SAMPLITUDE RULES (DAW Summing)

[zeoy]

Regarding my shorter renders, maybe the problem is that I checked the "only render marked region" setting and maybe the markers were a bit off.

If you feel up to it, could you put all the noise wavs a bit in into the arrangement (like half a second) and make a new render where you make sure atleast the entire noise*.wav gets played back? I can trim it to be precisely 22050 and aligned with the other mixdowns afterwards. This way, if someone would feel like hosting and writing a small conclusive text or something, could put the mixdowns from all the hosts tested online, and they'd all be exactly 22050 samples long (half a second of noise), phase aligned etc. So anyone could easily replicate the test if they wanted to.

I could do that (and most probably will do it tomorrow) but it would be better and more fair if someone with access to the latest Tracktion version did the test.

[daGuru]

OT: Hey Zeoy, who did yer sig/banner? It's kinda groovy

[stefancrs]

I could do that (and most probably will do it tomorrow) but it would be better and more fair if someone with access to the latest Tracktion version did the test.

I want _both_ It'd be nice to be able to spot any difference if there is one etc.

[zeoy]

I could do that (and most probably will do it tomorrow) but it would be better and more fair if someone with access to the latest Tracktion version did the test.

I want _both_ It'd be nice to be able to spot any difference if there is one etc.

OK stefan. cant say no . But that'll be tomorrow.

[stefancrs]

That's ok, I'll be off now

[zeoy]

OT: Hey Zeoy, who did yer sig/banner? It's kinda groovy

It's a friend of mine. He's a very talented graphic designer (IMHO). Maybe he should be in the VST-GUI business

[daGuru]

For sure, tell him I need a GOMP monster plz

[jens]

Problem is, i'll bet you none of the Samplitude guys will post their results here.

(Nevermind me,just being a bit cynical today)

I will! (Magix Audio Studio has the exact same engine and I use that one regularly for editing tasks and such)

[HanafiH]

The .RAR file

HERE 3.9MB

…contains a number of 24bit 48kHZ standard .WAV format files.

There are nine MONO .wav files, named w55.wav, w110.wav and so on. Each file contains 0.25 seconds of silence followed by a TRIANGLE wave at the frequency in Hertz suggested by the file name, which rises from infinity to –0.03dbFS over a period of five seconds. These tones were generated using the Signal Generation utility in Steinberg Wavelab 4.0.

Each file can be loaded into a separate track in a DAW of your choice, under the following conditions, and the following test performed. Do not play all nine triangle wavs at full volume on your system, doing so may overload your monitoring and/or damage your system.

All nine test WAVs should be loaded at origin in a 48kHz 24bit project.

Pan Law MUST be set to 0db.
Pan pots MUST be set to dead center.
All dithering, processing and plugins MUST be removed from the audio chain.
MASTER faders should be set to 0dbFS (normal full on 0db position)
All TRACK faders must be set to –15dBFS.

All sound between origin (0.0 seconds) and origin + 5.250 seconds should be rendered to an interleaved stereo file at 24bits resolution and 48kHz sampling rate.

The .RAR file also contains two outputs from this test. One from Samplitude 7 (stests7), another from Cubase SX1 (stestsx1) (the only two hosts I have). Comparison using Steinberg Wavelab 4.0 indicates the following:

The Samplitude file has a maximum loudness of –2.50dbFS, the SX file -2.51dBFS.

The Samplitude file has a DC Offset averaging –121.235 dbFS

The SX file has a DC Offset averaging –121.754 dbFS

Audio file comparison by Wavelab indicates there are 239,280 points of difference between these two files, and the maximum amplitude of that difference is -70.17dBFS.

[stefancrs]

Which indicates that the difference between SX1 and Samplitude7 is bigger than the difference between any of the following:
Podium, EnergyXT, Tracktion 1.6 and Vegas.

Which, imho, in turn indicates that either you are doing something wrong, or Samplitude or SX1 is. Have you tested doing the same but rendering to 32bit float? If not, how do the two hosts do the truncation to 24bit? (I seriously doubt that's the source of the error though, since Vegas output was very close to the other hosts)

It'd be good if someone could do the tests with the same test data as the ones we've used so far. The only significant difference between the two tests is that in "my" test, we're using more tracks.

[jupiter8]

Just tried it in Cubase SX 3 and it canceled out completly against the Podium reference track.

[stefancrs]

Completely? Could you zoom in on the x axis until you start seeing noise? Might depend on your audio editor, but in most you can. If the x axis scale is in dB you can see how small the difference is after you've zoomed in a fair bit.

[nuisance sonore]

Or better yet, send it to Stefan so he can check it in the same editor he used for the others.

HanafiH, I tried the files you made available in Vegas against the Sam and SX you provided and I'm getting huge differences; around -84dB against the Sam output, and -60dB against the SX1 (and the difference with SX1 is all over the place). Not really in accord with the result we were getting so far, could you try the files Stefan proposed?

[stefancrs]

Or better yet, send it to Stefan so he can check it in the same editor he used for the others.

Adobe Audition 1.5.

[jupiter8]

Ok, i just mailed it to Stefan. When i said completley i meant i could'nt hear anything and i did the Cubase Audio/Statistics it reported 0 dB min/max sample value.
DC offset 0.00% minus infinity dB.
It did however report about -173 dB rms power both min and max.