2018-12-21, 10:35 PM
Got to choose between two sources of the same movie for a restoration, one at 24fps and the other at 23.976fps.
I'd go with 24fps, because this version has two commentary tracks in AC-3, plus the lossless movie track in its original language; at the end, tracks from PAL DVD should be converted anyway, and *I think* 25->24 *should* lead to less "errors" than 25->23.976.
But, in other cases, if all audio tracks would be equal, I'd go with a 23.976fps version, because it seems that 99% of BDs (and WEB) versions follow this, AND it seems that some hardware players have some problems playing 24fps discs.
Opinions?
Posts: 459
Threads: 29
Joined: 2015 May
Thanks: 69
Given 85 thank(s) in 71 post(s)
Country:
deleted user
Unregistered
Thanks:
Given thank(s) in post(s)
24fps is what it plays as in cinema. It's also what Cinema DTS discs come in, timing-wise. But most Blu Rays and DVDs are synced to 23.976 (or rather, 24000/1001).
So it comes down to which audio you want to keep lossless from your sources. Maybe the best thing to go by is your lossy audio sources.
Let's say your main audio is lossless (DTS-HD MA or TrueHD), but your additional audio (which for example could be 24fps synced) is AC-3 and thus lossy, it may be actually less of a loss to retime your lossless audio to 24fps (because it's lossless to begin with and will also be delivered lossless, so your only loss is from the sample rate conversion), while in order to convert the AC-3 to 23.976 would either require wasting space on delivering it in a final lossless form (to not introduce more loss) or introduce generational loss by reencoding it in another lossy codec (or again as AC-3).
To give the example of Cinema DTS, I always retime that to 23.976 because the final format will be lossless anyway, so it doesn't matter much.
In the case of 25fps-synced stuff, that just doesn't leave you much of a choice, you will have to retime that anyway. The only choice there is the final delivery form. Lossless would be the only way to preserve the quality, but often the track just isn't important enough to justify that size, so you just accept the generational loss.
Whether to retime 25fps-synced stuff to 24 or 23.976, I think that honestly doesn't make much of a difference if you use a good SRC (sample rate converter). iZotope 64 bit SRC seems to be the best one from all I've read so far (its included in Sony/MAGIX Soundforge Pro).
From what you said about this project, I would also go with 24fps. It will preserve the commentary tracks nicely.
Btw, I recently figured out that if you use ffmpeg for AC-3 conversion, the default setting uses DRC. You need to add the parameter "-drc_scale 0.00000" to disable DRC. Else you end up with garbage audio.
Posts: 7,151
Threads: 601
Joined: 2015 Jan
Thanks: 1081
Given 1497 thank(s) in 969 post(s)
Country:
deleted user
Unregistered
Thanks:
Given thank(s) in post(s)
Posts: 496
Threads: 22
Joined: 2017 Dec
Thanks: 51
Given 253 thank(s) in 172 post(s)
Country:
2018-12-23, 05:49 AM
(This post was last modified: 2018-12-23, 05:58 AM by SpaceBlackKnight.)
|