Old 10-20-2015, 02:04 PM   #41
JamesPeters
Human being with feelings
 
JamesPeters's Avatar
 
Join Date: Aug 2011
Location: Near a big lake
Posts: 2,668
Default

Quote:
Originally Posted by Mr. PC View Post
The test shouldn't be between two sample rates, but two songs of the same sample rate, say 48k, however one has been *MIXED* at a higher sample rate, say 192k. Both mixes should use the same FX, eg. 10 compressors, 10 EQs, a limiter etc. (and these FX should work at the higher sample rate) and not over-sample. I plan to export such a test shortly.


BTW, my 8 year old student saw this thread, and asked me what it was about! It was a long and difficult explanation.

Here's a rough experiment.

1 - http://vocaroo.com/i/s1DyAG0ABYae
2 - http://vocaroo.com/i/s13BEeK3NIp6
3 - http://vocaroo.com/i/s1ycrbdY9ufD
4 - http://vocaroo.com/i/s1aM9FqncDhB

One is mixed at 192k, the other at 48k. Both exported as 16bit 48k flacs without oversampling. Then I have 2 more mixed at both rather *with* oversampling. (Reaper doesn't seem to allow higher than 192k).


I can't trust my own judgement here as I'm biased knowing which is which, but I don't think I'd really pass the double blind test.

However, this version has only a few tracks, and not too many effects. Even the over-sampled versions don't really sound too different (I aim to mix as 'lightly' as possible. I'll try this experiment with something using heavy compression and eq tomorrow (or maybe someone else would like to try).
Those sound the same to me.

Are you going to ask me how I feel about them next? (Joking, by the way. Also did you figure out why those 2 programs were rendering differently?)
__________________
http://petersamplification.com
Using REAPER for Linux
JamesPeters is offline   Reply With Quote
Old 10-20-2015, 02:13 PM   #42
JamesPeters
Human being with feelings
 
JamesPeters's Avatar
 
Join Date: Aug 2011
Location: Near a big lake
Posts: 2,668
Default

Quote:
Originally Posted by Joaquins Void View Post
...
The synth used is Synapse Audios Dune, but you can make this happen with just about any synth that feature phase modulation or oscillator sync. It's not because the dev is bad or anything. It's just that predicting what comes out of certain processes is borderline impossible. It's also a trade off situation. I.e you could preemptively over sample like mad, but then the synth would be much heavier than it needs to be for most use cases.
...

I've been thinking about this. I've used a lot of software synths and never noticed this before. Or perhaps I have noticed this and just presumed it was intentional distortion for a particular patch. I guess it's also possible a lot of synth plugins of this type do have oversampling to an extent, possibly automatically enabled if a certain function or combination of functions is activated.
__________________
http://petersamplification.com
Using REAPER for Linux
JamesPeters is offline   Reply With Quote
Old 10-20-2015, 04:21 PM   #43
Softsynth
Human being with feelings
 
Softsynth's Avatar
 
Join Date: Jun 2015
Posts: 5,520
Default

Quote:
Originally Posted by Mr. PC View Post
The test shouldn't be between two sample rates, but two songs of the same sample rate, say 48k, however one has been *MIXED* at a higher sample rate, say 192k. Both mixes should use the same FX, eg. 10 compressors, 10 EQs, a limiter etc. (and these FX should work at the higher sample rate) and not over-sample. I plan to export such a test shortly.


BTW, my 8 year old student saw this thread, and asked me what it was about! It was a long and difficult explanation.

Here's a rough experiment.

1 - http://vocaroo.com/i/s1DyAG0ABYae
2 - http://vocaroo.com/i/s13BEeK3NIp6
3 - http://vocaroo.com/i/s1ycrbdY9ufD
4 - http://vocaroo.com/i/s1aM9FqncDhB

One is mixed at 192k, the other at 48k. Both exported as 16bit 48k flacs without oversampling. Then I have 2 more mixed at both rather *with* oversampling. (Reaper doesn't seem to allow higher than 192k).


I can't trust my own judgement here as I'm biased knowing which is which, but I don't think I'd really pass the double blind test.

However, this version has only a few tracks, and not too many effects. Even the over-sampled versions don't really sound too different (I aim to mix as 'lightly' as possible. I'll try this experiment with something using heavy compression and eq tomorrow (or maybe someone else would like to try).
It could be interesting to hear the whole piece. I listened several times via headphones.
Anyway all four sound virtually identical to my ears. I could be imagining it but consistently the first version sounded a little clearer, a clicking background percussive sound property(if we can call it that) seems crisper, tighter, more open & airy.
It could be my imagination (My brains expectation bias; something SHOULD sound different).
I couldn't venture a guess as to what caused the difference and I wouldn't feel cheated if you said it was a test and they were actually all the same.
The difference is that small, but I hear it every time.
Softsynth is offline   Reply With Quote
Old 10-20-2015, 05:05 PM   #44
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 24,755
Default

Regardless of format, the above files are showing the following when downloading. Is this intentional?

Length: 0:22.517
Samplerate: 22050
Channels: 1
Bits/sample: 16
Bitrate: 193.4kbps, 54.8%

Even if it isn't, it's even more telling IMHO because there isn't enough difference between files to be any type of deal breaker IMHO even at that low sample rate.
__________________
If it requires a null test to find it, it is by definition minuscule.
karbomusic is online now   Reply With Quote
Old 10-20-2015, 05:12 PM   #45
Softsynth
Human being with feelings
 
Softsynth's Avatar
 
Join Date: Jun 2015
Posts: 5,520
Default

Quote:
Originally Posted by karbomusic View Post
Regardless of format, the above files are showing the following when downloading. Is this intentional?

Length: 0:22.517
Samplerate: 22050
Channels: 1
Bits/sample: 16
Bitrate: 193.4kbps, 54.8%

Even if it isn't, it's even more telling IMHO because there isn't enough difference between files to be any type of deal breaker IMHO even at that low sample rate.
Weird, the downloads are mono. Main files are stereo.
Softsynth is offline   Reply With Quote
Old 10-20-2015, 05:13 PM   #46
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 24,755
Default

Quote:
Originally Posted by Softsynth View Post
Weird, the downloads are mono.
And 22k.
__________________
If it requires a null test to find it, it is by definition minuscule.
karbomusic is online now   Reply With Quote
Old 10-20-2015, 05:14 PM   #47
serr
Human being with feelings
 
Join Date: Sep 2010
Posts: 8,256
Default

They sound the same because he did something to the files far beyond sample rate conversions. They've been reduced to 16 bit.

Can't very well start comparing subtle fidelity issues with sample rate conversion with that going on!

What's the point of hiding the results like that?
If you hear a difference, you can decide to set your sample rate switch to HD. I bet most people have so much drive space it's a moot point too.

Or not...
serr is online now   Reply With Quote
Old 10-20-2015, 05:16 PM   #48
Softsynth
Human being with feelings
 
Softsynth's Avatar
 
Join Date: Jun 2015
Posts: 5,520
Default

Quote:
Originally Posted by karbomusic View Post
And 22k.
Yep, strange.
Softsynth is offline   Reply With Quote
Old 10-20-2015, 05:20 PM   #49
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 24,755
Default

Quote:
They sound the same because he did something to the files far beyond sample rate conversions.
I'm not getting any closer to thinking any of this is actually mattering. If they won't even null at 22050 yet they sound very much alike at the same time. Just sayin' I can't wait for the big differences to be posted since I'm typically the one who isn't the A/B/X proponent. I am a proponent of context matters but maybe need a little more for this one.
__________________
If it requires a null test to find it, it is by definition minuscule.
karbomusic is online now   Reply With Quote
Old 10-20-2015, 05:54 PM   #50
serr
Human being with feelings
 
Join Date: Sep 2010
Posts: 8,256
Default

I honestly think the goofiest cases come up with recordings that have already had damage from generation loss (analog, digital, or whatever).

Just like a blurry photocopy you can barely read, you need a perfect capture now or you're not going to be able to read it.


I still think this can be simplified.

Use 24 bit always. Only exception being to make a CD format (in addition to the 24 bit version).

Record at the highest quality your system supports without bogging down.
24 bit 96k is a release format now. Why the hell not? And we have surround sound just as easy to deliver now too.

Enjoy the convenience of being able to get work done and get away with an awful lot of technically lossy conversions and still have above average audio. But don't start going lossy for no reason! Why the heck would you do that, right?

Simple.


Constantly evaluating how far you can compromise your audio based on the input source and always messing around with this? What the hell?!

Set the switch to full quality and just record and mix.
serr is online now   Reply With Quote
Old 10-20-2015, 05:56 PM   #51
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 24,755
Default

Quote:
Just like a blurry photocopy you can barely read, you need a perfect capture now or you're not going to be able to read it.
I'm not disagreeing, just waiting for real-world examples to show it to be a worthwhile difference. Even if it is a small difference but it causes problems in different contexts, I'm game but need to see/hear that scenario.
__________________
If it requires a null test to find it, it is by definition minuscule.
karbomusic is online now   Reply With Quote
Old 10-20-2015, 08:30 PM   #52
JamesPeters
Human being with feelings
 
JamesPeters's Avatar
 
Join Date: Aug 2011
Location: Near a big lake
Posts: 2,668
Default

Quote:
Originally Posted by karbomusic View Post
And 22k.
He did this on another thread, that I'd linked to above in my last reply to him.

Mr. PC--if that hosting site only allows 22KHz files, don't use it. Use any of a number of other free hosting sites for this purpose, and allow the proper files to be uploaded.
__________________
http://petersamplification.com
Using REAPER for Linux
JamesPeters is offline   Reply With Quote
Old 10-21-2015, 02:02 AM   #53
Softsynth
Human being with feelings
 
Softsynth's Avatar
 
Join Date: Jun 2015
Posts: 5,520
Default

Quote:
Originally Posted by JamesPeters View Post
He did this on another thread, that I'd linked to above in my last reply to him.

Mr. PC--if that hosting site only allows 22KHz files, don't use it. Use any of a number of other free hosting sites for this purpose, and allow the proper files to be uploaded.
That's only the download versions, not the files in the links, which you play from the site.
Softsynth is offline   Reply With Quote
Old 10-21-2015, 03:51 AM   #54
kenz
Human being with feelings
 
Join Date: Aug 2013
Posts: 339
Default

Quote:
Originally Posted by serr View Post
I still think this can be simplified.

Use 24 bit always. Only exception being to make a CD format (in addition to the 24 bit version).

Record at the highest quality your system supports without bogging down.
24 bit 96k is a release format now. Why the hell not? And we have surround sound just as easy to deliver now too.

Enjoy the convenience of being able to get work done and get away with an awful lot of technically lossy conversions and still have above average audio. But don't start going lossy for no reason! Why the heck would you do that, right?
Why not go 64-bit 192khz too for playback? Literally it has the exact same reasons for a rendering. I'm not sure what's so special about your 24-bit campaign or why you think it has different reasons than 64-bit for example. When both have been proven useless for playback, and it's just math. So if you find 64-bit ridiculous then almost same claim can be made for 24-bit and math will show the same and personally I'm not going into that again...

I don't even understand really, what's the point of 24-bit and why is it so special compared to *better* formats for editing? Don't get me wrong, for recording you should use 24-bit (or 20-bit), but for simplicity like you mention, I use 32-bit floating point to not worry about clipping either. I do a lot of sound design/manipulations so that's very useful. Obviously rendering is a whole different story, 16-bit suffices there and there's no worry of clipping anyway (or shouldn't be...).

The point is: 24-bit has no point. 16-bit for rendering, 32-bit for editing. I don't see much point for 24-bit. So with your logic, I should render at 32-bit? If not, why not? What's so special about 24-bit, the fact that players support it? It's useless for playback anyway.

Anyway, bit depth has nothing to do with samplerate conversion at all (by that I mean it doesn't impact it) not sure why you'd bring this up here. I realize you're being sarcastic about the bit depth impacting the sound quality after resampling, but it's actually true, it doesn't.
kenz is offline   Reply With Quote
Old 10-21-2015, 03:59 AM   #55
bladerunner
Human being with feelings
 
bladerunner's Avatar
 
Join Date: Sep 2007
Location: Kent, UK
Posts: 4,779
Default

Quote:
Originally Posted by zyisrad View Post
Just like everyone pines for that analog sound while using ITB plugs now, in the future people will be nostalgic for that 44.1khz sound.
Yes - profoundness here. Reminding everyone that anything to do with music making is art first and foremost.
bladerunner is offline   Reply With Quote
Old 10-21-2015, 04:08 AM   #56
LightOfDay
Banned
 
Join Date: Jun 2015
Location: Lower Rhine Area, DE
Posts: 964
Default

so I have a question: why is higher samplerate still considered to = better quality? that is complete nonsense. do a thorough research on the subject and dont read random internet-bullshit on this matter.

keywords here are Monty Montgomery and Lawo.

24 bit/96 khz are for people believing in snakeoil. for the not-knowing-what they are talking about.

24 bit is better than 16 bit is because you have a higher dynamic range that will cause you less trouble while mixing and mastering. its not because you will gain higher audio-quality.

but against all scientific the discussion about snakeoil gets on and on and on. really? come on, learn something.
LightOfDay is offline   Reply With Quote
Old 10-21-2015, 04:21 AM   #57
kenz
Human being with feelings
 
Join Date: Aug 2013
Posts: 339
Default

Higher sample rates are useful for same reason(s) 24-bit is useful while mixing: it gets less "errors" (in samplerate case, it is 'aliasing'), unless the plugin internally oversamples. But for playback, they aren't. (the "errors" in bitdepth is what would be the noise floor, so higher dynamic range)

However I fail to see a reason to use 24-bit instead of 32-bit floating point while editing/mixing, to be honest. Perhaps for mixing it's not useful but when doing sound edits/design things can go crazy and instead of trying to tame them at the input (which will change the sound if non-linear processing takes place!) just lower the output volume because there's no clipping. One less thing to worry about.
kenz is offline   Reply With Quote
Old 10-21-2015, 04:43 AM   #58
planetnine
Human being with feelings
 
planetnine's Avatar
 
Join Date: Oct 2007
Location: Lincoln, UK
Posts: 7,585
Default

It also used to be the case that for many converters, 24/96 did sound better, because 44k1 operation was compromised by antialiasing filters, and design and component shortcuts in all but the best converters.

This will take some time to be forgotten.



>
planetnine is offline   Reply With Quote
Old 10-21-2015, 05:43 AM   #59
Mr. PC
Human being with feelings
 
Mr. PC's Avatar
 
Join Date: Apr 2010
Location: Cloud 37
Posts: 905
Default

I guess that uploading site is converting my files. Damn. I just discovered the Reaper Resources thing, so I'll use that.

Somehow, I'm now getting aliasing on *all* my renders. I'm using Reasynth with both 48k and 192k,

https://stash.reaper.fm/25498/192.png
https://stash.reaper.fm/25499/48.png

In the DAW, there's a huge difference, but somehow after being rendered they sound the same.
Also, it seems there's no way to set sample-rate above 192 (is anyone able to do this?) I'm still getting aliasing even at 192k, but far less than at 48k. Most aliasing seems to come from the synth, rather than compressors or EQs.

kenz
--Would you suggest I mix at 192k? Any idea why the aliasing occurs on render, but now in the DAW? Here are my render settings,
https://stash.reaper.fm/25500/render.png

JamesPeters
--Actually, aliasing was my problem on that older mix; that's why I started looking into sample-rates.
__________________
AlbertMcKay.com
SoundCloud BandCamp
ReaNote Hotkeys to make Reaper notation easy/fast

Last edited by Mr. PC; 10-21-2015 at 05:55 AM.
Mr. PC is offline   Reply With Quote
Old 10-21-2015, 05:45 AM   #60
Softsynth
Human being with feelings
 
Softsynth's Avatar
 
Join Date: Jun 2015
Posts: 5,520
Default

Quote:
Originally Posted by Mr. PC View Post
I guess that uploading site is converting my files. Damn. I just discovered the Reaper Resources thing, so I'll use that.

Somehow, I'm now getting aliasing on *all* my renders. I'm using Reasynth with both 48k and 192k,

https://stash.reaper.fm/25498/192.png
https://stash.reaper.fm/25499/48.png

In the DAW, there's a huge difference, but somehow after being rendered they sound the same. It seems there's no way to set sample-rate above 192 (is anyone able to do this?)

kenz

Would you suggest I mix at 192k? Any idea why the aliasing occurs on render, but now in the DAW? Here are my render settings,
https://stash.reaper.fm/25500/render.png

JamesPeters

Actually, aliasing was my problem on that older mix; that's why I started looking into sample-rates.
Why are the downloadable versions totally different files though?
very strange.
Softsynth is offline   Reply With Quote
Old 10-21-2015, 05:46 AM   #61
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 24,755
Default

Quote:
Originally Posted by Softsynth View Post
That's only the download versions, not the files in the links, which you play from the site.
They sounded so close (insignificant) when trying to compare them on the site that I then wanted to download to do a better blind comparison, check nulling amount etc. I'm still waiting on something that is beyond insignificant. I'm not saying it isn't possible just waiting for it to be posted.

As far as 32bit is concerned, what kenz said. At least make the glue/freeze settings 32 bit and you'll never have to worry about clipping at the track level. Please practice good gain anyway because that is organized and smart regardless but this covers you if you go over and start clipping at the track level.
__________________
If it requires a null test to find it, it is by definition minuscule.
karbomusic is online now   Reply With Quote
Old 10-21-2015, 05:51 AM   #62
Softsynth
Human being with feelings
 
Softsynth's Avatar
 
Join Date: Jun 2015
Posts: 5,520
Default

Quote:
Originally Posted by karbomusic View Post
They sounded so close (insignificant) when trying to compare them on the site that I then wanted to download to do a better blind comparison, check nulling amount etc. I'm still waiting on something that is beyond insignificant. I'm not saying it isn't possible just waiting for it to be posted.

As far as 32bit is concerned, what kenz said. At least make the glue/freeze settings 32 bit and you'll never have to worry about clipping at the track level. Please practice good gain anyway because that is organized and smart regardless but this covers you if you go over and start clipping at the track level.
For that purpose I use 32bit 48khz.
Softsynth is offline   Reply With Quote
Old 10-21-2015, 06:04 AM   #63
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 8,696
Default

Quote:
Originally Posted by LightOfDay View Post
so I have a question: why is higher samplerate still considered to = better quality? that is complete nonsense. do a thorough research on the subject and dont read random internet-bullshit on this matter.

keywords here are Monty Montgomery and Lawo.

24 bit/96 khz are for people believing in snakeoil. for the not-knowing-what they are talking about.

24 bit is better than 16 bit is because you have a higher dynamic range that will cause you less trouble while mixing and mastering. its not because you will gain higher audio-quality.

but against all scientific the discussion about snakeoil gets on and on and on. really? come on, learn something.
96 kHz is also the industry standard for production and processing now.

You think it's an inferior standard? Well, inferior standards win all the time. Betamax was better than VHS. Get over it.

If you are a hobbyist, or a little local studio, and you do every part of the chain from recording to mastering, or you know exactly who is going to be mixing or mastering and they're cool with it, then sure, you can use 44.1kHz, or give your mastering engineer a 16bit file or do whatever you want. No-one will call the cops on you.

But the big boys and girls have been using 96kHz as a standard for quite a while now. 88.2kHz is the minimum sample rate expected to be handed to a label by a producer.

It has more to do with future-proofing for changes in delivery format and archiving for remasters than anything else, from what I've read.

It's all here, in black and white: http://www.aes.org/technical/documen....2.15-02_1.pdf

Last edited by Judders; 10-21-2015 at 06:12 AM.
Judders is offline   Reply With Quote
Old 10-21-2015, 06:42 AM   #64
kenz
Human being with feelings
 
Join Date: Aug 2013
Posts: 339
Default

Quote:
Originally Posted by Judders View Post
But the big boys and girls have been using 96kHz as a standard for quite a while now. 88.2kHz is the minimum sample rate expected to be handed to a label by a producer.
The "big boys" who still use antiquated equipment and ignore the noise it generates past 15khz+ or so? Yeah I'm sure they definitely care about quality above 20khz if they can't even be bothered to upgrade it and remove the noise within the audible range. Sometimes I still see the 15khz CRT line in new recordings, seriously? (but that's not even the biggest issue)

It's an absolute joke. How can you claim to even be able to hear a difference above 22khz when they definitely don't care even of 15khz range or up? A sonogram easily shows how bad it is. Note I'm obviously talking about the final render.

Big boys are just marketing babble.

FWIW I use 96khz when editing but that's mostly because I'm a perfectionist and want as little aliasing as possible.


Mr. PC: I definitely do not recommend 192khz. That's overkill and a lot of plugins may not work with it properly, depending on your plugin range I suppose. 96khz is well enough to get rid of aliasing in my opinion, at least within any audible sense before it gets downed in noise.

Same reason I see some people praise 24-bit on noisy recordings which have a noise floor far above even the 16-bit range... it's pretty ridiculous how much babble there is about them "hearing a difference" when they don't even care of perfecting the sound.

If you want quality, pay/work for it. People think magically flipping the 96khz/24-bit+ switch will make it sound good without the necessary clean equipment or work to clean it etc. No, seriously even if your target audience are dogs at least take the effort and clean the sound so it shows in the sonogram not like a joke.
kenz is offline   Reply With Quote
Old 10-21-2015, 06:46 AM   #65
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 8,696
Default

Quote:
Originally Posted by kenz View Post
The "big boys" who still use antiquated equipment and ignore the noise it generates past 15khz+ or so? Yeah I'm sure they definitely care about quality above 20khz if they can't even be bothered to upgrade it and remove the noise within the audible range. Sometimes I still see the 15khz CRT line in new recordings, seriously? (but that's not even the biggest issue)

It's an absolute joke. How can you claim to even be able to hear a difference above 48khz when they definitely don't care even of 15khz range or up? A sonogram easily shows how bad it is.

Big boys are just marketing babble.

FWIW I use 96khz when editing but that's mostly because I'm a perfectionist and want as little aliasing as possible.
What anyone can hear is beside the point.

Professionals are obliged to deliver in whatever format their clients pay them to deliver. It's as simple as that.

If you think that technology and industry standards are ever driven solely by logic and empirical results, I think you are being rather naive.
Judders is offline   Reply With Quote
Old 10-21-2015, 06:47 AM   #66
kenz
Human being with feelings
 
Join Date: Aug 2013
Posts: 339
Default

I didn't say you are wrong. I merely stated that their reasons for doing so are marketing babble, not coming from them, but from the "pro equipment" manufacturers. How else would they get to sell their new expensive gear to the pros? Look at Avid.
kenz is offline   Reply With Quote
Old 10-21-2015, 07:22 AM   #67
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 8,696
Default

Quote:
Originally Posted by kenz View Post
I didn't say you are wrong. I merely stated that their reasons for doing so are marketing babble, not coming from them, but from the "pro equipment" manufacturers. How else would they get to sell their new expensive gear to the pros? Look at Avid.
Maybe. I don't have any contacts in the gear-hawking industry.

Another thing to consider with the archiving of tape: even if it's a useless, inaudible mess above 22kHz, might we not want to still save that for historical reasons? When all the tape in the world has crumbled to dust, or is too precious to play, maybe that information might be useful on a purely academic level?
Judders is offline   Reply With Quote
Old 10-21-2015, 08:17 AM   #68
JamesPeters
Human being with feelings
 
JamesPeters's Avatar
 
Join Date: Aug 2011
Location: Near a big lake
Posts: 2,668
Default

Quote:
Originally Posted by Softsynth View Post
That's only the download versions, not the files in the links, which you play from the site.
Oh well...I'm not liking the idea of tabbing back and forth "playing media in the browser" to do comparisons. My memory for this level of subtlety in sound is too temporary for the amount of time spent switching between tabs and starting/stopping each file.
__________________
http://petersamplification.com
Using REAPER for Linux
JamesPeters is offline   Reply With Quote
Old 10-21-2015, 08:28 AM   #69
Mr. PC
Human being with feelings
 
Mr. PC's Avatar
 
Join Date: Apr 2010
Location: Cloud 37
Posts: 905
Default

Quote:
Originally Posted by kenz View Post
FWIW I use 96khz when editing but that's mostly because I'm a perfectionist and want as little aliasing as possible.

Mr. PC: I definitely do not recommend 192khz. That's overkill and a lot of plugins may not work with it properly, depending on your plugin range I suppose. 96khz is well enough to get rid of aliasing in my opinion, at least within any audible sense before it gets downed in noise.
Well, is there a way of finding out which plugins will work? Voxengo plugins work at all sample rates (do reaper plugs?) It seems Acon only works up to 96k, so maybe I'll drop it for Voxengo Reverbs.

These days, I'm only using,

ReaComp
Voxengo - Gliss EQ, Elephant, OldSkool Reverb.
Acon reverb*

I actually *am* hearing aliasing on my synths, even at 192k, and my synths don't over-sample (can ReaSynth oversample?). Another real-world problem (not totally related to this thread, but most of this thread is unrelated to this thread) is that my exports are aliasing, even at 192k, even though they don't alias (well, they do, but not as much as on export).
__________________
AlbertMcKay.com
SoundCloud BandCamp
ReaNote Hotkeys to make Reaper notation easy/fast
Mr. PC is offline   Reply With Quote
Old 10-21-2015, 08:40 AM   #70
kenz
Human being with feelings
 
Join Date: Aug 2013
Posts: 339
Default

Quote:
Originally Posted by Judders View Post
Another thing to consider with the archiving of tape: even if it's a useless, inaudible mess above 22kHz, might we not want to still save that for historical reasons? When all the tape in the world has crumbled to dust, or is too precious to play, maybe that information might be useful on a purely academic level?
Are you talking about saving the tape itself, or the noise above 22khz?

If it's a specific type of noise, then yes it's important to save one of these (just one suffices, one per type) for academic purposes / science (and know the cause of it too, of course). But remember that no matter how good restoration technology will get, you cannot extract a (specific frequency) signal if it's completely drown in noise, it is a mathematical impossibility. Not to mention it violates laws of the universe. It would be equivalent to "creating information from no information". So there's a "point of no return" so to speak.

As long as you know the noise profile/shape (not sure of the correct term here), no reason to store redundant noise (other than backup copies of it obviously). Noise itself is not information, that's why it's called noise. (the noise shape is, however)

But other than that, it's pointless to store a tape with more information than it has. This is a similar scenario to taking an 8-bit signal and upsampling it to 64-bit, then storing it as archive. What's the point instead of storing it as the original 8-bit? You can't "create quality" if the source material is bad, regardless of the format you use. Obviously you can even record one at 64-bit but if it's so bad and it has a -20db noise floor, it's not going to be of any use.

(that's what I meant with people wanting high samplerates without even bothering to clean their equipment defects... and if they can't hear it, then they clearly can't hear the higher frequencies either; personally I can barely hear those defects and only in very specific passages mostly when the sound is very soft, I'm sure I won't be able to as I age more sadly...)

Quote:
Originally Posted by Mr. PC View Post
Well, is there a way of finding out which plugins will work? Voxengo plugins work at all sample rates (do reaper plugs?) It seems Acon only works up to 96k, so maybe I'll drop it for Voxengo Reverbs.
Honestly no idea, besides trial-and-error and testing them, or just reading their User Manual, usually they specify the supported sample rates... (if it doesn't say anything, proceed to test and check if it works on *any* or not)
kenz is offline   Reply With Quote
Old 10-21-2015, 08:42 AM   #71
drtedtan
Human being with feelings
 
drtedtan's Avatar
 
Join Date: Jul 2010
Posts: 332
Default

Quote:
Originally Posted by kenz View Post
Why not go 64-bit 192khz too for playback? Literally it has the exact same reasons for a rendering. I'm not sure what's so special about your 24-bit campaign or why you think it has different reasons than 64-bit for example. When both have been proven useless for playback, and it's just math.
I don't want to put words in serr's mouth, but I believe his point was to work in (at least) 24 bit, not to fixate upon it. Besides, there aren't many A/D converters that work at 32 bit, so you're probably stuck with 24 bit for capturing unless you intend to compromise your dynamic range.

Then, at export, if you have a 24+ bit render, you can sell the HD audio plus you can always down convert as appropriate at a later date if and when necessary for a given medium.


Quote:
Originally Posted by kenz View Post
Obviously rendering is a whole different story, 16-bit suffices there and there's no worry of clipping anyway (or shouldn't be...).

The point is: 24-bit has no point. 16-bit for rendering
16 bit is fine for the playback of music with a limited dynamic range, but there are situations that exceed the limits of 16 bit playback, such as a symphony orchestra or a loud jazz band. These types of music can make use of 120 dB or greater dynamic swings, which 16 bit cannot play back without compromising the signal with compression. (While some people may prefer the compressed version, it is not an accurate capture of the live ensemble's performance, which is often the goal in these types of music).
drtedtan is offline   Reply With Quote
Old 10-21-2015, 08:56 AM   #72
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 8,696
Default

Quote:
Originally Posted by kenz View Post
Are you talking about saving the tape itself, or the noise above 22khz?
It's my understanding that tape can have signal above the noise floor in excess of 22kHz.

For any kind of archiving, a well preserved original is best, but facsimiles should capture as much of the source as possible. If I was archiving tape, having a plot of typical frequency responses of various formulas would not make up for having low-passed digital copies in my archive.
Judders is offline   Reply With Quote
Old 10-21-2015, 08:58 AM   #73
Softsynth
Human being with feelings
 
Softsynth's Avatar
 
Join Date: Jun 2015
Posts: 5,520
Default

Quote:
Originally Posted by drtedtan View Post
I don't want to put words in serr's mouth, but I believe his point was to work in (at least) 24 bit, not to fixate upon it. Besides, there aren't many A/D converters that work at 32 bit, so you're probably stuck with 24 bit for capturing unless you intend to compromise your dynamic range.

Then, at export, if you have a 24+ bit render, you can sell the HD audio plus you can always down convert as appropriate at a later date if and when necessary for a given medium.




16 bit is fine for the playback of music with a limited dynamic range, but there are situations that exceed the limits of 16 bit playback, such as a symphony orchestra or a loud jazz band. These types of music can make use of 120 dB or greater dynamic swings, which 16 bit cannot play back without compromising the signal with compression. (While some people may prefer the compressed version, it is not an accurate capture of the live ensemble's performance, which is often the goal in these types of music).
The vast majority of audiophiles (never mind regular folks) can count the number of people that they know that have systems capable of utilizing this amazing dynamic range on no fingers of either hand.
It is even less probable to have the opportunity to be able to listen at volumes levels that justify such a system AND to have a room that has good enough acoustics to support it AND recordings that actually offer such wide dynamic range, not just on paper but actually measurably utilizing that dynamic range.
Presuming you have a worthy musical performance (little point otherwise) and a high power handling system and room that can support it then you have to have young listeners with exceptional hearing without a hint or trace of tinnitus (otherwise that noise floor doesn't really mean much).

Even Classical 16bit CDs have been returned in their droves for having too wide dynamic range (quite bits are too quiet, loud parts too loud etc.). The best we can hope for is a happy medium. Barely acceptable dynamic range is a rarity with today's loudness wars. 16bit dynamic range isn't being utilized, let alone 24bit.

Last edited by Softsynth; 10-21-2015 at 09:05 AM.
Softsynth is offline   Reply With Quote
Old 10-21-2015, 10:20 AM   #74
Mr. PC
Human being with feelings
 
Mr. PC's Avatar
 
Join Date: Apr 2010
Location: Cloud 37
Posts: 905
Default

Also, when you're driving, and the horns come out and blast you, so you almost crash, then you turn it down, but a second later you can't hear the pianissimo section, so you're turning it up again.

There should be a law that all musical playback devices have some built in compression controls (same with T.V., and we should be able to mix the levels, because the voices are always too quite and sound FX too loud).

Anyhow, I actually listen to mp3 most of the time, and I would export 16 bit 48k, except that, as mentioned about, people will download my music in differing formats. Exporting 24bit gives them the option of converting to what they will (should I dither when exporting to 24bit btw?)

But everything in this post makes me think Mixing in 192k 32 bit is the way to go (unless I have a non-supporting plug; I've been reading manuals and my plugs are all good for 192).

Thanks! Now I have new mixing problems (I have some annoying 'hollow parts' in the stereo field, only at certain sections of the song. The reverb helps fill it, but makes it blurry at the same time. I'm trying a small stereo delay mixed with reverb. Maybe I'll do a reverb thread as it really drives me nuts).
__________________
AlbertMcKay.com
SoundCloud BandCamp
ReaNote Hotkeys to make Reaper notation easy/fast
Mr. PC is offline   Reply With Quote
Old 10-21-2015, 10:33 AM   #75
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 24,755
Default

I'm still waiting for any sample files that really audibly show something worth worrying about here.
__________________
If it requires a null test to find it, it is by definition minuscule.
karbomusic is online now   Reply With Quote
Old 10-21-2015, 10:54 AM   #76
cyrano
Human being with feelings
 
cyrano's Avatar
 
Join Date: Jun 2011
Location: Belgium
Posts: 4,587
Default

Quote:
Originally Posted by kenz View Post
Why not go 64-bit 192khz too for playback?
There's no hardware supporting that.

Only choice is 16 or 24 bit. Anything else is ludicrous for playback as it will involve resampling on it's way to your speakers, which might introduce new artefacts.
cyrano is offline   Reply With Quote
Old 10-21-2015, 12:31 PM   #77
The Telenator
Banned
 
Join Date: Dec 2011
Location: Oud West, NL
Posts: 2,335
Default

Quote:
Originally Posted by Judders View Post
96 kHz is also the industry standard for production and processing now.

It has more to do with future-proofing for changes in delivery format and archiving for remasters than anything else, from what I've read.

It's all here, in black and white: http://www.aes.org/technical/documen....2.15-02_1.pdf
I'll point to this post by Judders as the most important in the last 24 hours or so of posts as I'm looking through them now.

Along with this, I think everyone ought to acquaint themselves with what iTunes requires now and the changes proposed for the near future. Someone have the link for that ... the rendering-converting for iTunes doc? I include them here because they have become such a major player.

I initially approached this thread with full expectations that it would quickly and totally degenerate into a flame fest, but instead it's turned out to be a pretty fair discussion. If read through carefully, I think anyone coming to this topic seeking enlightenment or at least a better, fuller understanding won't be disappointed.

Last, I just want to mention without opening a side-discussion or highjack that I really wish all those dozens and dozens of websites that accept MP3 submissions would all up their standards. Why insist on 192kbps from artists when everyone has cloud storage and, as has been pointed out here repeatedly, storage space for music files just isn't an issue to the major players anymore?
The Telenator is offline   Reply With Quote
Old 10-21-2015, 10:04 PM   #78
insub
Human being with feelings
 
insub's Avatar
 
Join Date: Mar 2014
Location: Louisville, KY, USA
Posts: 1,044
Default

I may be wrong about this, but I think there is some confusion about what the Project Sample Rate is. Changing the project sample rate during mixing will have zero effect on the audio you're mixing that's already been recorded. It may affect VSTi, but you should probably have already rendered to audio before mixing anyway. Most VST effects I believe will run in any sample rate the host can utilize. You're recorded audio files will be determined by your interface's abilities.

So, are you asking if you should record at higher sample rate? The answer is yes, but not as high as all get-out. A white paper I read about the Nyquist Frequency indicated that a samplerate of around 60kHz will accommodate today's technology for A/D converters' ultrasonic filtering.

I believe your VST effects will run at the audio file's samplerate, not the project samplerate. Although, the two should be matched.

Sure, using all audio at 88.2kHz will reduce the need for oversampling in your plugins. I think the reason VSTs have select-able samplerates is because not everyone is using files of the same samplerate. Example: If you recorded @44.1 you may need 4x, but If you recorded @88.2 you only need 2x, and for 196kHz you'll need no oversampling at all for the same process.

But, not all processes need oversampling at all. So, all 196kHz files in a project is probably overkill. It was easy to understand that distortion/saturation needs oversampling because of how it works, but I was surprised to learn that compression could need oversampling as well among other processes.

So many times we over-complicate the issue. It is really quite simple. Does the plugin offer oversampling? If yes, then you may need it. When? Turn it on and see if there is an improvement. If not, turn it back off, or try a higher multiplier. Can you even hear any aliasing? Melda has a nice video demonstrating the need for oversampling for some audio effect processing. Go check it out.

I think VSTi have driven a lot of this issue. When people were working outside the box there was conversion, or it was analog (no samplerate at all). We recorded using microphones (which typically don't capture much energy in the ultrasonic range) and those mics were aimed at amplifiers which also could not produce much ultrasonic energy. And then that went through an A/D converter when we moved to digital. But, the A/D converter has a built in LPF to remove ultrasonics prior to digitizing specifically to prevent aliasing. But, these VSTi are maybe not filtering the ultrasonics before the output. WHY? I don't know. Crappy programming perhaps, laziness, or ignorance? Whatever. Perhaps that is part of the difference between good VSTi and crappy ones.

I work mostly with audio captured using mics through an interface. So, I hardly encounter aliasing. Most aliasing in these situations are below the audible threshold. "Oh, but they can add up as the track count increases," you say. But, they could also decrease depending on the phase relationship. It seems we go through a lot of trouble using saturation and distortion to add information (noise) to the upper, audible frequency range and then start complaining about aliasing noise in the same range... What???

I think some good info came out in this thread I started about if we should even give a crap about the Nyquist Frequency. Read it here:
http://forum.cockos.com/showthread.p...nyquist+matter
I think there is a link to that white paper among other excellent information sources about digital audio including bit depth, sample rate, and the Nyquist Frequency. The fact that CDs are 16bit/44.1kHz is not a coincidence. Much smarter people than myself came up with that one, and for good, scientific reasons as well.

Please, correct me if I'm wrong.
__________________
Everything you need to know about samplerates and oversampling... maybe!
My Essential FREE 64bit VST Effects, ReaEQ Presets for Instruments
Windows 10 64 bit; MOTU 828 MKII, Audio Express, & 8PRE; Behringer ADA8000
insub is offline   Reply With Quote
Old 10-22-2015, 08:25 AM   #79
drtedtan
Human being with feelings
 
drtedtan's Avatar
 
Join Date: Jul 2010
Posts: 332
Default

I don't want to continue derailing this discussion of sample rates by talking about bit depth, but I will respond to your points, Softsynth, before going back on topic.


Quote:
Originally Posted by Softsynth View Post
The vast majority of audiophiles (never mind regular folks) can count the number of people that they know that have systems capable of utilizing this amazing dynamic range on no fingers of either hand.
It is even less probable to have the opportunity to be able to listen at volumes levels that justify such a system AND to have a room that has good enough acoustics to support it AND recordings that actually offer such wide dynamic range, not just on paper but actually measurably utilizing that dynamic range.
I'm not sure who you have been hanging around to arrive at that number, Softsynth, but I've known many musicians who keep their PA system(s) in their house when not using them for live sound reinforcement. And even a mediocre PA system can handle this dynamic range in a typical living room or media room environment (the larger the room, the more power you'll need to attain higher volume levels, so smaller rooms in the home are not a particular challenge). No special acoustics or system required. And as technology continues to develop, it is likely that even the average Joe will have access to such systems in his home in the near future.

But this is missing the greater point that I intended to make, which is not that we should release at 24 bit, necessarily, but that we should track, mix and master at 24 bit (or greater) so that way we have the option of utilizing that bit depth for either 1) projects that require it, or 2) in the future as it is becoming an industry standard audio format. From there, you can always create lower resolution versions, for lack of a better term, but you can't increase the "resolution" if you start with a 16/44.1K file.


Quote:
Originally Posted by Softsynth View Post
Presuming you have a worthy musical performance (little point otherwise)and a high power handling system and room that can support it
My comment was in regards to recording orchestras, such as the London Symphony Orchestra or the Prague Philharmonic, so the worthy performance should be given in that context. But you do have a good point as it applies to the music that the average Joe makes in his bedroom.


Quote:
Originally Posted by Softsynth View Post
then you have to have young listeners with exceptional hearing without a hint or trace of tinnitus (otherwise that noise floor doesn't really mean much).
Tinnitus is not actually sound, it is imaginary, and as such, it does not interfere with our ability to hear real sounds.


Quote:
Originally Posted by Softsynth View Post
Even Classical 16bit CDs have been returned in their droves for having too wide dynamic range (quite bits are too quiet, loud parts too loud etc.). The best we can hope for is a happy medium. Barely acceptable dynamic range is a rarity with today's loudness wars. 16bit dynamic range isn't being utilized, let alone 24bit.
As I said in the post you quoted, people may well prefer the reduced dynamic range of a compressed version of the performance, but that does not make it an accurate transcription of the performance as it originally occurred.

Which brings me back to my main point - if you record, mix, master at 24 bit (and 96K, which is becoming an industry standard) you can also create numerous other versions for different clients. That way you can deliver the 24/96K file to the director who wants to use your music in his movie and the compressed version to Mr. PC so he won't be scared if he forgets to turn the volume back down before the horns kick in.

But if you start with a lower bit rate version, you can't get those dynamics back (sure you can run it through an expander, but it won't be the same).
drtedtan is offline   Reply With Quote
Old 10-22-2015, 08:34 AM   #80
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 24,755
Default

Quote:
Tinnitus is not actually sound, it is imaginary, and as such, it does not interfere with our ability to hear real sounds.
I'd like to see data to back that up since I have a very light touch of tinnitus and it absolutely interferes with real sounds of the same frequency.
__________________
If it requires a null test to find it, it is by definition minuscule.
karbomusic is online now   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 08:45 PM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2019, vBulletin Solutions Inc.