Don't let Live resample your audio!

kineticUk · Post by **kineticUk** » Tue Aug 07, 2007 9:27 pm

Just thought I'd add, looking at the results of the BIAS Matlab tests...
Adobe Audition seems to get this conversion down in all their tests. Excellent.
Edit: Those tests posted by funkstar http://src.infinitewave.ca/ seem to be a lot better than the BIAS ones, my bad.

Funkstar De Luxe · Post by **Funkstar De Luxe** » Tue Aug 07, 2007 9:31 pm

kineticUk wrote:Just thought I'd add, looking at the results of the BIAS Matlab tests...
Adobe Audition seems to get this conversion down in all their tests. Excellent.
Edit: Those tests posted by funkstar seem to be a lot better than the BIAS ones, my bad.

Yes, but the pre/post filter setting only. No filter seems to have some high freq problems

Tone Deft · Post by **Tone Deft** » Tue Aug 07, 2007 9:32 pm

I'll try this later.

may I comment in a polite manner that the thread title is a bit strong for the reality of the situation. finding plots is one thing, understanding the ramifications is another. that's kinda the rubbing point for me on this.

Funkstar De Luxe · Post by **Funkstar De Luxe** » Tue Aug 07, 2007 9:36 pm

Tone Deft wrote:I'll try this later.

may I comment in a polite manner that the thread title is a bit strong for the reality of the situation. finding plots is one thing, understanding the ramifications is another. that's kinda the rubbing point for me on this.

It is a bit strong... But then again Ableton is an audio tool, that's very expensive and completely out classed by a free download. No one is upset, and no one expected any better, but if you are using Live for any kind of professional audio recording, I suggest you do not let it resample. It's not devistating, but it can be the difference between a beautifully natural sounding record and a cold digital sound

Tone Deft · Post by **Tone Deft** » Tue Aug 07, 2007 9:42 pm

Funkstar De Luxe wrote:it can be the difference between a beautifully natural sounding record and a cold digital sound

no, it's the difference between a perfectly plotted THD graph and a strangely plotted THD graph. if you posted two files for people to listen to and vote on which sounded 'off' and the resampled version clearly won, then your findings could back that up. your ears are the only thing that really matter.

I guess where I'm headed with this is along the lines of 44.1kHz sampling vs. 192kHz sampling, the numbers are there, but do they REALLY matter in terms of sound quality?

it's a difficult topic, human hearing is far from perfect.

edited due to cross posting...

Funkstar De Luxe · Post by **Funkstar De Luxe** » Tue Aug 07, 2007 9:47 pm

Tone Deft wrote:
Funkstar De Luxe wrote:it can be the difference between a beautifully natural sounding record and a cold digital sound
no, it's the difference between a perfectly plotted THD graph and a strangely plotted THD graph. if you posted two files for people to listen to and vote on which sounded 'off' and the resampled version clearly won, then your findings could back that up.

I guess where I'm headed with this is along the lines of 44.1kHz sampling vs. 192kHz sampling, the numbers are there, but do they REALLY matter in terms of sound quality?

it's a difficult topic, human hearing is far from perfect.

Sadly it is not. As you can see from the graph a lot of the artefacts are in the audible range. Resample the files I posted and listen to them. You can do all this yourself, there no real argument here.
All audio that is going to be processed should be recorded at the highest possible rate. It's just common sense. But yes, 44.1/16 is perfectly adequate for playback if the audio has been downsampled correctly.

Do you work for Ableton or something?

Tone Deft · Post by **Tone Deft** » Tue Aug 07, 2007 9:52 pm

yeah, I should listen. :rolls eyes at self:

I don't work for ableton but wouldn't mind the extra paycheck. I am an annoying fanboy.

FWIW I work on audio hardware for a living, when I test stuff I always find the limitations of the hardware, nothing is ever perfect but I have to be able to quantify quality which never feels good because perfection is the only answer. the other part of that is learning what's acceptable and that takes time and exposure to lots of situations.

I never know the age, maturity or experience of other posters, I don't mean to imply you don't have a point or understand the results.

Meef Chaloin · Post by **Meef Chaloin** » Tue Aug 07, 2007 9:53 pm

this is fuct

Tone Deft · Post by **Tone Deft** » Tue Aug 07, 2007 9:54 pm

Meef Chaloin wrote:this is fuct

can you explain why?

Meef Chaloin · Post by **Meef Chaloin** » Tue Aug 07, 2007 9:56 pm

Tone Deft wrote:
Meef Chaloin wrote:this is fuct
can you explain why?

yeah

Funkstar De Luxe · Post by **Funkstar De Luxe** » Tue Aug 07, 2007 10:00 pm

Tone Deft wrote:yeah, I should listen. :rolls eyes at self:

I don't work for ableton but wouldn't mind the extra paycheck. I am an annoying fanboy.

FWIW I work on audio hardware for a living, when I test stuff I always find the limitations of the hardware, nothing is ever perfect but I have to be able to quantify quality which never feels good because perfection is the only answer. the other part of that is learning what's acceptable and that takes time and exposure to lots of situations.

I never know the age, maturity or experience of other posters, I don't mean to imply you don't have a point or understand the results.

My whole point of posting this was to say "hey, Ableton isn't too great at this; here's a free tool that's better". I work in quite a specialized area of audio, so these topic are right up my street (not to mention I also have a nerdy interest in them). No nothing is perfect, but doesn't mean we shouldn't try

However, I DO think that when using a DAW, rendering audio should be taken as accurately as possible. It is the single most important step. I doubt this is a big problem though, I mean how many people actually render at a different rate for recording?

Tone Deft · Post by **Tone Deft** » Tue Aug 07, 2007 10:01 pm

Meef Chaloin wrote:
Tone Deft wrote:
Meef Chaloin wrote:this is fuct
can you explain why?
yeah

totally.

I'm not even going to sell my copy of Live to someone, that would be unethical.

downloading Logic Pro now, their purchasing system is QUICK!!

Tone Deft · Post by **Tone Deft** » Tue Aug 07, 2007 10:04 pm

Funkstar De Luxe wrote:
Tone Deft wrote:yeah, I should listen. :rolls eyes at self:

I don't work for ableton but wouldn't mind the extra paycheck. I am an annoying fanboy.

FWIW I work on audio hardware for a living, when I test stuff I always find the limitations of the hardware, nothing is ever perfect but I have to be able to quantify quality which never feels good because perfection is the only answer. the other part of that is learning what's acceptable and that takes time and exposure to lots of situations.

I never know the age, maturity or experience of other posters, I don't mean to imply you don't have a point or understand the results.
My whole point of posting this was to say "hey, Ableton isn't too great at this; here's a free tool that's better". I work in quite a specialized area of audio, so these topic are right up my street (not to mention I also have a nerdy interest in them). No nothing is perfect, but doesn't mean we shouldn't try However, I DO think that when using a DAW, rendering audio should be taken as accurately as possible. It is the single most important step. I doubt this is a big problem though, I mean how many people actually render at a different rate for recording?

very cool. it shouldn't matter but knowing you have a background in all this is good to know, I can forget some of the questions I might have. you just never know what the person behind the letters on the screen is about.

it'd be interesting to do a head to head comparison on this. I do appreciate people pointing out the flaws in Live, it's an education and I should know my tool better.

laird · Post by **laird** » Tue Aug 07, 2007 10:17 pm

Judging by Funkstar's data, this is more than a "quality" issue but a massive artifact (as in: changing what should be there, not just adding to what is there) issue.

His sine wave, recorded at 96KHz, which descends in pitch from higest to lowest... when converted to 44.1 Khz, is no longer a sine wave (but a complex waveform with lots of weaker over and undertones), and no longer descends in pitch... it actually rises in pitch for several seconds.

I will have to test this out tonight, out of curiosity.. but from the pictures, the converted wave file isn't just "colored" but radically changed. FWIW, I've never used the sample rate conversions.... just bit depth conversions... something about knowing what my final rate is going to be at the end yadda yadda didn't someone mention something similar to that?

But.. hearing will be believing (or disbelieving).

[edit] As for the strength of the forum title... maybe you should rename it "Fruity Loops is better than Live", you'll get less controversy

[/edit]

Tone Deft · Post by **Tone Deft** » Tue Aug 07, 2007 10:25 pm

ugh, can't open the .rar file.