44 to 48kHz conversion

[dani]

88.2 -> 44.1 involves a low pass at 22khz to avoid aliasing, then discarding every other sample.

high quality x khz -> 44.1 involves up-sampling in the MHz range (interpolation - of which there are different methods), low passing, then discarding the unnecessary samples - obviously this is more complex, and is dependant on method of up-sampling as well.

and yes, this isn't the typical sneap forum discussion! aha.

thanks,

 

[MetalSir]

that said, i think it's important to note that, although the DIs are recorded at 44.1, there will still be more information in a 48khz recording of a reamp, as opposed to a 44.1khz reamp. most (if not all) of the 10khz+ signal from an amp is produced by the amp, not the guitar/DIs. basically, what MetalSir said.

this is EXACTLY what i'm talkin' about..

and yes, this isn't the typical sneap forum discussion! aha.

..and yeah.. ain't talkin' about "hey bro how TF can i get the cannibal_fuck*in'_corpse's snare sound?? because i love 'em and they kick ass!!" ..that is basically the reason why i collected just 105 posts since i signed in in late 2008.


BTW:
almost ALL here said: "hey man! you stupid retarded! do everything at 44100hz and spend more time with your girlfriend in stead of askin' so sptupid audiophile questions".

i can accept this.

 

[RedDog]

Correct about the law of halves, up sampling an audio file from, say 44 to 48 would not merit any quality increase. All the conversion is doing is adding roughly 180 (if math serves me) sample zeros. And then, by downsampling, the error would occur trying to truncate the added information down to a smaller rate. But unlike bit depth conversion, you're not changing the word length so you need not apply dither.

 

[MetalSir]

Correct about the law of halves, up sampling an audio file from, say 44 to 48 would not merit any quality increase. All the conversion is doing is adding roughly 180 (if math serves me) sample zeros. And then, by downsampling, the error would occur trying to truncate the added information down to a smaller rate. But unlike bit depth conversion, you're not changing the word length so you need not apply dither.

dudez PLEASE. i'm not saying ANYTHING about getting more definition by upsampling an audio file! Gesus i got a DEGREE in audio ENG in the most famous italian conservatory!!

i need to upsample because of for reamping at 48khz i HAVE TO BUILD UP a 48khz protools session! so, when i import the DI audio files (recorded at 44khz) into my 48khz session OBVIOUSLY protools will ask for CHANGING the SR. This is the ONLY reason why i need to upsample: importing 44100hz files into a 48000hz PT session.

ain't talking about getting magically more definition by creating on the thin air bits during the upsampling. NO WAY.

only one guy here understood what i'm talking about: a DI guitar/bass (line, not distortion at all) can be recorded with HUGE margin at 22khz without ANY affection for the sound content AT ALL. it's a theorem, not air.
but AFTER a reamp with distortion AS HELL you'll get just a BIT more detailed audio file.. with just a bit more spectrum width.. so.. please don't tell me that using 44100 SR is THE SAME of using a 48khz one. 'cmon, please.
and the funniest thing is that I'm not batman!

 

[John_C]

To answer the original question:

If it's using the same upsampling procedure then it's completely the same, obviously.
I don't see any reason why pro tools would do it differently on import or export. Try a null test and see if you're really that worried about it

 

[RedDog]

Ohhh, I got you. I couldn't quite grasp your english. Yeah, upsample the DI, reamp that. Boom, done.

 

[MetalSir]

Ohhh, I got you. I couldn't quite grasp your english. Yeah, upsample the DI, reamp that. Boom, done.

ohhhh! finally!!
now, let's go back to the OPENING QUESTION (and the only reason why i opened this 3d):

FOR YA GUYZ, IS BETTER:

1. CONVERTING FROM 44->48 EXPORTING THE FILES FROM THE ORIGINAL SES (AT 44KHZ) AND THEN USING THEM IN THE NEW SES (AT 48KHZ)
2. EXPORTING THE FILES AT THE SAME SR OF THE ORIGINAL SES (44KHZ) AND CONVERTING 'EM WHILE IMPORTING THE FILES INTO THE NEW SES (AT 48KHZ)

this was the only thing that i questioned.

 

[AEL89]

I already answered your question - File > Save Copy In then choose your desired sample rate and it will save the entire session at that. You'll still have the old session folder too so you can either import session data or import regions between each session, converting only what you need to.

Once again: There will likely be a negligible difference between converting the DIs up then reamping at 48 then converting down and doing everything at 44.1.

(You also misquoted me on the previous page by the way)

 

[Uros]

FWIW you should be doing the sample rate conversion in a dedicated SRC software, instead doing it in your DAW. Try Voxengo r8brain. It's free (there's also a professional version, but the free one will suffice).

 

[Pedro Teixeira]

What's the point in this? If you're upsampling a DI that was previously recorded at 44.1khz you're adding bits of zero'ed information, as in, nothing, nada. So even if you re-amp at 48khz the source of the guitar was originally captured at 44.1khz thus this being quite useless.

 

[TheWinterSnow]

ITT: people who all think they are right and actually don't know jack shit. I was going to help out in this thread, but the more I read through it, the more I realized it was full of fail and aids.

yeah pointless response for a pointless thread.

OP: you can't hear a damn difference between 44.1 and 48, and if you think you do, its a placebo effect.

 

[MetalSir]

ITT: people who all think they are right and actually don't know jack shit. I was going to help out in this thread, but the more I read through it, the more I realized it was full of fail and aids.

yeah pointless response for a pointless thread.

OP: you can't hear a damn difference between 44.1 and 48, and if you think you do, its a placebo effect.

your 2,263 post. basically you +1 on here.

 

[BLUElightCory]

ohhhh! finally!!
now, let's go back to the OPENING QUESTION (and the only reason why i opened this 3d):

FOR YA GUYZ, IS BETTER:

1. CONVERTING FROM 44->48 EXPORTING THE FILES FROM THE ORIGINAL SES (AT 44KHZ) AND THEN USING THEM IN THE NEW SES (AT 48KHZ)
2. EXPORTING THE FILES AT THE SAME SR OF THE ORIGINAL SES (44KHZ) AND CONVERTING 'EM WHILE IMPORTING THE FILES INTO THE NEW SES (AT 48KHZ)

this was the only thing that i questioned.

#1. Keep it simple.

 

[MetalSir]

#1. Keep it simple.

thanks man!

 

[Lasse Lammert]

you all should read this before ever posting about this topic again:

http://www.google.de/url?sa=t&rct=j...KQK5_0IvQ&sig2=lNEBeLBEY0GD9Q_XmYIi9A&cad=rja

 

[nervirasme]

I just wonder how many people would still respond to the subject of sample rate vs. audible differene if there was a blind test in there somewhere...

 

[TheWinterSnow]

you all should read this before ever posting about this topic again:

http://www.google.de/url?sa=t&rct=j...KQK5_0IvQ&sig2=lNEBeLBEY0GD9Q_XmYIi9A&cad=rja

Sources like this make me cringe. To simplify a lot of what that article says is that you only need two samples to provide an accurate representation of the signal without any distortion. Well two samples of a sine wave is now a square wave (a lot of distortion) and you get the sine wave back by running the signal through and anti-aliasing filter.

There are a few problems with that. One, the filters aren't that effective, two, the filters if operated the way they WHERE designed will convert what was supposed to be a square wave or near square waves into sine waves. Still, at 44.1KHz/48KHz, modulated ringing begins to take place at about 11.025KHz. This is why the golden number for the low pass filters on guitars is 12KHz. 96KHz has a lot less distortion aliasing but is still there however, virtually unnoticeable in a mix. 192KHz, has no audible aliasing. When it comes down to it, in the real world, 44.1 and 48 has distortion, you can cite as much theory as you want. don't believe me, run a sine sweep through a spectral analyzer through the various sample rates, you will see what I mean.

I also remember a thread awhile back that had spectral analysis of just about every DAW to show intermodulated distortion at 44.1KHz from their engines and some fared better than others. Summing in DAWs will add distortion to the signal as well on top of the already aliased signal. Anti-Aliasing filters aren't good enough to fix the image of 20KHz until we reach a sample rate of 192KHz from my studies and research. And by studies and research, I actually mean, building sample based digital platforms, anti-aliasing filters etc and actually testing them with scopes and analyzers. These circuits and technology are far from perfect by all means.

Dr Nyquists research and theories are just that, theories. The final real world result never works out as perfectly as it did on paper.

 

[MetalSir]

I just wonder how many people would still respond to the subject of sample rate vs. audible differene if there was a blind test in there somewhere...

ok. i surrender. after having explained WHAT and WHY i need 3 or 4 times i fucking surrender.

a question on a LOGISTIC matter (on import/export before or after resampling) turned on a CRUSADE against resampling, that i PERSONALLY don't give a shit about. i decided to reamp a 48. and that's it.
- i'm not asking to you if i am right or not.
- i'm not asking to you if this make or not sense.
- i'm not asking to you if i gain something or not doing this.
- i'm not asking to you anything about what you're fighting for.

my question was so ESASY that no one of you understood it. and this is ridiculous.

before searching on google FUNNY article you'd just READ the posts BEFORE the last one, and understanding what people ask. increasing your post score posting just for a +1 isn't the right way.

let me ask this somewhere else.

 

[TheWinterSnow]

What you want to hear != what you need to hear. But I will say this. Common sense will tell you to go the route that provides as minimal re sampling as possible. Since the quality gain between 44.1 and 48 is abysmal (48 was not used for quality gain, it was what broadcasting decided to use, audio went for a 44.1 format) it really doesn't matter which you are recording in. With that in mind, your best bet is then to resample your session to 44.1 and reamp and mix in 44.1 due to the fact that the final mixdown will be 44.1. This path gets you to your final destination with as little resampling of each individual track. Its constant resampling that will cause more audible distortion, so go the path that provides for less resampling and you are in the clear.

From now on try to remember this. if you are recording audio for TV (and Movies/Movie Scores) or Broadcast, record in 48, if you are recording audio for music production (CD's, radio etc) record in 44.1. The difference in sample rate has nothing to do with quality, it was rather the decided protocol for that application.

 

[Trevoire520]

FOR YA GUYZ, IS BETTER:

1. CONVERTING FROM 44->48 EXPORTING THE FILES FROM THE ORIGINAL SES (AT 44KHZ) AND THEN USING THEM IN THE NEW SES (AT 48KHZ)
2. EXPORTING THE FILES AT THE SAME SR OF THE ORIGINAL SES (44KHZ) AND CONVERTING 'EM WHILE IMPORTING THE FILES INTO THE NEW SES (AT 48KHZ)

this was the only thing that i questioned.

I don't think it's going to make a difference at all. Assuming you're using the same program to resample each time it doesn't really matter if it's resampling on export of the original session or resampling on import to the new session. The maths/processing is still the same either way.