Login

So, 96 or 44? What REALLY is the best?

Tips & tricks, working results, technical support

So, 96 or 44? What REALLY is the best?

Postby russianpolecat » Mon Apr 22, 2013 1:45 pm

OK, so on another thread I was talking about workflow and the fact that I was wondering about the best way to incorporate libraries that have different sample rates (I mainly work at 96 - because I mainly have 96k libraries). Now, a number of responses got me thinking because they suggested that there really isn't that much of a difference between working at 44 and 96 in Nebula (if at all - in fact some people think certain music sounds better, or prefer the sound at 44). Now, my take on this is that I don't disagree with the above. For rock and pop etc. I doubt you'd notice much of a difference at 96 over 44 (if at all). But that view is predominantly to do with the 44 versus 96 debate that exists outside of Nebula. I was always under the impression that the reason most people sampled stuff at 96 and recommended working at that sample rate was because it is more accurate to the nuances of the hardware - in other words a separate (context specific) 44 versus 96 debate. So, what are peoples thoughts about this? Is there really a difference with Neb 96? Or is this the same 44 versus 96 debate that exists 'out there'? Or to put it another way, why work at 96 at all?? (other than with respect to reverbs maybe)

;)
russianpolecat
User Level XI
User Level XI
 
Posts: 167
Joined: Thu Sep 08, 2011 10:16 pm

Re: So, 96 or 44? What REALLY is the best?

Postby david1103 » Mon Apr 22, 2013 6:05 pm

There are a whole bunch of tangled up arguments around 96 vs 44.1 that need to be separated and addressed independently.

I wish someone would have standardized DAW sample rates at about 60k and we would not be here talking about this!

1. Sound recording/playback at 96k vs 44.1

Many think 96k sounds better, but a common explanation is that this is because of the filter in the DAC. Digital audio needs a steep filter at half of its sampling rate, this is at its 'folding frequency'. At 44.1 this filter is closer to human hearing so its possible to hear artifacts.

A crummy DAC is going to sound better at 96k as its crummy Nyquist filter will pollute the sound more at 44.1 as it is closer to human hearing.

The only way to test the 'real' difference well is with a very high quality DAC and see if you hear a difference switching between rates with 44.1/96k recorded identical material.

2. Computation in plug-ins at 96k and aliasing

Plug-ins should sound better at 96k if they are able to support it for reasons of aliasing. Nebula does not suffer that badly with aliasing, I have not really sat down and tested an identical natively sample 44.1 vs natively sampled 96k program, but aliasing will be different for different types of program depending on the prog rate.

3. Sample Rate Conversion

Best avoided in Nebula if possible, so if it is needed to load your programs this has to be factored into the balance.

Does 96k have a special benefit outside of these factors? I think people will have different opinions. I spoke to AITB and those guys are 44.1 all the way, AlexB stated in a thread he though 96k was much better.

I think its a shame most developers went straight to 96k as the advantage is quite small vs CPU use.

Letting Nebula itself down-sample is not good enough for my purposes. When a developer (like cupwise) uses SoX to downsample the original 96k test tone results to create 44.1 programs I think the results are great.

To sum up, 'what REALLY' is the best?' Technically 96k is best sound quality but worst work flow. The question is how much better sound quality? Maybe that depends on the ADC used to sample it.

If you are using 44.1 in your project then native 44.1 is best. 96k then native 96k is best. I would always go native sample rate as the first quality consideration, but then the quality of the program itself has so many more factors!

As ever... more scientific tests needed ;)
User avatar
david1103
Beta Tester
Beta Tester
 
Posts: 516
Joined: Wed Mar 31, 2010 12:26 am

Re: So, 96 or 44? What REALLY is the best?

Postby ngarjuna » Mon Apr 22, 2013 8:52 pm

david1103 wrote:There are a whole bunch of tangled up arguments around 96 vs 44.1 that need to be separated and addressed independently.

I wish someone would have standardized DAW sample rates at about 60k and we would not be here talking about this!

1. Sound recording/playback at 96k vs 44.1

Many think 96k sounds better, but a common explanation is that this is because of the filter in the DAC. Digital audio needs a steep filter at half of its sampling rate, this is at its 'folding frequency'. At 44.1 this filter is closer to human hearing so its possible to hear artifacts.

A crummy DAC is going to sound better at 96k as its crummy Nyquist filter will pollute the sound more at 44.1 as it is closer to human hearing.

The only way to test the 'real' difference well is with a very high quality DAC and see if you hear a difference switching between rates with 44.1/96k recorded identical material.

2. Computation in plug-ins at 96k and aliasing

Plug-ins should sound better at 96k if they are able to support it for reasons of aliasing. Nebula does not suffer that badly with aliasing, I have not really sat down and tested an identical natively sample 44.1 vs natively sampled 96k program, but aliasing will be different for different types of program depending on the prog rate.

3. Sample Rate Conversion

Best avoided in Nebula if possible, so if it is needed to load your programs this has to be factored into the balance.

Does 96k have a special benefit outside of these factors? I think people will have different opinions. I spoke to AITB and those guys are 44.1 all the way, AlexB stated in a thread he though 96k was much better.

I think its a shame most developers went straight to 96k as the advantage is quite small vs CPU use.

Letting Nebula itself down-sample is not good enough for my purposes. When a developer (like cupwise) uses SoX to downsample the original 96k test tone results to create 44.1 programs I think the results are great.

To sum up, 'what REALLY' is the best?' Technically 96k is best sound quality but worst work flow. The question is how much better sound quality? Maybe that depends on the ADC used to sample it.

If you are using 44.1 in your project then native 44.1 is best. 96k then native 96k is best. I would always go native sample rate as the first quality consideration, but then the quality of the program itself has so many more factors!

As ever... more scientific tests needed ;)

This is a pretty good summation. Personally I don't find Nebula SRC to be a very big deal; some people hate it. /shrug I would say that sample rate considerations (including Nebula's internal SRC) are probably the least significant factor of your mix. Moving your head an inch to the left will likely have a more pronounced effect.

I'd do a blind listening test between a program at 96 and then downsampled to 44.1. If you can pick the difference blind then you have your answer (and likewise if you can't).
User avatar
ngarjuna
Expert
Expert
 
Posts: 778
Joined: Tue Mar 30, 2010 5:04 pm
Location: Miami

Re: So, 96 or 44? What REALLY is the best?

Postby richie43 » Tue Apr 23, 2013 12:26 am

To add to ngarjuna's response;
I don't worry about it too much. I try to use presets with the same sample rate as my project, but I certainly wouldn't bother with changing sample rates in the project, rendering, and changing again. No thanks. I also don't think twice if I really want a specific preset that happens to be 96kHz and the project is in 44.1..... I imagine that the biggest difference will be how you feel about it. I just want to make and produce great audio. Go make music.
The Sounds of the Hear and Now
http://soundyaudio.com/
richie43
Expert
Expert
 
Posts: 4856
Joined: Fri Feb 04, 2011 8:47 pm
Location: Minnesota, USA

Re: So, 96 or 44? What REALLY is the best?

Postby babiuk » Tue Apr 23, 2013 12:37 am

Well, I do think this way too. I´m not able to pick the difference because I don´t even try it, I pay attention to music and that´s all. Maybe I´m losing something but it´s true that there are things which really affects in a really noticeable way.

ngarjuna wrote:Moving your head an inch to the left will likely have a more pronounced effect.


+1 to this. :D
User avatar
babiuk
User Level XI
User Level XI
 
Posts: 199
Joined: Tue Oct 19, 2010 11:16 am

Re: So, 96 or 44? What REALLY is the best?

Postby mathias » Tue Apr 23, 2013 1:20 am

great answers all! :!:

i would add, that for me the important point is, that the music i am mixing, can tell its story. and there are so many ways it can sound and still telling the same story ...
so i am going with what i and my clients like and that's it. mission done.

the thing that may be of bigger interest for the whole work is, if you have a hearing situation, that is clear/transparent/balanced enough, so you can really judge what you hear and make the right decisions.
this is so much more important, than all the 44,1 versus 96 khz hairsplitting.

best,
mathias
system 1: windows 8 32 bit - samplitude prox, tracktion6, reaper
system 2: mac osx yosemite - reaper(32+64bit), tracktion6(32+64bit)

both systems on: macbook pro (late 2009), core 2 duo 3,06 ghz, 4 gb ram, graphic: nvidia geforce 9600M GT 512 MB
mathias
Expert
Expert
 
Posts: 2114
Joined: Wed Mar 31, 2010 12:25 am
Location: South-West Germany

Re: So, 96 or 44? What REALLY is the best?

Postby russianpolecat » Tue Apr 23, 2013 11:27 am

Yeah, good stuff here :D
russianpolecat
User Level XI
User Level XI
 
Posts: 167
Joined: Thu Sep 08, 2011 10:16 pm

Re: So, 96 or 44? What REALLY is the best?

Postby russianpolecat » Tue Apr 23, 2013 11:29 am

david1103 wrote:There are a whole bunch of tangled up arguments around 96 vs 44.1 that need to be separated and addressed independently.

I wish someone would have standardized DAW sample rates at about 60k and we would not be here talking about this!

1. Sound recording/playback at 96k vs 44.1

Many think 96k sounds better, but a common explanation is that this is because of the filter in the DAC. Digital audio needs a steep filter at half of its sampling rate, this is at its 'folding frequency'. At 44.1 this filter is closer to human hearing so its possible to hear artifacts.

A crummy DAC is going to sound better at 96k as its crummy Nyquist filter will pollute the sound more at 44.1 as it is closer to human hearing.

The only way to test the 'real' difference well is with a very high quality DAC and see if you hear a difference switching between rates with 44.1/96k recorded identical material.

2. Computation in plug-ins at 96k and aliasing

Plug-ins should sound better at 96k if they are able to support it for reasons of aliasing. Nebula does not suffer that badly with aliasing, I have not really sat down and tested an identical natively sample 44.1 vs natively sampled 96k program, but aliasing will be different for different types of program depending on the prog rate.

3. Sample Rate Conversion

Best avoided in Nebula if possible, so if it is needed to load your programs this has to be factored into the balance.

Does 96k have a special benefit outside of these factors? I think people will have different opinions. I spoke to AITB and those guys are 44.1 all the way, AlexB stated in a thread he though 96k was much better.

I think its a shame most developers went straight to 96k as the advantage is quite small vs CPU use.

Letting Nebula itself down-sample is not good enough for my purposes. When a developer (like cupwise) uses SoX to downsample the original 96k test tone results to create 44.1 programs I think the results are great.

To sum up, 'what REALLY' is the best?' Technically 96k is best sound quality but worst work flow. The question is how much better sound quality? Maybe that depends on the ADC used to sample it.

If you are using 44.1 in your project then native 44.1 is best. 96k then native 96k is best. I would always go native sample rate as the first quality consideration, but then the quality of the program itself has so many more factors!

As ever... more scientific tests needed ;)
That's great. Really informative!
russianpolecat
User Level XI
User Level XI
 
Posts: 167
Joined: Thu Sep 08, 2011 10:16 pm

Re: So, 96 or 44? What REALLY is the best?

Postby Tim Petherick » Wed May 01, 2013 2:15 am

Edit - I just noticed David saying similar things to what I'm posting on this thread. I didn't see that!

The Filters is one of the reasons why 44.1 does not sound quite right.

But going too high you can hear dynamics getting softened because of distortion.

A good way to tell is by doing direct a/b's,so source to speakers then source recorded then playback.

In general i tend to believe that 48khz and 88.2 are the best ,you want to stay clear of that filter but not go so high that you get distortions. 48 sounds better in dynamics and 88.2 slightly smoother. 192khz you think your hearing more analog but really it's more of the convertor going ahhhhh. It's basically sounding softer so you may think it's more analog.


I don't get the 44.1 and 96 debate completely , why make that jump and forget the other sample rates? I think 48 would be better as a standard and maybe 88.2

And yes conversion sounds bad if your doing it to multi stems. The more tracks you have the more you'll notice it. depending on conversion either, dynamics suffer or you getting ringing from steep filters neither sound good.
Last edited by Tim Petherick on Wed May 01, 2013 10:07 pm, edited 2 times in total.
User avatar
Tim Petherick
Expert
Expert
 
Posts: 1351
Joined: Sat Apr 17, 2010 4:07 pm
Location: Bath , Uk

Re: So, 96 or 44? What REALLY is the best?

Postby yr » Wed May 01, 2013 9:36 am

I guess it was mentioned before, but integrating an HQ src in NAT could solve many problems. There are a lot of libraries that are available only as 96khz for instance.
Reuven | post-production & sound-design | scenography |
website | nebula presets
User avatar
yr
Beta Tester
Beta Tester
 
Posts: 438
Joined: Mon Mar 29, 2010 9:04 am
Location: Amsterdam

Next

Return to Working with Nebula

Who is online

Users browsing this forum: No registered users and 6 guests