|
|
|
05-15-2022, 04:21 PM
|
#1
|
Human being with feelings
Join Date: Jun 2016
Posts: 267
|
best practices rendering 192kHz to 48kHz (downsampling)
Hello!
I'm curious if anyone would share best practices they've developed -- or other good ideas they've caught on to -- when downsampling. Thought I'd find something on the forum about this, but couldn't.
In my case, I have 161 audio files at 192kHz, and I'm wanting to create 48kHz versions. This to be treated as a library of audio files.
For what it's worth, these 48kHz versions are not final masters -- they will eventually be further mixed together to create a final master mix.
So, for example, would *you* use the Batch File/Item Converter and set the sample rate to 48kHz, using the resample mode "r8brain free (highest quality, fast)" ... and away you go? Or would you go about it another way? Are there more options/processes/things to do?
Thanks!
Pietro
|
|
|
05-16-2022, 10:49 PM
|
#2
|
Human being with feelings
Join Date: Jun 2016
Posts: 267
|
Over 70 views and no replies? Hmm. It seems strange that no one has anything to say about this. Is it really such a niche topic?
|
|
|
05-17-2022, 12:48 AM
|
#3
|
Banned
Join Date: Apr 2022
Posts: 635
|
Quote:
Over 70 views and no replies?
|
Haha sorry about that- was warping time again... so hmmm..i think most users on this forum do not seem to care much for sampling higher than 44 or 48khz..? (deeming it unnecessary bloat?)
I would also just use the batch file/Item converter,check results are fine..and be on my way,if that helps any..
|
|
|
05-17-2022, 05:13 AM
|
#4
|
Human being with feelings
Join Date: Feb 2008
Location: So Florida
Posts: 1,395
|
Quote:
Originally Posted by pietro79
Over 70 views and no replies? Hmm. It seems strange that no one has anything to say about this. Is it really such a niche topic?
|
I would also just use the batch file/Item converter,check results are fine..and be on my way,if that helps any..[/QUOTE]
Well it's usually best to convert files in half (in your case 88.1k)
But I have not found rbrain converter to be as good as the highest slow converting...(I personally go with all the slowest, highest quality converting options) Hope that helps.....
That's my 2 cents anyhoo...Jeff
__________________
Win 10 I9 20 core 48g Ram always on the latest update
|
|
|
05-17-2022, 06:03 AM
|
#5
|
Banned
Join Date: Apr 2022
Posts: 635
|
Quote:
But I have not found rbrain converter to be as good as the highest slow converting...(I personally go with all the slowest, highest quality converting options)
|
Hey- so this is interesting..have not actually had real chance to do real tests with the rbrain options yet..do you have any negative results to share on them conversions please?
|
|
|
05-17-2022, 06:41 AM
|
#6
|
Human being with feelings
Join Date: Feb 2017
Posts: 4,823
|
Quote:
Originally Posted by pietro79
Over 70 views and no replies? Hmm. It seems strange that no one has anything to say about this. Is it really such a niche topic?
|
probably not much to say, just convert them. I doubt many people are dissatisfied with the results, and if they are they probably just try the other options until they are happy.
what problems are you having?
I think the source material may have a bearing on which option is most efficient etc. so perhaps a little bit of trial & error is needed, or just always use the highest quality settings.
|
|
|
05-17-2022, 08:20 AM
|
#7
|
Human being with feelings
Join Date: Jun 2016
Posts: 267
|
Quote:
Originally Posted by onewayout
Well it's usually best to convert files in half (in your case 88.1k)
|
I wonder, why's that Jeff?
|
|
|
05-17-2022, 08:31 AM
|
#8
|
Human being with feelings
Join Date: Feb 2017
Posts: 4,823
|
Quote:
Originally Posted by pietro79
I wonder, why's that Jeff?
|
numbers that are more easily divisible means less rounding errors?
|
|
|
05-17-2022, 08:33 AM
|
#9
|
Human being with feelings
Join Date: Feb 2017
Posts: 4,823
|
Quote:
Originally Posted by onewayout
Well it's usually best to convert files in half (in your case 88.1k)
|
i think half of 192 is 96
|
|
|
05-17-2022, 09:10 AM
|
#10
|
Human being with feelings
Join Date: Jun 2016
Posts: 267
|
Quote:
Originally Posted by domzy
i think half of 192 is 96
|
If that's true, then 192 divided by 48 is 4, so there would be no rounding going on.
It's the first time I hear of "rounding errors" for downsampling, though. I know dithering (or not) is an process to consider when reducing bit depth, but wasn't aware of any similar process for downsampling sample frequency.
|
|
|
05-17-2022, 09:17 AM
|
#11
|
Human being with feelings
Join Date: Jun 2016
Posts: 267
|
Quote:
Originally Posted by domzy
what problems are you having?
|
Great question! I'm not having problems yet. I'm asking because it's not a workflow I've done much of yet and am considering the fact that it may take time for me to notice mistakes I may be making. In other words, just asking about your experiences so that I might tune in to certain errors sooner than later. I agree with the idea that it's ultimately about perception and if I can't hear a difference it doesn't matter much if at all.
But also, I know that perception sometimes takes a bit to wake up to the nuances... you know, like I go back to early audio I've produced and now I notice all sort of soft mistakes that made negatively affected the sonic textures of my early projects.
|
|
|
05-17-2022, 09:18 AM
|
#12
|
Human being with feelings
Join Date: Jun 2016
Posts: 267
|
Quote:
Originally Posted by Reapology
i think most users on this forum do not seem to care much for sampling higher than 44 or 48khz..? (deeming it unnecessary bloat?)
|
Yup, in many cases it is. I recorded 192 to have more information with which to do extreme sound design processing, like pitching down without losing high frequencies
|
|
|
05-17-2022, 09:25 AM
|
#13
|
Human being with feelings
Join Date: Feb 2017
Posts: 4,823
|
Quote:
Originally Posted by pietro79
But also, I know that perception sometimes takes a bit to wake up to the nuances... you know, like I go back to early audio I've produced and now I notice all sort of soft mistakes that made negatively affected the sonic textures of my early projects.
|
yeah, i agree with that.
Dunno what to suggest other than always use the highest quality you can or do the hard yards and learn what works best for you personally.
|
|
|
05-17-2022, 09:30 AM
|
#14
|
Human being with feelings
Join Date: Feb 2008
Location: So Florida
Posts: 1,395
|
Quote:
Originally Posted by domzy
i think half of 192 is 96
|
Oops, I answered during coffee this morning.. (-'
__________________
Win 10 I9 20 core 48g Ram always on the latest update
|
|
|
05-17-2022, 10:22 AM
|
#15
|
Human being with feelings
Join Date: Jul 2010
Location: Silicon Valley, CA
Posts: 2,779
|
Quote:
Well it's usually best to convert files in half
|
That's mostly a myth...
The mathematically-difficult part is filtering, and filtering is always a part of down-sampling (to prevent aliasing). You are using the exact-same algorithm if you cut the sample rate in half and you can't simply "throw away" every-other sample (or 3 out of 4 samples). In the real world, all of the "remaining samples" will be slightly-changed.
I don't do much conversion but personally I've never heard a difference after up-sampling or down-sampling, no matter what software or settings I was using (as long as I'm at "CD quality" or higher).
You should dither whenever you reduce the bit-depth. But that's not even a real-big deal because at 16-bits and "normal conditions" you can't hear dither, or the lack of dither, or the effects of dither anyway. So I don't think it's too-important what dither you use.
If you go down to 8-bits it probably does help because you can hear quantization noise at 8-bits. I've never tried dithering an 8-bit file.. I've almost never work with 8-bit audio.
Last edited by DVDdoug; 05-17-2022 at 10:37 AM.
|
|
|
05-17-2022, 03:28 PM
|
#16
|
Human being with feelings
Join Date: Feb 2007
Location: Hampshire, UK
Posts: 376
|
Quote:
Originally Posted by pietro79
It's the first time I hear of "rounding errors" for downsampling, though. I know dithering (or not) is an process to consider when reducing bit depth, but wasn't aware of any similar process for downsampling sample frequency.
|
This was an issue with very early sample rate conversion algorithms which were coded very simply to save processor cycles but hasn't been an issue for many years.
I'd just use the R8Brain setting and go direct to 48kHz.
|
|
|
Thread Tools |
|
Display Modes |
Linear Mode
|
Posting Rules
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
HTML code is Off
|
|
|
All times are GMT -7. The time now is 06:36 AM.
|