Boosting sample rate of loops - is this possible?

[Dags]

Hi everyone.

Before I get into it, a couple of preliminaries:
* I'm not asking if there's any perceived benefit in working at a higher sample rate, this is just a question relating to how to get around a theoretical technical problem.
* I'm not trying to extend the Nyquist frequencies of something that has already been recorded at 44.1kHz
* I have tried to do a search on the forum for a similar question, but was unable to turn up anything (maybe I just asked the wrong question)
* This is just something I was playing around with last night (briefly and unsuccessfully) and it got me wondering if anyone on the forum has worked out how to do it.

Ok, let's begin

Let's say that I wanted to record some new audio at 48 or 88.2kHz in order to maintain whatever transient upper harmonic frequencies are inherent in the source.
So I set my software's sample rate to 88.2 (because dithering back to 44.1 later on will be in multiples of 2) in order to record this new audio.
(In case you're wondering, I'm using Logic V7 with a Delta 1010 card)

Hello! What's this? Now all my pre-recorded loops or audio I would like to work with are now playing back faster because Logic assumes that all audio in this session is recorded at an 88.2kHz sample rate.
Hmmm.......interestink.

I go into the audio window for the files already in the track and set them to be 44.1.....apparently this also resets the sample rate for the whole session again. Bummer.
Bouncing to disk only permits 16/24 (and mp3) bit rate adjustment as well.
Hmmm......a bit of a quandry.

So, finally getting to the point, has anyone discovered a way of converting 44.1kHz files to a higher sample rate in order to be able to use them in a higher sample rate recording session?

Your thoughts, experiences, and solutions are all welcomed!

Dags

 

[Vullkunnraven]

I've encountered something similiar yet quite different:

I use Pro Tools and Reason. Reason's loops, soft synths, etc are all sampled at 44, however I record in Pro Tools at 96. To keep the whole session at 96 I rewire the Reason sounds, bus their outputs, and record them on a new track in pro tools...

Would rewire be possible with yur software?

If not... you can try reamping (though this takes some good d/a converters and micing)... some people record the loops in their session and reamp them to give it that 'ambient room' sound

Hope some of this helps at least...

 

[Dags]

To keep the whole session at 96 I rewire the Reason sounds, bus their outputs, and record them on a new track in pro tools...

Would rewire be possible with yur software?

Just did a quick search and it seems that Rewire is possible in Mac OSX with Logic.
I don't have Reason. but you actually just gave me a thought somewhat along the same lines - if I were to load up the loops inside Logic's soft sampler, maybe it would permit me to re-record them out of the sampler using the session's higher sample rate?
Basically just capturing the audio from the 'sampled' instrument (in this case a drum loop) at a higher sample rate.
Or would the sampler's 44.1 dependency screw things up?
Will have to try this out on the weekend when I'm back in the studio.

If not... you can try reamping (though this takes some good d/a converters and micing)... some people record the loops in their session and reamp them to give it that 'ambient room' sound

So, playing the loops out of the monitors and recording them back into the software? Not a bad idea, and yes it would require some really accurate micing.
This is a possibility if I can get hold of another computer with a decent soundcard to play back the samples.
I still have the problem of only being able to have one global sample rate within a recording session. I wouldn't be able to play back the 44.1 loops while recording in 48/88.2 using my normal software.

Hope some of this helps at least...

Certainly has!
You have started me thinking about trying out loading the loops into the sampler - I probably wouldn't have thought of that if you hadn't mentioned Reason.

Its a bit of a workaround to what I was attempting to achieve/see what was possible within the realms of current audio processing, but if it works, who cares?
Thanks Vullkunnraven

Dags

 

[Ford Van]

So I set my software's sample rate to 88.2 (because dithering back to 44.1 later on will be in multiples of 2) in order to record this new audio.

First, you don't "dither" back to 44.1, you do "sample rate conversion" to 44.1 for another sample rate.

Second, how do you think a sample rate conversion works from 88.2 to 44.1? Do you think it will just drop every other sample and call it good?

It is NO easier to sample rate convert from 48, 88.2, or 96. All of the audio is "resampled" anyway.

Personally, if I was 88.2 capable, I would just do 96KHz. Why not. But, I suppose you could save 5% on hard drive space by going 88.2.

But, be advised, and don't be confused, doing a SRC from 88.2 isn't any better than doing a SRC from 96 except for the possibility of how "good" the original 88.2 or 96 files sound compared to each other (which I HIGHLY doubt ANYBODY could hear a difference between those two sample rates).

If you don't believe me, make three recordings of a CD playing back into your A/D converters, one at 44.1, one at 88.2 and one at 96. Edit to make sure each track is the same exact material and length. Then, SRC them both to 44.1. Now, have somebody play back all three files for you in a way where you don't know which file is currently being played. Have this friend switch between them randomly, and see if you can hear ANY difference.

That is kind of lame that Logic won't SRC on the fly! Sonar will work with any mix of sample rate and bit depth files within the same project.

 

[Ford Van]

And before any morons come along trying to argue about how 88.1 to 44.1 conversions are somehow "easier" or "better" than 96 to 44.1, please go read TECHNICAL papers about sample rate conversion and how it is done. Just because somebody on a BBS said that 88.2 to 44.1 is "easier" or "better" do not make that true.

ALL audio that is SRC'ed is up-sampled anyway then recalulated to the target SR.

 

[Dags]

First, you don't "dither" back to 44.1, you do "sample rate conversion" to 44.1 for another sample rate.

Oops - should have checked my terminology. I knew what I meant

It is NO easier to sample rate convert from 48, 88.2, or 96. All of the audio is "resampled" anyway.

That is kind of lame that Logic won't SRC on the fly! Sonar will work with any mix of sample rate and bit depth files within the same project.

That's pretty much why I am asking the question. I would have thought that Logic would be able to handle any sample rate on the fly as well!
Actually, Ford Van, searching on your term 'sample rate conversion' I have found a brief comment that Logic's sample rate conversion tool is within the sample editor window. Not *that* obvious at first, but now that I know where to find it it does make sense.
Its all about asking the right question, isn't it?
Silly me.
I'll give that a whirl when I'm back in the studio.

But yes, it is a bit of a bummer that the software doesn't just automatically compensate for various sample rates. Maybe we'll see this in version 8?

Thanks for your replies everyone. They have given me much food for thought!
(speaking of which, its lunch time!)

Dags

 

[snow lizard]

And before any morons come along trying to argue about how 88.1 to 44.1 conversions are somehow "easier" or "better" than 96 to 44.1, please go read TECHNICAL papers about sample rate conversion and how it is done. Just because somebody on a BBS said that 88.2 to 44.1 is "easier" or "better" do not make that true.

ALL audio that is SRC'ed is up-sampled anyway then recalulated to the target SR.

Not really. You could line up 3 SRC converters that all approach the task differently.

I think Dan Lavry actually has a unit that can be adjusted to a mode where it basically drops every other sample. It's supposed to work well. Weiss converters work a bit differently, but will generally upsample to a common multiple in all conversion cases. Other SRC's might basically do a D/A conversion and resample, which could introduce even more problems, or at the least negate any percieved benefits of going higher in the first place. Cheap converters will chump you no matter what.

There are other issues to consider.

(Is "chump" really a verb?)

Given that most people aren't using mastering grade converters anyway, it doesn't make sense to use anything other than the target rate.

That doesn't apply to bit depth I guess.


Moron,

sl

 

[bennychico11]

But yes, it is a bit of a bummer that the software doesn't just automatically compensate for various sample rates. Maybe we'll see this in version 8?

it's not that it can't compensate. Most all programs will convert the files you import to the original sample rate your session was created at. You should just create your session at 48kHz and then there probably is a File-->Import Audio command that will import audio and convert them to your session's rate. I haven't used Logic for a long time, so I can't give you specifics.

You just can't have one audio file at 44.1 and another at 48 in the same session. It's like telling the computer "Okay, I want the length of 1 second to be exactly 44,100 samples....no wait, I want it to be 48,000 samples....no wait, play it at 44,100!!!" If you had both of them in there at the same time playing at different sample rates, nothing would sync up properly. You need to decide on what sample rate you are going to work at and record you audio at that and/or convert all your other audio to it as well.

 

[Dags]

it's not that it can't compensate. Most all programs will convert the files you import to the original sample rate your session was created at. You should just create your session at 48kHz and then there probably is a File-->Import Audio command that will import audio and convert them to your session's rate. I haven't used Logic for a long time, so I can't give you specifics.

Ahhh......Ok I'll have a look for that. I just tried setting the global higher sample rate for the song then adding an audio file which, of course, played back at the wrong speed as a result.

You just can't have one audio file at 44.1 and another at 48 in the same session. It's like telling the computer "Okay, I want the length of 1 second to be exactly 44,100 samples....no wait, I want it to be 48,000 samples....no wait, play it at 44,100!!!"

LOL! And here I was thinking that computers were clever

Thanks for the useful info, bennychico11

Dags

 

[Ford Van]

Sonar 5 just does a SRC while it imports the file. You don't even have to ask it to do that.