KVR Audio

stefancrs · Post by **stefancrs** » Thu Feb 17, 2005 1:38 pm

With regard to the "SAMPLITUDE RULES" thread;
FWIW we are conducting a test
I have generated 16 stereo wavs of white noise at 44.100 samples / second in 32bit float.

We will perform two different mixdowns:
1) Just all the noise wavs playing simultaenously with channels set to -6dB. The master at 0dB though.

2) Same as above, but also utilising two send busses. All channels should send themselves to both bus 1 and bus 2 at -6dB.

Mixdown should be created in 32bit float. No normalizing, it should not be needed, and if it is for some strange reason, it will be done during the "nulling test".

Anyone care to host the files? I can do the test in Podium, popsych will do the test in samplitude. It would be good if we could put up a page with the source files needed, this instruction, and the resulting files and comparisons.

To refer people to in the future I mean.
If anyone's up to the hosting tasks, it's a 1Mb rar for the source files.

Btw, anyone can feel free to do the test in hosts already tested, if the results differ, we know someone made a mistake

edit: if anyone thinks there's something wrong with this method of comparing different hosts audio engines, do tell.

dramsenik · Post by **dramsenik** » Thu Feb 17, 2005 2:00 pm

you can download the test files here
http://www.oskarlegroupe.ch/dramsenik/audiotest.zip

stefancrs, i changed the *.rar in *.zip

hem i run through duplicate post...

Sascha Franck · Post by **Sascha Franck** » Thu Feb 17, 2005 10:28 pm

Logic 5.5.1 won't import or export 32bit float files.

Thumbs up for anybody else using antique sequencers!

C00kie · Post by **C00kie** » Tue Feb 22, 2005 2:33 pm

How's it going with this test, Stephan?

Is there a list of hosts that have either failed or passed the test?

dramsenik · Post by **dramsenik** » Tue Feb 22, 2005 2:39 pm

So far this test doesn't seems to grip...

Stefan can do this test in podium.

Personnaly i can't do it: 'cause my host is PT and i can't export in 32bit.

C00kie · Post by **C00kie** » Tue Feb 22, 2005 2:49 pm

I have another idea to test rendering capabilities of any host. Quite easy to perform and totally objective:

* Download the RightMark Audio Analyser
* Make exports in RMAA of its test tones at the frequency & bit depth your host can cope with
* Put only the test tones on a track in your host.
* Set all faders to unity (or have some fun and set the track to +2dB and master to -2dB)
* Render an output wav in your host
* Let RMAA analyse the file rendered by your host.

Please report any results: both hosts which analysis does and does not result in a nice flat line.

lex_strooder · Post by **lex_strooder** » Tue Feb 22, 2005 3:16 pm

stefancrs wrote: edit: if anyone thinks there's something wrong with this method of comparing different hosts audio engines, do tell.

Sorry to sound harsh (pun indeed), but will white noise really represent a normal audio file used in a host ?
Why not grab some samples of voice, drums, bass and a synth...

And why not include some freeware Vst like a reverb, a phaser and a limiter.

otherwise I would like to know why hosts sound different (and they really do in my ears) so please make this a good test.

/ Lex

stefancrs · Post by **stefancrs** » Tue Feb 22, 2005 5:19 pm

C00kie wrote:How's it going with this test, Stephan?

Is there a list of hosts that have either failed or passed the test?

Oh, heck, I'm a lazy bitch. I've done the tests for podium and I've got the test files for samplitude. I will try to get them online with the help of dramsenik tomorrow.

stefancrs · Post by **stefancrs** » Tue Feb 22, 2005 5:22 pm

lex_strooder wrote:
stefancrs wrote: edit: if anyone thinks there's something wrong with this method of comparing different hosts audio engines, do tell.
Sorry to sound harsh (pun indeed), but will white noise really represent a normal audio file used in a host ?
Why not grab some samples of voice, drums, bass and a synth...

And why not include some freeware Vst like a reverb, a phaser and a limiter.

otherwise I would like to know why hosts sound different (and they really do in my ears) so please make this a good test.

/ Lex

White noise is the meanest signal to put through any process that is supposed to _not_ alter the sound.
The difference in sound is probably due to level differences, panning law differences, and effect differences.

We might do a test with effects as well. Reverb won't work though, unless it's a pure convoluting reverb. Most reverbs has a degree of randomness in them, and if so, they aren't predictable so they can't be used in a comparison test.

BertKoor · Post by **BertKoor** » Thu Mar 10, 2005 11:05 pm

Downloading this test, will report later... Difference or no difference, I want to see some proof!

Has anybody seen those MythBusters on Discovery Channel?

Scotrick · Post by **Scotrick** » Thu Mar 10, 2005 11:43 pm

bert, buddy, I admire your drive...
for your reading interest, check out the forums at this place:

http://www.3daudioinc.com/

where many minds much smarter than mine have debated all of this esoteric world..

I came from a four track tape cassette recorder, and now I record with a blue mic and tl audio preamp into a motu 828mkii into cubase... I monitor on event monitors, tr8xl, and to my ears what I put into the machine is what comes out, whether I'm recording my own materials or others...

I used to be an 'audiophile,' with special speakers and cables and amps and on and on and finally I realized I was wasting my life chasing some really really minute differences in the audio chain when I should be sitting back either making music or enjoying it

C00kie · Post by **C00kie** » Sat Mar 12, 2005 4:01 pm

Yep, hosts differences are regarded as futile. Still the debate surfaces time after time...

Jafo · Post by **Jafo** » Sat Mar 12, 2005 4:18 pm

My spies inform me that C00kie wrote:Yep, hosts differences are regarded as futile. Still the debate surfaces time after time...

Howdy. No offense intended, but this isn't a debate. This is an experiment. They're testing a myth here on KVR, that all hosts sound the same. Will different hosts, when given the same data and transformations, give the same results? If they give different results, myth busted. If not, myth confirmed. Simple. No room for debate, period.

I suppose that if there are small enough differences, we can argue about what constitutes "negligible difference" -- and this will probably happen...

FWIW, I find clear, if subtle, differences between Tracktion, SONAR, Multitrack Studio, and Audacity. But this should provide definitive answers!

C00kie · Post by **C00kie** » Sat Mar 12, 2005 4:41 pm

I'm interested to see outcomes of this experiment, but have seen nothing at all yet. Anybody done the test, and willing to publish outcomes?

stefancrs · Post by **stefancrs** » Sat Mar 12, 2005 5:08 pm

I have but I still am a lazy-ass. I got the files at work. Podium renderings. Difference when compared to samplitude was at -150dB or something. You could see the floats running out of precision (taking giant steps instead of being able to change their values smoothly), so unless two hosts render everything with exactly the same volume (master volume or whatever) you will get the same error. You get the same error when you render from just one host but two times with slightly different master volumes. So... But I'll try to get it online soon

Host comparison test