96k vs 44.1k... big difference

[Revson]

For all the nerds in the audience:

http://www.hardwaresecrets.com/printpage/How-Analog-to-Digital-Converter-ADC-Works/317

 

[John_C]

For all the nerds in the audience:

http://www.hardwaresecrets.com/printpage/How-Analog-to-Digital-Converter-ADC-Works/317

thanks a lot. The guys at hardwaresecrets are legends.

 

[locus_coeruleus]

Well, all I hear is like the 96khz version was not lowpassed.

 

[s34nsm411]

here is another test of a synth

http://dl.dropbox.com/u/2093596/gladiator44.mp3
http://dl.dropbox.com/u/2093596/gladiator88.mp3

pretty huge change in the high end... sounds better to me at the higher sr

 

[Uros]

My ultimate question is - does recording recorded at 96/88.2khz and then downsampled to 44.1 khz sound superior to recording made straight at 44.1khz in practice?

 

[006]

Yeah my problem with this is that you took 44.1 and upsampled to 96. Not only does the audio need to go back to 44.1 anyway after that to be played on consumer shit, but also simply converting it to 96 is not the same as capturing something at 96, so this example is really pointless, IMO. To get the best idea of the differences you need to have two separate performances, one done in 44.1 and one in 96 - not the same 44.1 capture upsampled.

 

[s34nsm411]

Yeah my problem with this is that you took 44.1 and upsampled to 96. Not only does the audio need to go back to 44.1 anyway after that to be played on consumer shit, but also simply converting it to 96 is not the same as capturing something at 96, so this example is really pointless, IMO. To get the best idea of the differences you need to have two separate performances, one done in 44.1 and one in 96 - not the same 44.1 capture upsampled.

except that it already is back down to 44.1/16bit in the files I posted...

this test is to show the differences in vst behavior at different sample rates and NOT to compare sampling an analog source at different SR

either way if you listen to the test I did with a synth there is no captured audio, just midi, and the same effect is demonstrated

 

[006]

I am aware that it's back down to 44.1/16, which was my point. You went from 44.1 to 96 to 44.1. Still not the same as starting off at 96 and going down to 44.1.

 

[s34nsm411]

yeah not the same of course but thats just another variable in this little test. so far ive only determined that the processing is different in 96 (or 88) compared to 441

raw guitar is so lofi sounding anyway I doubt there would be much advantage to tracking DI's at 96 vs 44.1

 

[fatalforce]

good experiment sean. that is a difference i can hear! I will be trying this in the future to see if I can duplicate the results.

 

[JeffTD]

yeah not the same of course but thats just another variable in this little test. so far ive only determined that the processing is different in 96 (or 88) compared to 441

raw guitar is so lofi sounding anyway I doubt there would be much advantage to tracking DI's at 96 vs 44.1

I'd have to disagree, I'm guessing you'll hear a bigger difference than you think when tracking at 96.

 

[Uros]

http://www.lavryengineering.com/documents/Sampling_Theory.pdf

Good link on the subject.

 

[John_C]

http://www.lavryengineering.com/documents/Sampling_Theory.pdf

Good link on the subject.

 

[Soundlurker]

I did a test as well. The 44khz and upsampled 96khz files are of the same performance. The real 96khz guitars are a different performance.
Here we go:

44khz DI
Upsampled 96khz DI
96khz DI

----

44khz Rhythm Guitars
Upsampled 96khz Rhythm Guitars
96khz Rhythm Guitars



As you can hear, not much difference between the 44khz and upsampled 96khz guitars but a quite evident difference when compared to the real 96khz in both DI and distorted guitars. How much of that difference is due to the different performance it is hard to tell but it is clear for me that the 96khz DI sound more pleasant and the 96khz distorted guitars sound smoother.


Schecter with 81->E-mu 1616m->Poulin Hybrit->Guitarhack between -1
No processing whatsoever.


Edit: 192khz 192

 

[John_C]

there's a really dramatic difference there.....

 

[s34nsm411]

Shadow_Walker I was pretty amazed how similar all 3 sounded in your test so I had to try it again myself on the assumption that something you had in the chain does not react nearly as wildly to SR change than I did.


http://dl.dropbox.com/u/2093596/test/guitarhack96.mp3
http://dl.dropbox.com/u/2093596/test/guitarhack441.mp3
http://dl.dropbox.com/u/2093596/test/lextacy96normalized.mp3
http://dl.dropbox.com/u/2093596/test/lextacy441normalized.mp3

all files are using the same 44.1 DI you provided

signal chain for the first two is lextacy into guitarhack JJ between -1 impulse in LeCab

signal chain for the second two is lextacy into 'aradaz ourcabinet'


from this test we can see that the ones using the guitarhack impulse are not too different, but the ones with the other impulse are

conclusion: poulin sims sound pretty much the same at different SR, it is the mostly the impulse that sounds different at higher SR.

maybe the aradaz ones were recorded at higher SR? maybe the aradaz interface somehow upsamples the impulse wrong at higher SR? Not sure what, but something is clearly going on here.


EDIT: actually I used the lecab default impulse on mistake here! not the guitarhack

 

[Soundlurker]

Why do the Guitarhack ones sound so different from mine? I did use the older "between -1", but the JJ one doesn't sound that different usually.

 

[TimatWaves]

I know the threads focus is primarily on the fidelity difference, but there is something else to consider.

Going from 44.1/48k to 96k will require up to 4 to 5 times more CPU power for certain plugins. If you are running plugin heavy sessions, this can become a factor even on powerful machines.

 

[SimonTaddio_Qc]

^
good point

 

[s34nsm411]

Why do the Guitarhack ones sound so different from mine? I did use the older "between -1", but the JJ one doesn't sound that different usually.

hmm not sure, im pretty sure i used this one:

GuitarHack JJ BETWEEN-1.wav
80.7 KB (82,690 bytes)

and i used lextacy anyway, i think you said you used hybrit?