The sound of different DAWs

PaulBrewer

Has anyone done any tests on the 'sound' of DAWs ?

I guess I mean the sound of its summing as opposed to plugins or converters.

Anecdotal evidence suggest some sound better than others.

Anyone any thoughts ?

Neurojazz

PaulBrewer wrote: »

Has anyone done any tests on the 'sound' of DAWs ?

I guess I mean the sound of its summing as opposed to plugins or converters.

Anecdotal evidence suggest some sound better than others.

Anyone any thoughts ?

There was a few years back (and discussed on this forum)

All the outputs summed equally (so no tonal changes)

krd

Did you know. The guy who wrote Cool Edit (now owned by Adobe), wrote the sound engine for Win 95. He actually wrote the whole thing by himself and handed it over to Microsoft. And it was the same sound engine for years.

Cool edit was written later. And there was noticeable quality difference in anything done in Cool Edit, than anything that had to go through the Windows sound engine. I think he knew there was bug in it, and he got around it in Cool Edit.

Whatever platform you're using, you will always have to pass through the OS sound engine - could be a combination of OS and third party hardware. What set Apple apart for years, is they put so much work into getting their engine right. MS, were more focussed on doing office machines.

I don't know if there would be much difference nowdays. Hardware is a lot faster, and less corners need to be cut to get the stuff working.

Anima

No it doesn't apply today. With ASIO, the DAW or any audio software is talking directly with the hardware (or close to it) which is why there is better latency.

WASPI is what Windows 7 uses and is something similar.

I don't see how there could be any difference. If there is an apparent difference it must be down to loudness between tests. There couldn't be any tonal change between them as far as I can see.

PaulBrewer

There could easily be differences - how each company combines the huge numbers that summing digital audio entails could easily be a source of that difference.

Anima

No, that is what people say when this question gets asked but it's just rhetoric. Huge in what sense? Adding together various audio streams is simple addition and possibly scaling of the signals to reduce clipping.

The only problem, that I'm aware of, is that adding floating point numbers can accumulate a certain amount of error. This is negated by the fact that DAWs use 64-bit summing these days which is much more than you'd even need anyway (you'd never hear the difference). Even more so, CPUs internally have 80 bit extended FP arithmetic.

If there is an algorithm for some DAWs, it's not going to be complicated and hardly going to produce a particular sound. When you factor in that any track passes through x amount of plugins of various quality, this whole argument is just pointless anyway.

Neurojazz

If there was any part in the chain that was analog - then sure there would be a difference Paul.

But you are dealing with computers, and it's sadly a case of 128=128 (with no added magic fairy dust from either Mac or PC or it's DAW) affecting the sound.

Alternate ending:

But for Mac/Logic/Protools users, there are in fact tiny elves inside your machine rubbing the audio on the buttocks of scantily clad exotic slaves - that's why they charge you more.

PaulBrewer

Anecdotal evidence suggest otherwise - and what I'm asking is has anyone actually done any checks ?

powerplant

Neurojazz wrote: »

There was a few years back (and discussed on this forum)

All the outputs summed equally (so no tonal changes)

Since we don't really know how these DAWs are doing their maths (due to their proprietariness) you can't be completely sure. A samply by sample comparison would be the final decider though, is that what you were referring to in this output summing test Neurojazz?

flyswatter

The same mix will sound the same in any DAW.

Neurojazz

powerplant wrote: »

Since we don't really know how these DAWs are doing their maths (due to their proprietariness) you can't be completely sure. A samply by sample comparison would be the final decider though, is that what you were referring to in this output summing test Neurojazz?

Yep - remember some of the boards members did it in this forum, in the MP forum i think (not sure) - Can't remember if they took sine and then showed different DAW cancelling the phase or just ended up leading to Gearslutz threads. Was about 4 years ago if i remember correctly.

PaulBrewer

Neurojazz wrote: »

Yep - remember some of the boards members did it in this forum, in the MP forum i think (not sure) - Can't remember if they took sine and then showed different DAW cancelling the phase or just ended up leading to Gearslutz threads. Was about 4 years ago if i remember correctly.

I think if it's that long ago it might have lost it's value insofar as most DAWs will have moved on since then, possibly changing how they sum - certainly PT10 has been rewritten.

I remember my first dealings with digital mixing was with a Yamaha 02 and it always annoyed me how various parts of a mix would sound good soloed.

Also I recall the guys in SSL saying Logic summed either better or worse (I can't recall but I'll check!) than Ptools.

PaulBrewer

What made me think about this subject again was a recent review of Magix Samplitude in SOS .

The designers spoke of 'maintaining' quality and how on one up-date they made a 'mistake' which impacted overall audio quality.

madtheory

It seems the most common problem with ITB summing is poor use of dither. It should not be an option, it should be automatic. However, Pro Tools by default will truncate when going from 32 bit float to 24 bit (or 48 bit to 24 bit in HD). In HD there is an optional dithered mixer that gets around this, but for native you need to use a third party plugin to dither to 24bit. However, the artifacts are at around -128dBFS, so it's not always going to be audible. Personally, I prefer to avoid truncation as much as possible!

Funnily enough it turns out that the noise, distortion and crosstalk added by a typical well made analogue console can help to make a mix more coherent, and typically digital mixers lack all of those flaws/ features. Unfortunately no one has the time, money or interest to investigate the psychoacoustics of this. That said, the recent review of Slate VCC by Hugh Robjohns in SOS is most illuminating on the matter.

The coders of our DAWs are not at liberty to tell us exactly how this stuff is done, so we have to rely on our ears and measurements- which frankly is a PITA, because it often turns into hearsay and mythology, and stupid ego wars on the tinternot. Such as Paul's example above- what was the mistake? Who made it? What did it do? No one knows, but assumptions are often put forward as fact.

PaulBrewer

madtheory wrote: »

It seems the most common problem with ITB summing is poor use of dither. It should not be an option, it should be automatic. However, Pro Tools by default will truncate when going from 32 bit float to 24 bit (or 48 bit to 24 bit in HD). In HD there is an optional dithered mixer that gets around this, but for native you need to use a third party plugin to dither to 24bit. However, the artifacts are at around -128dBFS, so it's not always going to be audible. Personally, I prefer to avoid truncation as much as possible!

Funnily enough it turns out that the noise, distortion and crosstalk added by a typical well made analogue console can help to make a mix more coherent, and typically digital mixers lack all of those flaws/ features. Unfortunately no one has the time, money or interest to investigate the psychoacoustics of this. That said, the recent review of Slate VCC by Hugh Robjohns in SOS is most illuminating on the matter.

The coders of our DAWs are not at liberty to tell us exactly how this stuff is done, so we have to rely on our ears and measurements- which frankly is a PITA, because it often turns into hearsay and mythology, and stupid ego wars on the tinternot. Such as Paul's example above- what was the mistake? Who made it? What did it do? No one knows, but assumptions are often put forward as fact.

I threw away the SOS with the review in of Magix !

As I recall one of the designers said they had a complaint from 'an American' (!) who reckoned that the then new revision sounded worse than the old one. Designer guy said he couldn't hear a difference but they could measure one and it was traced to a 'mistake'.

Looks like I'll have to fire up a couple of DAWs myself and have a listen !

studiorat

In the past I've found Logic to sound different to ProTools. That is all the same flies playing with the faders at 0dB on both sessions.

I reckon the main difference is in the difference between the Pan law on each DAW.

There have been some tests done on each, because there are differences between the summing in terms of the interpolation algorithms used and the truncation for word length conversion. But these differences were in the region of -126dB, so they are hardly audible differences really.

Adyx

PaulBrewer wrote: »

I threw away the SOS with the review in of Magix !

As I recall one of the designers said they had a complaint from 'an American' (!) who reckoned that the then new revision sounded worse than the old one. Designer guy said he couldn't hear a difference but they could measure one and it was traced to a 'mistake'.

Looks like I'll have to fire up a couple of DAWs myself and have a listen !

In that case it was the dithering being left out when converting to 24 bit. When you have software errors like that of course it's going to impact on the sound quality. However, all other things being equal and assuming the coding is up to scratch any difference in sound is purely subjective.

It's the same as Monster Cable selling a €100 HDMI lead as having better picture quality than a €2 lead - assuming they both work the picture will be identical.

madtheory

PaulBrewer wrote: »

a complaint from 'an American' (!)

Which one?

woodsdenis

studiorat wrote: »

In the past I've found Logic to sound different to ProTools. That is all the same flies playing with the faders at 0dB on both sessions.

I reckon the main difference is in the difference between the Pan law on each DAW.

There have been some tests done on each, because there are differences between the summing in terms of the interpolation algorithms used and the truncation for word length conversion. But these differences were in the region of -126dB, so they are hardly audible differences really.

Correct, the Pan law is different in PT and Logic and it does make the same files sound different. I dont know if a mono sine wave would be any different though. I cant help but think that there would be subtle differences due to just coding and how each DAW flies the bits around internally. If there is, its minimal.

tweeky

If it was possible (or you could be *rsed) to import a bunch of stereo stems ie. drums, bass, gtrs, keys,vocals from any third party system (analog to DAT maybe?) and with all faders at zero db, pans at hard left and right and print the same mix in logic and protools/nuendo whatever. Combine the 2 mix files you want to compare with one pair out of phase and one in phase, the resulting file is the difference. But who cares! Some people prefer logic (me) and some people protools. I prefer the way logic "works" for me and the miniscule sonic difference between it and whatever else doesn't bother me. The pan law can be changed in logic, i prefer the -3dB Compensated option but i cant tell you why but i do.

madtheory

It's already been done- years ago. Glenn Fuston's Awesome DAWSum.

BTW did you mean DAW not DAT?

tweeky

I know i'd seen that before.
I meant DAT (with timecode) as an option for a non DAW digi recorder.

PaulBrewer

The differences I'm sure are indeed small - however with more and more people working exclusively 'Sa bhosca' (SB from here on in) it's worth a look.

madtheory

SB- love it!

Radiosurfer

Above posts on pan law are on the money here. This has been discussed exhaustively on Gearslutz and each of the forums related to each DAW and has been tested ad nauseum in countless shootouts. The same wav file imported into each major DAW you can think of will sound exactly the same when you equal all other factors (no plugins, same vol level, etc.) until you begin panning. Then the differences become noticeable.
While on the subject of pan laws, it seems that Harrison Mixbus have introduced a very nice sounding DAW that a number of people I have seen comment online in different forums say really comes into it's own when you start to pan your sounds around.
I have not used other DAW's extensively enough to comment on them all but I have seen enough of these threads and shootout results to say that at this point, the only difference you will find between DAW's is when you pan, and it's a subjective difference (i.e. none are "better" than the other, just different) so it's back down to your ears again as to which one you will prefer. There is no "best" in this situation (not anymore anyway).

woodsdenis

Radiosurfer wrote: »

. There is no "best" in this situation (not anymore anyway).

Exactly, however I think its fair to say that different DAWs will suit different types of recording. ProTools is a much better at handling/editing Audio than most, but sub par on a lot of other things. Logic is fantastic for midi and VI's but needs to get better at handling/editing Audio. That said if you become proficient you can always work around the deficiencies of your DAW and like any tool its important to learn how to use it. I remember seeing a very famous
Swedish producer tuning vocals syllable by syllable in Logic's Sample editor. I kinda mentioned why didn't he use Melodyne. His response was he had been doing like that for years and preferred it.

I know my ideal would be a hybrid of Protools and Logic taking the best out of each system.

PaulBrewer

I'd seen the 'per syllable' tuning approach in Logic coincidentally with a Swedish producer too !

The results were certainly very 'pop' ...

Radiosurfer

woodsdenis wrote: »

Exactly, however I think its fair to say that different DAWs will suit different types of recording.

This is so true. PT seems the least fuss and the most dedicated to actual "recording" but workflow has been the biggest influence on the DAW I use. As soon as I started using Ableton Live on a regular basis I just couldn't bring myself to go back to PT or Logic as everything seems to take twice as long to accomplish and I'm an impatient bastard.
Ableton Live is totally unsuitable for recording, but that's the one I use now. It's also nowhere near Logic when it comes to mixing (no bezier curves, seriously!) but I mix in Ableton. The single most important factor to me when I'm working is workflow. How fast can I accomplish what I want to do and how "in the way" does my DAW get. And for me, when it comes to workflow, I just can't go back from Ableton (despite it's many shortcomings).
Although, as stated throughout this thread, while Ableton Live has many shortcomings, once you know how to set it up properly, sound quality is not one of them. Anyone who claims any of the main DAW's is not up to scratch sonically does not know what they are talking about. That's not say you can't prefer the sound of one over the other, just that it's completely subjective.

woodsdenis

Daniel James who is really talented film/composer uses Ableton Live for his stuff.
The kind of music he does you would think would be more in the Logic mode, but he swears by it.

As you said its about workflow and the bloody thing not impeding that. I think it's a fair assumption that they all sound equal, and with relatively cheap high qualiity interfaces , sound quality isn't an issue. How you use them is the key.

Seziertisch

woodsdenis wrote: »

I know my ideal would be a hybrid of Protools and Logic taking the best out of each system.

It's called Cubase ;-)

woodsdenis

Seziertisch wrote: »

It's called Cubase ;-)

I haven't seen it personally in operation for years, though there are a few Nuendo devotees in Nashville. I am sure it is a robust program now but anyone who lived with Cubase Audio in its inception still has nightmares, ask tweeky. Mind that was a loooooooong time ago.
The one that looks interesting is the Prosonus Studio one, anybody used that.

madtheory

My second computer based sequencer was Cubase 3.0 on the Atari. The first studio I worked in full time, we bought Cubase Audio to go with the ADATs and 02R. It just never worked! We soon changed to Cubase VST, which by 3.0 was great. I stuck with that well past its sell by date because SX was awful. Changed to Pro Tools.

What I'm seeing of Cubase now makes me want to go back!

woodsdenis

madtheory wrote: »

What I'm seeing of Cubase now makes me want to go back!

Seriously! what has changed. I do think Protools is in a state of flux ATM. I would look at all alternatives at this point.

madtheory

The interface is intuitive like it used to be in VST, and it has all the pitch to MIDI, audio quantising, tempo map creation stuff that is a PITA in PT. Plus you've access to a huge amount of VST format plugins. The bundled eq is quite decent,as is the reverb. Haven't really tested the compression. All of the MIDI editing that I was used to in VST is there as well, PT is seriously lacking in comparison. Stuff like editing sustain pedal, and being able to multiply and reverse MIDI parts. Those are my personal faves anyway.

I'm loathe to change though, it would take a while to get up to the speed I have in PT.

woodsdenis

madtheory wrote: »

I'm loathe to change though, it would take a while to get up to the speed I have in PT.

I hear you, I am trying to justify the cost of upgrading HD to HDX and it's very difficult. Is Cubase crippled on a Mac because it's a PC centric program?

The real PITA is that all the major ones are lacking somewhere in functionality.

Radiosurfer

I mentioned it previously, but has anybody checked out Harrison Mixbus? It's supposed to be amazing according to the threads on Gearslutz. Based on Harrison consoles (which may not suit ITB people at all, but there you go) and has what sounds like an attractive "everything is on screen" layout. No hidden dropdowns or seperate windows to open.

madtheory

woodsdenis wrote: »

madtheory wrote: »

I'm loathe to change though, it would take a while to get up to the speed I have in PT.

I hear you, I am trying to justify the cost of upgrading HD to HDX and it's very difficult. Is Cubase crippled on a Mac because it's a PC centric program?

The real PITA is that all the major ones are lacking somewhere in functionality.

Ya, you end up using two. Can get awkward.

Haven't tried the Mac version of Cubase though we do have it here. Have you looked at Reaper? Audio is seriously good. No idea of the MIDI. But the whole thing is very intuitive.

Adyx

madtheory wrote: »

Ya, you end up using two. Can get awkward.

Haven't tried the Mac version of Cubase though we do have it here. Have you looked at Reaper? Audio is seriously good. No idea of the MIDI. But the whole thing is very intuitive.

I'm used to doing most of my midi work in Ableton Live, but I do most of my audio recording in Reaper. Maybe it's just what I'm used to but I don't think Reaper is as midi user-friendly as Live. I do love Reaper but when it comes to midi it's closer to Pro Tools (from what I've heard) and I'll take Live or Cubase any day. It's by far the best value for money software you'll ever get though.

noodle650

all daws will null when exact same settings are applied. The reason they may sound different to the users is down to workflow imo

PaulBrewer

Based on what evidence Noodle ? Anecdotal evidence suggests otherwise !

madtheory

PaulBrewer wrote: »

Based on what evidence Noodle ? Anecdotal evidence suggests otherwise !

Yes, and I think the evidence is more than anecdotal at this stage. We know a couple of facts... As said earlier pan law is one major variation. There would also be differences with types of dither, or in some cases no dither resulting in truncation artifacts. Note that all dither/ truncation issues are right down in the noise floor, which is about -128dBFS for 24 bit files (depending on the dither). These would show up in a null test, but in a mix would probably only be audible to trained ears using good headphones or a superb room. Another caveat is that one would tend to do the mix with these issues in place, so it is possible to work with them. It seems that can be a source of irritation though.

PaulBrewer

How does one select the dithered mixer in PT ?

madtheory

It's only in HD. LE/ M-Powered doesn't have any 24 bit dither, you need a third party plugin.

For HD last time I looked it was in the unused plugins folder IIRC. haven't looked at it since v8, and AFAIK it's fixed in v10. We don't have that here yet so I can't verify that for myself.

You really need to use it if you've a lot of channels and the mixer is taking up more than one TDM slot, because the bus between them is 24 bit, even though they're 48 bit internally. The dither is used when piping audio from one slot to another. One thing I haven't verified is whether you still need to put 24 bit dither on the master for the final mix, or if it's automatic in the dithered mixer.

Or you could RTFM!

PaulBrewer

RTFM ? Oh

woodsdenis

PaulBrewer wrote: »

RTFM ? Oh

The default is the stereo mixer in the plugins folder .The stereo dithered mixer is in the unused plugins folder. Swap them over.

PaulBrewer

Aha! Thanks.