ok I m not sure anymore : sample rate

[dum]

Hold on boys, hold on, just one more minute then you can go at it.

Dum, so as an example of what you do, you may use the sample rate creatively as an effect on certain VST instruments but also becasue some effects and instruments simply hit their sweet spot at higher sample rates.

Yes to the sample rate creatively bit (but then I'll render the wav to whatever the lower sample rate is, take it to wave-editor and convert it's sample rate to whatever I'm working at - 88.2 in this case), and Yes to the FX & Instruments hitting a sweet spot at higher sample rates. but when processing audio recordings, I suggest converting them to the higher sample rate first with a nice SRC. or recording them at the high-sample rate to begin with if possible. And most importantly of all, using your ears to tell you if any of that is worth doing. I have no idea what kind of material you work with, really.

Then you render a high bit depth high sample rate stereo file, dump it into wave editor and run it through Ozones SRC to end up with a high bit depth, 44.1 stereo file ready for the mastering?

when I'm ready to render the finished wav in ableton, I'll render at 88.2
then I'll do whatever mastering tweeks I need to do, in either ableton or wave-editor - rendering again at 88.2
Now that it's finished, I'll reduce it to 44.1hz 16 bit in Wave-Editor with the nice SRC and dithering options.
in your case, I guess you could use Ozone. as it's the Izotope algos that I'm using wave-editor for (besides the fact it's an awesome wave editor, of course)

[Tone Deft]

a good sample-rate-converter will know what half of the information to bin

that's the point of choosing a good SRC

it doesn't know which half it just does it right to begin with. there is no 'half.'

btw, can you elaborate on this. Besides the obvious 'half' misnomer, dave was talking purely in terms of data there.

what do you mean 'it just does it right to begin with' ?
Converting the sample rate of an 88.2hz wave file to 44.1hz wave file does indeed bin information. And the quality of the algo will determine what information is binned.

So...what exactly are you talking about here ?

agreed on misnomer. the semantics can be awkward.

it's the ultrasonic conspiracy theories. info above our hearing level being crucial to the processing and final output. for example, a large advantage to higher sample rates is using filters with a slow cutoff which in turn might allow you to make a more phase accurate filter or fewer MIPs. all those decisions that go into coding a DAW that can make one different from another. the entire audio chain would have to be designed to pass ultrasonic frequencies.

again, since I'm posting... the errors in those files are way too loud, the test is probably flawed. BUT it's data and should be reproduced and props for being one of the few if only person to do the work. we have the tools to answer this, right? I gotta go to work.

[UKRuss]

Hold on boys, hold on, just one more minute then you can go at it.

Dum, so as an example of what you do, you may use the sample rate creatively as an effect on certain VST instruments but also becasue some effects and instruments simply hit their sweet spot at higher sample rates.

Yes to the sample rate creatively bit (but then I'll render the wav to whatever the lower sample rate is, take it to wave-editor and convert it's sample rate to whatever I'm working at - 88.2 in this case), and Yes to the FX & Instruments hitting a sweet spot at higher sample rates. but when processing audio recordings, I suggest converting them to the higher sample rate first with a nice SRC. or recording them at the high-sample rate to begin with if possible. And most importantly of all, using your ears to tell you if any of that is worth doing. I have no idea what kind of material you work with, really.

Then you render a high bit depth high sample rate stereo file, dump it into wave editor and run it through Ozones SRC to end up with a high bit depth, 44.1 stereo file ready for the mastering?

when I'm ready to render the finished wav in ableton, I'll render at 88.2
then I'll do whatever mastering tweeks I need to do, in either ableton or wave-editor - rendering again at 88.2
Now that it's finished, I'll reduce it to 44.1hz 16 bit in Wave-Editor with the nice SRC and dithering options.
in your case, I guess you could use Ozone. as it's the Izotope algos that I'm using wave-editor for (besides the fact it's an awesome wave editor, of course)

Great, thanks. I'll do some experimenting.

[davepermen]

when a DSP algorithm is created the author can make the decision to band limit the process to save processing power, cut out noise, whatever. BUT the decision has to be made to leave THE ENTIRE audio path to not be band limited. you keep talking about more info being in the signal, but all that extra info is above what we can hear. I don't buy into this.

if i play back at half speed, it's NOT above what we can hear. at least if you play back with vinyl mode, the numbers and actual hearing parts are that simple. you go to half speeds, all frequencies go to half, too. all the unhearable frequencies below the 88.2 niquist (means 44.1khz) get down to hearable frequencies below the 44.1 niquist (means 22.05khz).

now a time stretch algorithm would pitch that up again obviously. but still, it would have all that frequency information from 11.025khz to 22.05khz on the stretched sample to better pitch up. and while it all might get lossed, it depends on the stretching alorithm, of course.

the point stands for vinyl/pitch mode.

and no matter how to spin, you lose exactly half the samples. instead of having 88200 samples per second, you have 44100 samples. wetter those additional samples are of USE or not, that's another discussion. but the more, the better. if your disk space and your soundcard and your cpu can handle it, it's better to have it all. even if it might never make a difference in your use case.

[The Carpet Cleaner]

All right all right all right.

You guys didn't really paid attention to what I uploaded before on the first page.

So I'm going to make it even simpler :
I've extracted the Lead track from my project and made it a new Project.
So you can download this project here , it's 90kb ! not even 1mb so come on. Also I'm with Suite so there is multiband dynamic, not sure it will work with regular Live version.
http://dl.dropbox.com/u/2968492/Test%20 ... roject.zip

It's just a 1 second .wav sample at 44.1khz 16 bits into a Sampler, with some effects. On my imac, the sound is drasticaly different between 44.1khz and 88.2khz.

Do you have the same result. And explain why then (if you can ).

Also, now : tkx guys, I now understand that the sample rate setting in Live prefs, and during the export are two different things. that's great.

Lastly : during my test, there is no more difference once I've export everything at 44.1khz 16bit.
So I'm wondering, is it common to send un-mastered track to mastering studio, with a setting at 88.2khz/24bit. Could they do something much much better than me with Live in term of downsampling to 44.1khz before release ?

Cheers !

[Tone Deft]

if i play back at half speed, it's NOT above what we can hear. at least if you play back with vinyl mode, the numbers and actual hearing parts are that simple. you go to half speeds, all frequencies go to half, too. all the unhearable frequencies below the 88.2 niquist (means 44.1khz) get down to hearable frequencies below the 44.1 niquist (means 22.05khz).

I hear where you're coming from, I think it's speculation. please provide an example.

I tried it with Operator, I spit out white noise to an audio track. I played it back and watched Spectrum. repitching down, there were no higher frequency components filling in the top end. BUT Operator's white noise might be band limited. SO... I tried it with a microphone but my voice doesn't make 20kHz components.

so... where to get or create wav files with frequency content over 20kHz?

it's too early for me.


1ino1eum - cool, good to kick the tires on this siht.

[davepermen]

not having any equipment to record myself, no, haven't made any tests in live. i mentoyed in the original thread that i was talking about timestretching/repitching for tracks at non-original speed in traktor, and mixing them. there was a noticeable difference, audio quality enhanced.

and i know that numerically, my stuff is sound (no pun intended).

question is really, how much additional signal info is in those samples. have to get some recording equipment to test..

and yeah, i guess operator doesn't create sounds above audible frequencies much.

recording operators noise, repitching it to twice the speed should create a sample with the high frequencies. then resampling, and repitching down should show if live is actually able to store the information in those higher frequencies. if not, 44.1khz sample rate is enough, as live is the "bottleneck" there.


oh, and, good morning

[dum]

[leedsquietman]

Both Sample rate conversion and dithering (more specifically the process of reducing wordlength, i.e. going from 32 or 24 to 16 bit) bin information.

The processes are designed to be as transparent as possible, so as to have a minimal effect on audio artifacts.

Izotope's dithering algorithm is available in both Ozone 4 and Soundforge 10 btw.


Russ - you basically have it right in the processes you described earlier. Katz really goes into it in detail.

http://www.digido.com/mastering-audio-book.html

This gives you some idea of the content of the book, which is recognized as top notch by industry pros such as Bernie Grundman, Doug Sax, Roger Nichols etc.

BTW - Shop around - I got it on sale from Amazon for $22.99 instead of the $39.99 quoted.

[Tone Deft]

Dave - yeah, thanks, good weekends make for awful Monday mornings. maybe max4Live can help us out, [phasor~] is used as a control signal, surely it can go well beyond 20kHz. I'd be curious to see if Live would record an ultrasonic signal. if we can get a signal like that into Live it could be interesting to see what doesn't pass the signal out. ultimately a google search for "40kHz wav file" might work, a quick google only showed the standard ones.

This gives you some idea of the content of the book, which is recognized as top notch by industry pros such as Bernie Grundman, Doug Sax, Roger Nichols etc.

BTW - Shop around - I got it on sale from Amazon for $22.99 instead of the $39.99 quoted.

so what does Katz say about sample rates? 50kHz is ideal, so 48kHz is the best option to use.

one of his citations makes some interesting conclusions:
http://www.lavryengineering.com/documen ... Theory.pdf

While this article offers a general explanation of sampling, the author's motivation is to help dispel the wide spread misconceptions regarding sampling of audio at a rate of 192KHz. This misconception, propagated by industry salesmen, is built on false premises, contrary to the fundamental theories that made digital communication and processing possible.

The great value offered by Nyquist's theorem is the realization that we have ALL the information with 100% of the detail, and no distortions, without the burden of "extra fast" sampling.

While there is no up side to operation at excessive speeds, there are further disadvantages:
1. The increased speed causes larger amount of data (impacting data storage and data transmission speed requirements).
2. Operating at 192KHz causes a very significant increase in the required processing power, resulting in very costly gear and/or further compromise in audio quality.

this thread also seems to be a lesson in rendering with Live. I never have sat down to practice rendering to see what works best.

[leedsquietman]

Katz explains the whole sample rate process, aliasing and other factors which can affect quality, the importance of filtering on DACS and ADCS, how a great ADC/DAC can sound better at 44.1 compared to a consumer grade one at high sample rates, the importance of finding a good sample rate converter and names various positives and negative scenarios for working at 44.1/48 or 88.2/96 and he also mentions the Lavry paper you quote.

It's Chapter 20 in the book and VERY THOROUGH. He basically approves of high sample rates (88.2 or 96) so long as the converters, sample rate conversion and dithering (if reducing bit depth to 16 bit etc) are sufficient quality.

Bob Katz actually mastered the world's first 24/96 DVD-A recording incidentally and with a team of top engineers was one of the first to build a high quality DAC capable of running 96 Kbps back in the late 1980s.

[Tone Deft]

6 pages of a wide layout book isn't THAT much detail, he hits the highlights with good technical detail. this is a HUGE subject.

'sufficient quality' meaning specialized gear, custom DACs. this is beyond the scope of gear we use.

page 253 - "how good is 44.1k? a lot better than I used to think..." "can 44.1k ever sound equal to 96kHz? it may be impossible to find out without building a custom, discrete DAC."

Lavry states:

The above suggests that 88.2 or 96KHz would be overkill. In fact all the objections regarding audio sampling at 44.1KHz, (including the arguments relating to pre ringing of an FIR filter) are long gone by increasing sampling to about 60KHz.

on page 253 Katz states:

Jim (Johnston) has experimentally calculated that the minimum Nyquist filter gentle enough to elude the ear is 50kHz, so if he is right, then the 48kHz professional rate is nearly sufficient."

I read all this as
192k - overkill to the point of harm.
441.k - not so bad.
48k - damn near the perfect 50kHz rate as determined by experiments.
88.2k/96k - includes the little extra goodness that 48kHz misses by not being 50kHz but it's a very very small improvement.



this is the crossroads. there's knowing where the flaws are and then there's knowing whether those flaws matter or not and understanding where the real weak points are in your signal chain.

IMO sample rate is much more important when you consider the entire signal chain, use what makes sense. use whatever requires the fewest SRC conversions, use what gives you the latency you can work with, common sense stuff. chasing down minute errors with sample rates is just silly, especially since there are a hundred other places you can improve your signal chain.

not to mention that just about every CD you've ever listened to has been a lowly 16 bit 44.1kHz recording and those can sound amazing!





tw1nstates - you never did answer what monitors you use.

[leedsquietman]

Good info Tone. I don't have my copy to hand, so I was going from memory.

If you have a poor signal chain in the first instance 96 Kbps won't be able to polish a turd.

You have poor preamps, poor converters, lots of jitter/poor clocking, etc working 32 bit float at 192 Khz won't save your bacon ....

he does also mention about aliasing in that chapter which some people see as one advantage (anti-aliasing of plugins etc) and how upsampling works.


It's still a try it out for yourself deal as far as I'm concerned.

[Tone Deft]

Good info Tone. I don't have my copy to hand, so I was going from memory.

you don't drink nearly enough.

"It's still a try it out for yourself deal as far as I'm concerned."

absolutely, that's part of the reason I'm beating this dead horse, let's work this out. 1ino1eum's wav files intrigue me, it would be great to break Live, it would also be great to learn a new trick.

[UKRuss]

This is great. Intrigued to see the outcome of what we shall now call '1ino1eum's paradox'

However, I thought that it was generally considered to be of little value to increase the sample rate if you could not reduce it by an exact multiple when going to your final product, so for CD at 16bit 44.1, there is little value in recording at 48, it would need to be 88.2 or you may as well not bother.

A touch of googling shows other forum polls with the most popular set up by far as being 24bit at 44.1 recording becasue it provides the best compromise between quality recording and keeping you computer running without issue. taking it to 24/48 caused lartger projects to break down relatively quickly. So, forget 48, stay at 44.1 but use the deeper bit depth and then simply dither to 16bit with a decent algo.

Thats not to say the theory isnt sound, but it has to be married with the practicality of your own set up.

I note, fwiw, that I start to have issues much earlier in my projects, freeaing required etc., working at 24 bit than i did at 16bit, but the quality in the end produce is pretty plain so it is worth it.

I am scpetical i would notice the same difference working at 48 instead of 44.1 sample rate.

But as i say, interested in the results.

lets solve '1ino1eums Paradox!'