Live 6 versus SX 3.1 sound quality test

Robert Henke · Post by **Robert Henke** » Sun Aug 27, 2006 7:41 pm

Hi Tarekith,

you must have made some systematic error. It is easliy possible to cancel out your drums but the guitar is still present. This is very odd and inidcates a flaw in your test.
Latency compensation ? Plug Ins with Latency ? Sends open ?

I would not say I it is impossible that details are different in Cubase as in regards to Live butyou would certainly not get an okay sounding guitar track but no drums anymore if you subrtract both files.

The differences which could occur due to rounding errors and such would sum up to something very low in level and very unpleasant to listen to if amplified enough.

Robert

Tarekith · Post by **Tarekith** » Sun Aug 27, 2006 8:05 pm

I'm not sure Robert, the results are indeed odd. I tried to rule out as many variables such as those you mention, but it is indeed possible I missed something.

I need to come up with a test that either doesn't involve any plug ins, or has just VST plug ins. Then I can load the same presets into each app, and just the wav files in common.

Robert Henke · Post by **Robert Henke** » Sun Aug 27, 2006 8:20 pm

Oh, some detail worth knowing:

most controls in Live work with unquantised floating point precission internally.
This means you can type in "-0.0003" in a volume slder and it will do this.
But: you cannot see it because the display will tell you " 0.0 dB".

Most of the time the potential discrepancy between the displayed value and the real one does not matter, but for your test they do. So, type in the exact value you want in each volume slider instead of simply dragging it.

For the future we plan to change it, so that the visible resolution of all paramerters always equals the exact internal state. But this is quite a task because we have to ensure old Live sets will still sound the same.

Robert

Tarekith · Post by **Tarekith** » Sun Aug 27, 2006 8:23 pm

Thanks Robert, it seems after looking at the files, that the sounds that are always audible in the polarity reverse file, are those that are using Live plug ins, and not VST plug ins. This would explain it.

RePeter · Post by **RePeter** » Sun Aug 27, 2006 8:26 pm

Robert Henke wrote:
most controls in Live work with unquantised floating point precission internally.
This means you can type in "-0.0003" in a volume slder and it will do this.
But: you cannot see it because the display will tell you " 0.0 dB".

Robert

i love it when i come on this forum and actually learn something!

raapie · Post by **raapie** » Sun Aug 27, 2006 8:35 pm

I think not using VST plugins would be the best way to do another test. I have done several tests between Nuendo and Live5 and didn't find a difference when using fixed bits. VST plugins don't down sample as far as I know, I have heard many artifacts because of this.

Tarekith · Post by **Tarekith** » Sun Aug 27, 2006 8:45 pm

Ok, I modified the test. This time, I rendered all tracks in my loop as seperate audio files. That way I'd be working with the exact same files in both programs, and no plug ins would be involved. Then, I dragged these audio files into both a new Live Session, and new Cubase Project. All settings were equal and the defaults. Rendered in both apps into 24/44.2 mixdowns.

The mixdown were then compared in SX with one tracks polarity reversed again. This time the result was total silence. Nothign showing on inspector either, they cancelled completely. Tried my rewire experiment too, same results, total cancellation when compared to either the SX or Live render. I won't post the files unless someone really wants to hear them

No need as there's nothing to hear.

Ok, so any sound differences aren't down to the mix bus, so it must just be the fact you need to use warping in Live to work, or maybe Live's plug ins? Hmmm, interesting again....

The Phat Conductor · Post by **The Phat Conductor** » Mon Oct 02, 2006 9:40 pm

any updates here?

leisuremuffin · Post by **leisuremuffin** » Mon Oct 02, 2006 9:44 pm

timely bounce, phat conductor!

.lm.

Tarekith · Post by **Tarekith** » Mon Oct 02, 2006 9:48 pm

Not sure what else I can add. As long as you don't use warping, the mix engines are identical. Obviously we all know warping audio has at least a small effect on the sound quality, and of course better plug ins will sound...well... better

I guess the only thing I can say is that now that I know it's not a mix engine issue, I have no issues with using Live for all my production work. Well, at least from a sound quality standpoint, and I tend to take that stuff more seriously than perhaps I should.

There's still some features missing that would making mixing easier, but those aren't holding me back from dumping SX and going to Live for all my DAW needs.

If there's something else you'd like me to test, let me know though!

mercyplease · Post by **mercyplease** » Mon Oct 02, 2006 10:13 pm

Tarekith wrote:Ok, I modified the test. This time, I rendered all tracks in my loop as seperate audio files. That way I'd be working with the exact same files in both programs, and no plug ins would be involved. Then, I dragged these audio files into both a new Live Session, and new Cubase Project. All settings were equal and the defaults. Rendered in both apps into 24/44.2 mixdowns.

The mixdown were then compared in SX with one tracks polarity reversed again. This time the result was total silence. Nothign showing on inspector either, they cancelled completely. Tried my rewire experiment too, same results, total cancellation when compared to either the SX or Live render. I won't post the files unless someone really wants to hear them No need as there's nothing to hear.

Ok, so any sound differences aren't down to the mix bus, so it must just be the fact you need to use warping in Live to work, or maybe Live's plug ins? Hmmm, interesting again....

This was a great thread that actually had a results. So many of these buss tests go on and on without any conclusion. Its cool that you did your original test because it forced you to do a better one that gave you the expected results and assured any one worried about the live mix engine theres nothing to worry about.

Tarekith · Post by **Tarekith** » Mon Oct 02, 2006 10:28 pm

Thanks. I had a personal interest in this one, as I was seriously considering switching from Cubase after years and years of using it.

Jajah · Post by **Jajah** » Tue Oct 03, 2006 1:00 am

RePeter wrote:
Robert Henke wrote:
most controls in Live work with unquantised floating point precission internally.
This means you can type in "-0.0003" in a volume slder and it will do this.
But: you cannot see it because the display will tell you " 0.0 dB".

Robert

i love it when i come on this forum and actually learn something!

Just trust your ears:

difference · Post by **difference** » Tue Oct 03, 2006 10:19 pm

Tarekith wrote:Thanks Robert, it seems after looking at the files, that the sounds that are always audible in the polarity reverse file, are those that are using Live plug ins, and not VST plug ins. This would explain it.

Hang on a sec....

You were trying to do a comparison with all the same settings etc but you used Live's own plug-ins on one of the files?

MrYellow · Post by **MrYellow** » Wed Oct 04, 2006 12:08 am

Your test is flawed.

silence is what you'd expect to hear.

No it's not. I expect to here the different in the artifacts produced by the
effect (warping) that was applied in both apps.

This is 100% expected and logical.

What you need to do is a simple mix, no warping, no effects just wav files
playing back at their original pitch and speed. THEN you will result in silence
no matter what DAW you use. As long as panlaw is accounted for.

-Ben