Do DAW's really sound the same?

Yeah - I'm gonna try it a little later...Should we all pick a common track? Stereo? Mono? If there really is a difference, it shouldn't matter, right? Or - I think you mentioned it was a stereo track - maybe we stick to that.

What was the difference of the two signals? Some DAW's will apply different dither when rendering.

I think one should check pan law settings first before the subtlety that different dither might bring. It's in the Project Settings/Setup (? -in the edit menu) of Cubase...I don't remember where it is in Logic...

Bouncing files while changing nothing, though, isn't exactly the same as whether they sound the same on the mixer. I can't get Cubase to render a mono track null-able. It's always changing the gain.

Years ago, they sounded completely different to me. Nuendo 1 and DP I forget 3 or 4...sounded best. But, I also made an ASSUMPTION then that I no longer make--which is that the host would properly compensate for the third party plug ins I used to "make the test fair".

Fast forward, I did the same test a couple years back with Reaper, Cubase, and PT10. I didn't go to the "do they null" level of testing...but, none sounded any different to me WHILE PROPERLY FUNCTIONING. Which has a caveat...PT latency compensation was garbage. I was constantly engaging and disengaging plug ins for things to recalculate...it was WORSE when I hooked up hardware inserts and auxes. They literally MOVED while I was mixing. Why is it getting mushy? Oh, let inactivate and reactivate the plug ins? Cool-fixed. Reaper's hardware insert didn't work at all...I've had enough bad experiences with Logic on OSX and Sonar on Windows, they didn't get considered, but...It made my decision easy. But, it wasn't because any of them sounded different. I was even testing the 64bit version of Reaper in hopes it would get closer to the 56bit FIXED headroom I had...nope-sounded just like PT10 and Cubase6 at 32bit.

I think the whole importing a file and rendering it out is a completely VALID test of rendering (which apparently as of 6.07 update, my cubase is having trouble with)...but, it's not terribly related to the "sound of the DAW"--which would be more a combo of mixer and pan law implementation...and plug ins and IO compensation. Any time it's slipping and sliding track in time (via poor compensation) is "sounds worse" even when you can't hear it as X track is behind.

Wouldn't the exact same word clock have subtle diffs in timing every conversion? I would think so, so naturally they should NOT null, and in my experience, not all daws are created equal, I base this statement on personal shootouts, and my ears alone.

Former Longtime Protools user. Been using Harrison Mixbus.
I've been using my Harrison console since 1981, so this was a natural choice.
Mixbus sounds great. More analog-like, things mix together easier,
and the mixer is laid out like a real console.
I hear a difference for the better.

Through a 32bit mixer. The truncation still happens....dither noise application or not. One would assume it should be able to bounce a 24bit file to another 24bit file without altering it, but that would be an assumption based on common sense.

....which is also I think the issue with Cubase's export--it works PERFECTLY from exporting audio mixdown....which actually IS the intended point--but that's also how they intend for BWAV exports to happen...and it's a PIA to do...and for mono tracks always results in a gain change for me. For stereo, it doesn't--I render from channel outs all the time full stereo mixdown files and it's fine...for a while I set up a mono group buss and assigned whatever mono I wanted to export to it, because that's what was recommended. It worked but what a stupid PIA. I wish they'd just make their "track export" have an option to just render as a contiguous timestamped WAV...rather than the silly combo of a not timestamped wav (or wavs) and a proprietary XML file with the project placement and Cubase "info".

So the general consensus is all daws are NOT created equal? My head is spinning trying to keep up with Popman but he knows a lot more about it than I do. This is not a deal breaker for me nor never will be. Once something goes outside the computer into the analog realm, all this means nothing. Conversion doesn't matter that much either when mixing outside. But it is nice to know the answer to this age old digital question. The null test proved it for me.

Created equal or sound the same? Kinda different because a floating point software mixer will, IME, sound the same when everything is working properly...which becomes the caveat, right? The "less equality" is in how consistently and in how many different IO and plug in situations it all works properly.

I left PT7.4 for Repear for two reasons: sound and the whole proprietary hardware bullcrap, back then. I'm now in studio one and it is touted to have a better sound engine than other DAW's. So even if they are wrong, there is a recognition of differences.

BTW, here is a cool Studio One Trick - a plugin provides insert access to outboard on a track or buss and can be toggled on or off to A/b, instantly. That's pretty slick. Select three tracks, right click and create a buss with all the sends in place. Drag an effect to an insert and a buss with sends is created instantly. I think I'm going to be using this a while.

This was going to be a part of my reply. Not only aliasing noise and distortion, but also AA filters being made differently. Looking even just at the impulse response difference between Logic and Cubase you see that Logic has longer pre and post ringing which means they'll never match up in the time domain. Logic uses more aggressive filtering, which shows when you look at the sweep - the distortion is reduced in comparison to cubase - meaning you don't have identical harmonic content either. Now by not identical I mean within the 20hz-20Khz rang it's down -100dB or so. It'll never be audible, but it's another reason they wont null.

Unless a DAW touts Non-Linear Summing you can expect the mechanics of the DAW's to be trivial in relation to sonics. The way a wav is generated by summing signals is Sample 1= (Channel 1's first sample +....Channel N's first sample)/ Master Fader Gain. The clock can't affect it because it's not a real time thing, and even with an Online Render you're looking at clock drift in the nano-seconds with a 2.5Ghz processor (coupled to a crystal clock with +/0.5% accuracy). Drivers should obviously be working in this decade too - though I've heard a Digi 192 just freak out before..

I did a test like this with Logic, PT and Ableton Live. Ableton Live has a fixed pan law, so I set Logic and PT's pan law to match abletons.

My test consisted of taking a sinewave generated from a stand-alone app (i was reading Learning CoreAudio and their first tutorial is a commandline app that generates a sinewave), importing it into each daw and bouncing it down to a stereo file.
Then, in each daw, I imported all three files and tried nulling them against each other. PT added a little bit of silence at the start of the file compared to logic and ableton. Once i compensated for that, the files from each daw nulled out for each pair of files, and nulled out in each daw. 1 + -1 = 0.

I just realized, I'm 49 posts away from redemption!