Topic: Are There Any Sonic Differences In DAW Software?

[stevetoney]

I know people have their software preferences based mainly on features and/or workflow tendenices, but I've always wondered if there have been either empirical or anecdotal comps done to determine if any of the software's processing has been shown to affect sound quality any better or worse than another package? 

To perhaps clarify this question a bit, I'm not talking about whether one package has a better sounding set of special features, like effects packages...say reverb or whatever.  I'm talking about some of the basic functions we do here on probably close to 100% of our recordings...dithering, normalizing, and for some people EQ. 

[bryonsos]

Marking thread.

I don't know anything about this topic, but I'm guessing that since math is involved there could be subtle differences depending on how many significant digits are used, rounding up or down etc. Different packages might even use different equations for some functions 

[Gutbucket]

My impression is that there is little if any difference at this point for basic operations but once was a legitimate concern when available computing resources were far less robust and not all software worked natively in higher bit depths with floating precision.  In combination with vastly increased computing resources, software routines have advanced and matured, dither routines improved, etc.  Time marches on.  Basically there is no excuse anymore- at this point if the basic operations are not completely transparent, the software is broken.

That applies to the modern versions of mainstream computer audio programs.  Lightweight mobile apps may cut corners and suffer for it like older software.

Like you mention, I think sonic differences at this point are more likely to be audible in more complex effects like compression, EQ and other more advanced manipulations. The constraints on the ultimate audio quality in those things are still resource driven to a large extent- both in required computing resources and monetary expense (the better effects cost more and consume more processor time and memory space).

But generally, I think a bigger issue now is usability and getting things to do what you want easily and intuitively.

That's all my 'know enough to be dangerous ' opinion.  Software programing experts feel free to correct me.

[F.O.Bean]

I would say they're mostly all on par with each other. But that said, I've ALWAYS used WaveLab, so I will continue to do so!

[stevetoney]

But generally, I think a bigger issue now is usability and getting things to do what you want easily and intuitively.

Lee...thanks for the previous answer...I hope this followup question doesn't derail the thread at all, because I'd love to hear some more expert opinions on my initial question, but a followup question is...

Has the digital technology we use now (I guess that would be using 24/96 or 24/192 resolutions with better processing algorithms discussed above) gotten us beyond the concern that Neil Young famously expressed way back when when he said that digital technology was ruining music?

[stevetoney]

I would say they're mostly all on par with each other.

Basis?

[bryonsos]

My impression is that there is little if any difference at this point for basic operations but once was a legitimate concern when available computing resources were far less robust and not all software worked natively in higher bit depths with floating precision.  In combination with vastly increased computing resources, software routines have advanced and matured, dither routines improved, etc.  Time marches on.  Basically there is no excuse anymore- at this point if the basic operations are not completely transparent, the software is broken.

That applies to the modern versions of mainstream computer audio programs.  Lightweight mobile apps may cut corners and suffer for it like older software.

Like you mention, I think sonic differences at this point are more likely to be audible in more complex effects like compression, EQ and other more advanced manipulations. The constraints on the ultimate audio quality in those things are still resource driven to a large extent- both in required computing resources and monetary expense (the better effects cost more and consume more processor time and memory space).

But generally, I think a bigger issue now is usability and getting things to do what you want easily and intuitively.

That's all my 'know enough to be dangerous ' opinion.  Software programing experts feel free to correct me.

I think you're basically correct. In my job we use a lot of statistics software. My students use several different software packages depending upon ease of use. Awhile ago I noticed that SAS would routinely give a slightly different answer than other packages. I learned that SAS carries forward 8 significant figures while others only carry 4. This means that SAS is technically more accurate, but uses more memory, is not as fast etc. Mostly the different answers are splitting the split hair of the split hair and are irrelevant. For DAW software, I'm curious if these types of artifacts are audible or irrelevant. I'm also interested to learn if there are, for instance, different ways of calculating a normalization function, or if it's a standardized set of equations.

[F.O.Bean]

I would say they're mostly all on par with each other.

Basis?

Well, over the years I've used samplitude, sound forge, and WaveLab, and IMO, they all sounded pretty similar to me even tho I've never done a direct comp. I use WaveLab because I have always LOVED Apogees dithering scheme and WL crystal resampler sounds amazing IMO. But I use WL because I've used it mainly for the last 10 years and I know how to us it better than the others I've mentioned!

[Todd R]

Hmm, good question (well initial question, personally, I don't care much about Neil's belief that digital has killed audio -- everything I do is digital, and it won't change, so from my standpoint, it is what it is).

Not to go against Lee and Bean, but I would have thought that it matters.  Somewhere on my home computer I have the link that has been posted several times (maybe by Lee?) of various spectral analyses of a whole bunch of dither routines.  It seems that the quality of dither routines matter, I would have thought the same for DAW packages.

I think Lee is right and that the packages rely on the available computer resources and native operations that are available (say, what is available inside the Mac for the DAW s/w to work with).  I wouldn't imagine there is much difference in easy tasks like normalization or amplification, but compression, dither, and EQ could be a different ball of wax.

I only know enough to allow myself to sound like an idiot, but there are differences between a typical multi-band EQ and parametric EQ, and companies making EQ plugins talk about phase coherency and time coherency or whatever. Then there are essentially proprietary compression schemes like Waves L1 and Waves Ultramaximizer, etc.

I'd imagine that most of the known DAW software packages sound pretty good, but I'd also imagine there might be differences in the quality and availability of various compression and EQ effects.  So if you are using say a more basic 12-band equalizer to accomplish something in one software package, someone might be getting much better results using a more advanced parametric EQ in another package.  All that said, if you are using VST and AU plugins to accomplish your workflow, other than user interface, those shouldn't have a difference in sound regardless of which DAW software you use.


Who knows?  I use Amadeus Pro since it is pretty cheap, runs on a Mac, does most everything I want it to, has a good feature set (say for both editing, tracking, and saving to flac, etc), seems pretty fast, and is familiar to me since I've been using it for years.

[Gutbucket]

Has the digital technology we use now (I guess that would be using 24/96 or 24/192 resolutions with better processing algorithms discussed above) gotten us beyond the concern that Neil Young famously expressed way back when when he said that digital technology was ruining music?

Hmmm, where's my 10' pole? 

IMO, we are well past the point where digital technology is advanced enough to do no harm, much less 'ruin' music, yet people inevitably find new ways to use it to ruin music anyway. "Digital doesn't kill music, people kill music"

Tools and guns, hammers square pegs and round holes, forests and trees and all that.

[Gutbucket]

To break it down a bit, there are different things in play.  Take dither and resampling for instance, two very different things in several ways-

As I understand it, at least computationally, dithering is simple.  It's just adding low level random noise.  The question of it's quality is one of understanding human perception and adapting the particulars of that noise to best satisfy that.  Doing a good job there takes (took? does much development go into new dither schemes anymore?  it's basically a answered problem now, isn't it?) research and lots of testing, but not much computational resources.  Once those basic questions are well understood, the world moves on and it becomes a question of how to implement them.  We build on the foundation.

Resampling is more of a straight mathematical problem, even though the ultimate test is it's audibility (or lack of it). The sample rate conversion problem is more easily understood on a straight mathematical basis and shortcomings addressed by simply throwing more processing resources at it- better mathematical schemes and the use of more significant places past the decimal point.

I suppose all processing routines are a combination of those two basic issues, some weighted more to one side than the other.  When the problem is well understood mathematically, the questions become- how precise do we need to be? and how much does it matter? Those are more difficult to answer perceptual questions.

In addition there is the cost of things.  Previously written routines are less costly than developing new ones, and better routines are sold at a premium price if the market will bear it.  Some of that performance difference is real in terms of audio perception, some is simply market driven brand marketing perception, Some marketing plays that up.  No surprise there.

Like dither, consider lossy file size compression encoding schemes which are costly in development because of the not super well understood human perception issues, but are computationally trivial at this point to implement.  The know-how for how to do the processing part of it is mature.

[stevetoney]

Hmm, good question (well initial question, personally, I don't care much about Neil's belief that digital has killed audio -- everything I do is digital, and it won't change, so from my standpoint, it is what it is).

I wish I'd have phrased the question differently.  The intent of the question was to keep it somewhat germane to the initial question...so I'll re-phrase without the Neil Young reference...

Does the technology we have available to us today (e.g. 24/96 and 24/192 resolution along with great algorithms and fast computer processing) generally allow us to make our digital recordings that would be the sonic equivalent of an analog version?  Would an average human being hear a difference or is there farther to go on this (say 32/384 or whatever).

(I'm not really even sure if that's the best way to ask this question...I wasn't really looking to start an analog vs. digital debate although maybe that's inevitable, but was really just curious and looking for a yes/no answer from you guys that read up on this stuff.  My thoughts were that the answer is 'yes, it's generally accepted that most people don't hear a sonic difference', but I haven't done any research and wanted to hear from others that have.)

[Gutbucket]

I think so, but I'm a committed digital dude like Todd, and also more pragmatist than idealist.  I'm very particular about the sonics but have concluded that recording and storing my stuff at 24/48 is sufficient, though processing at higher rates is advantageous. 

That certainly doesn't mean that's what you get commerically though!  The delivery format becomes less of a question than what is stored on it, and what was done to it.

[stevetoney]

I think so, but I'm a committed digital dude like Todd, and also more pragmatist than idealist.  I'm very particular about the sonics but have concluded that recording and storing my stuff at 24/48 is sufficient, though processing at higher rates is advantageous. 

That certainly doesn't mean that's what you get commerically though!  The delivery format becomes less of a question than what is stored on it, and what was done to it.

All true.  I think almost all of us are digital, so I guess it really is a moot point.  I couldn't agree with you more about being pragmatic, although for some reason that position feels at odds with the fact that I'm a professional engineer.    

[Chuck]

I'd have to agree with ToddR and Gutbucket regarding digital recording and editing being better than analog. Digital recorders/ compressors/ limiters/ delays etc... are much better than the old analog counterparts in the fact that they are more linear. When you run something through analog equipment distortions are introduced via op-amps, resistors, grounding and on and on... Not so much with digital.

Digital gear also introduces distortions, but in my experience it generally doesn't introduce as much audible distortion.

Some of the distortion introduced by analog gear is perceived as sounding good. I think that is where the disconnect is. Some people like the tape saturation and other audible distortions introduced with analog gear better than the cleaner sound of digital.