Login

Sample Rate Discussion (off of AITB Germanos EQ Thread)

If you need to write about anything else please do it here..

Re: Sample Rate Discussion (off of AITB Germanos EQ Thread)

Postby yr » Wed Feb 16, 2011 6:56 pm

I'm offering a real world relevant test which might be useful to (some) users. There is enough additional measured information provided in this thread alone to make it more then anecdotal. Niklas reffered to the Nebula src as "unacceptable" demonstrating phase-shifts, and I showed the way this src changes the harmonic structure in some presets and creates aliasing noise. Your decision to ignore this data is a personal choice that has little to do with "science".
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

Re: Sample Rate Discussion (off of AITB Germanos EQ Thread)

Postby ngarjuna » Wed Feb 16, 2011 7:00 pm

yr wrote:I'm offering a real world relevant test which might be useful to (some) users. There is enough additional measured information provided in this thread alone to make it more then anecdotal. Niklas reffered to the Nebula src as "unacceptable" demonstrating phase-shifts, and I showed the way this src changes the harmonic structure in some presets and creates aliasing noise. Your decision to ignore this data is a personal choice that has little to do with "science".


Science? Here's the science:

http://www.lavryengineering.com/documen ... Theory.pdf

You can argue with Dan Lavry (I have seen many try) but please do post your credentials so we can evaluate since the agenda you're pushing is pretty much at odds with his conclusions.
User avatar
ngarjuna
Expert
Expert
 
Posts: 778
Joined: Tue Mar 30, 2010 5:04 pm
Location: Miami

Re: Sample Rate Discussion (off of AITB Germanos EQ Thread)

Postby yr » Wed Feb 16, 2011 7:18 pm

I find your argumentation is becoming more and more personal and irrelevant. Let me state this once more: I'm not arguing people should use any specific sample rate or that 96kHz is better then 44.1kHz. I don't think that this is the real issue of this thread.

The issue (as I see it) is whether multiple src conversions done by Nebula (typical case scenario) will give inferior results to a 2 step conversion of the audio material. This has little to do with Mr. Lavry and everything to do with the way this thread could help people make some sense of the Nebula src debate. Let me remind you that 2 of the most respected 3 party developers seem to be at a complete disagreement about this issues, so it might be a good idea to check things for yourself.
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

Re: Sample Rate Discussion (off of AITB Germanos EQ Thread)

Postby ngarjuna » Wed Feb 16, 2011 7:26 pm

yr wrote:I find your argumentation is becoming more and more personal and irrelevant. Let me state this once more: I'm not arguing people should use any specific sample rate or that 96kHz is better then 44.1kHz. I don't think that this is the real issue of this thread.

The issue (as I see it) is whether multiple src conversions done by Nebula (typical case scenario) will give inferior results to a 2 step conversion of the audio material. This has little to do with Mr. Lavry and everything to do with the way this thread could help people make some sense of the Nebula src debate. Let me remind you that 2 of the most respected 3 party developers seem to be at a complete disagreement about this issues, so it might be a good idea to check things for yourself.


Right, but you're using SRC from your host (the final conversion to 44.1) on top of any SRC that you're evaluating inside of Nebula. You're comparing apples to oranges. It's quite possible (esp with no evidence to the contrary) that it's the host's SRC to 44.1 which is causing what you don't like; that's just as likely as it being the Nebula SRC. In which case working at 96k will do absolutely nothing for you.

The problem is: there is no need for an SRC Nebula debate. If anyone feels the issue is so pressing, then they can upsample and go to all the extremes necessary to process at whatever sample rate they want. The only reason for a "debate" is to try and evangelize others to a particular viewpoint, a move which is unnecessary and potentially confusing to the many users here who don't have 10+ years of dealing with digital audio (and may be unfamiliar with the quagmire of misinformation spouted off about sample rate and SRC comparisons).

When you throw around misinformation - such as claiming to hear differences when said differences could have in fact been caused by multiple deficiencies in the testing methodology - not only does it not assist anyone in understanding the complexities and ramifications of sample rates, it further obfuscates the truth and reinforces various unfortunate myths (even when that was not the poster's intention).
User avatar
ngarjuna
Expert
Expert
 
Posts: 778
Joined: Tue Mar 30, 2010 5:04 pm
Location: Miami

Re: Sample Rate Discussion (off of AITB Germanos EQ Thread)

Postby yr » Wed Feb 16, 2011 7:48 pm

I've re-read everything I wrote and I can't see any trace of misinformation. When I hear a difference I report it as my subjective experience and when I see a bug (or a shortcoming of the Nebula src) I provide measurements. I also never said I was relying on my host DAW for the conversion. I use iZotope RX to upsample and down-sample the material so I'm comparing my (best) src converter using a 2 step conversion to the Nebula src. Any user can download the free voxengo src if they don't have other choices. In short: a 44.1kHz project (and files) using 96kHz presets vs 96kHz (externally upsampled) files with the same presets in a 96kHz project. The resulting file is converted externally to 44.1kHz.

What you suggested is exactly what I do- I upsample audio material prior to Nebula when I can, to bypass the internal src. I believe that Giancarlo himself described this src as very good but still a compromise between speed and accuracy. If you agree that any src will degrade sound, you will need the Nebula src to be much superior to any external src solutions, given the fact that there will be many more conversions in a typical project. Simple logic anchored in real world experience...
Last edited by yr on Wed Feb 16, 2011 9:06 pm, edited 1 time in total.
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

Re: Sample Rate Discussion (off of AITB Germanos EQ Thread)

Postby Stolle » Wed Feb 16, 2011 8:55 pm

I'm definitely going to devote an hour to test this with my own setup. I use lots and lots of Nebula instances so it's easily worth an hour of my time.
Stolle
User Level VI
User Level VI
 
Posts: 68
Joined: Mon Mar 29, 2010 3:27 am

Re: Sample Rate Discussion (off of AITB Germanos EQ Thread)

Postby biomuse » Wed Feb 16, 2011 11:57 pm

Giancarlo,

Would it make sense (when you have time) to implement an oversampling parameter in Nebula, such that the input signal could optionally be upsampled/processed/downsampled within the plug to take advantage of Nebula 96k patch SR, no matter what the host session SR is?

If this makes sense, advantages to this approach could include:

1) the ability to use SRC appropriate for audio instead of Nebula patch files, which would possibly allow SRC that offers a better quality/speed bargain;

2) easier to manage CPU load in 44.1/48Khz sessions, since the user could simply switch on Nebula oversampling when it's time to bounce or freeze

On the other hand, if this makes no sense (i.e., if it is no better than what we already have), then what is the difference between "oversampling" in plugs such as The Glue and Aether 1.5, where switching on "oversampling" for the plug produces a clear improvement in sound (even in a 44.1 host session), vs. something like this if implemented in Nebula?

In other words, is there a difference between "oversampling" as implemented in algorithmic plugs vs. convolution plugs?
biomuse
User Level X
User Level X
 
Posts: 119
Joined: Tue Dec 14, 2010 9:37 am

Re: Sample Rate Discussion (off of AITB Germanos EQ Thread)

Postby rrrobo » Thu Feb 17, 2011 4:15 am

ngarjuna wrote:SRC/sample rate changes have never made a good mix sound bad or a bad mix sound good.

I see your point but if I can be pedantic.. sample rate changes may have made a good mix sound slightly less good and.. sure they can't make a bad mix sound good but they could make the fidelity of the audio in the bad mix sound slightly better.

I can't understand how I can hear a difference between 96-96 and 44-44. Surely, at age 31, my hearing can't reach those upper frequencies. What gives?


Hi Stolle,

This is great - more people should do what you did and check things out for themselves rather than blindly recycle the old "I can't hear above 20khz so why bother" doctrine. I think you're right, there's something there that humans can perceive which may not be just frequency related (although I think part of it must be) and we would do well to try to understand what exactly it is. People are always surprised when I show them waveforms in spectral views of what goes on above 20khz, and that they can perceive a difference when A/Bing a full range 96khz wav against a 20khz brickwall castrated (as Alex put it) version. Closely A/Bing 24 bit wavs with reference monitoring, doing lots of testing, being inquisitive and cleaning out your ears are all good, positive, and worthwhile things to do.

PS I'm 35 and I still have pretty good hearing, I guess it depends on the person though.

Cheers
Mark
rrrobo
User Level V
User Level V
 
Posts: 52
Joined: Mon Jun 07, 2010 10:20 am

Re: Sample Rate Discussion (off of AITB Germanos EQ Thread)

Postby Stolle » Thu Feb 17, 2011 4:45 am

rrrobo wrote:
Hi Stolle,

This is great - more people should do what you did and check things out for themselves rather than blindly recycle the old "I can't hear above 20khz so why bother" doctrine. I think you're right, there's something there that humans can perceive which may not be just frequency related (although I think part of it must be) and we would do well to try to understand what exactly it is. People are always surprised when I show them waveforms in spectral views of what goes on above 20khz, and that they can perceive a difference when A/Bing a full range 96khz wav against a 20khz brickwall castrated (as Alex put it) version. Closely A/Bing 24 bit wavs with reference monitoring, doing lots of testing, being inquisitive and cleaning out your ears are all good, positive, and worthwhile things to do.

PS I'm 35 and I still have pretty good hearing, I guess it depends on the person though.

Cheers
Mark

Hi Mark and thanks for your reply. It truly is an important lesson to use your own judgement and not blindly accept what others say and write. It's a lesson I find myself in need of learning over and over again through life. In this particular case though, I accept Yr's argument that the likeliest reason for the big difference between the files comes from the fact that the material was originally recorded in 96 and then downsampled. In order to really experience the difference between 96 and 44.1 I'd need to follow Yr's instructions and record something repeatable at the different SR.

I did accept blindly Alex' statement that a library sampled at 96 sounds better downsampled to 44.1 than one sampled natively at 44.1. All the data in this thread points to this not necessarily being the case, and thus call for further investigations. I will sure give it an hour of my time this weekend. I love to learn.
Stolle
User Level VI
User Level VI
 
Posts: 68
Joined: Mon Mar 29, 2010 3:27 am

Re: Sample Rate Discussion (off of AITB Germanos EQ Thread)

Postby highvoltage » Thu Feb 17, 2011 12:53 pm

For me Nebula sounded awesome and superior to any other plugins before this topic came out, and it still sounds the same after reading this.

I must be deaf but i don't give a shit for these nuances. You talk about these issues like 'oh my god something is terribly wrong' when in all honesty, you really have to concentrate to hear the difference, and thats SO much smaller than making a 0.5 db cut/boost with an EQ.

I have a suspicion that most of us couldn't tell which is which in AlexB's examples in a blind test.
Fancy a poll? :)
highvoltage
Member
Member
 
Posts: 235
Joined: Tue Mar 30, 2010 9:44 pm

PreviousNext

Return to Other things

Who is online

Users browsing this forum: No registered users and 4 guests