I'm sure his comment was referring to summing and such, not just playback. Personally, I haven't heard much difference, and I've had experience with all the popular ones, but the only one that i've seen consistent claims of better sound for, coming from people I tend to trust, is Samplitude/Sequoia. Whatever the truth is, greta results are possible from any, so I really don't think this should be fussed over. What you do is infinitely more important that what you do it on... just ask any woman.

I started out way back when with Mixman Studio "Pro." I never thought twice about audio quality in those days...but when I outgrew it's track-count and bought Acid "Pro" 2, I noticed a certain -- uh, crappiness to the sound. This, I suppose, could be attributed to the sampling rate, the stock sound card, and the effects of pitch-shifting sounds down like 78 semi-tones. When I first began with Cubase SX, I thought some kind of veil had been lifted...but this change was accompanied by a new compy, interface, and speakers. Down at The Spot there are some cats who use Logic exclusively...but I think the quality is affected way more by the processes it undertakes than a program that effectively routes it to those processes and sums it.

But honestly, it's a cross between an elephant and a rhino.

Speaking of ridiculous amounts of pitch-shifting...Do you know of any VST plugs that specialize in slowing sounds down to their hilly essence or vice versa? I think ACID put a "limiter" on their pitch-shifting at some point...And I think Cubase only allows 16 semi-tones.

if you have a decent soundcard / monitor setup just try and listen for yourself... i personally have lots of experience with differences in how software itself sounds, i think it's logical, just look at plug-ins the difference there is sometimes above "huge" the same applies for any type of programming, the way apps are programmend does affect the sound... that's what i hear and experienced, however...

but that's jsut my two ears...

Probably the most overlooked aspect of this issue (summing results from different host audio apps) is dithering.

Summing 18+ tracks of 24-bit audio down to just 2 tracks (stereo) at 16-bit means large amounts of audio data is removed. While all the main music hosts include fairly decent dithering algorythms - using them correctly really is crucial for getting a useable production master.

What's worse is, most of the major host apps do not go into nearly enough detail when dealing with dithering.

For example, Cubase SX - did you know that only the last two inserts of the main stereo outs are post fader? Sure its in the manual, but not why that is so important!

-> if you use, lets say, Waves L2 (mastering limiter) as your master limiter and use it's "IDR" dithering feature, it must be placed in the last insert slot of the Main Out, otherwise your perfect mix get's run through the laundry machine! Like this....

multitracks -> main out bus -> insert with dither to 16-bit (accidentally placed in a pre-fader slot) -> conversion back to 32-bit (Cubase's internal resolution) for the master fader level -> outputs

Note: that main output actually has no dither! Just hack-n-slay truncation to 16-bit unless the L2 with dithering is placed in one of the 2 post-fader insert slots (that's the last two slots, they've also got black instead of white numbers).

So definitely definitely no matter which audio host software used: be sure to give attention to dithering!!! The dither should be the last process to touch the sound. -Post fader-.

Its crucial to getting a vibrant mixdown that plays back on a wide range of CD players without artifacts or distortion.

Hey, hey, hey ... what the hell is "dithering" anyway?!?!

So far Izotope is the only company that gives dithering a priority in its manuals. These two PDF files from the werbsite are probably the best dithering guides written for musicians, sound engineers....

;)

[url]http://www.izotope.com/products/audio/ozone/OzoneDitheringGuide.pdf

The following was originally posted by user "Nick Mulder" in Andy Sneap's forum:

--------
Hi all,

I have to advise my school about a new to build DAW. All over the net I read claims like "DAW A sounds much better than DAW B", escpecially since I read the ($2500) SAWstudio forum.

So I did this: I took 4 short 16/44 mono tracks, imported them in PT. Track 1 at -11 db, track 2 at -17, track 3 at -23, track 4 at -27.

I bounced them in PT. I phase-inverted that bounce.

I set up an identical mix in SX2. I imported the inverted PTbounce an added it to the SX2 mix. Result: silence, the inverted PTmix totally zero'd the SXmix.

PT summing and SX summing is identical.

So I took the PTmix to SAWstudio, same story there.. Identical 4 tracks, identical volumesettings, with the inverted PTbounce in the mix I again had the purest silence I ever heard.

Funny thing is that the coder of SAW claims that the 'superior sound' of his coding is because of his integer summing engine, well du'h, it does exactly the same math as ProTools and Cubase (Nuendo) since the result of the different DAW's zero eachother.

So I thought "maybe it's not just the mixing, it's the way the different programs handle the plugins".

I added a Waves C1 with preset#2 to the tracks (and learned in the meantime that PT doesn't correctly handle plugindelaycompensation, when I inserted a C4 and put it on bypass the mix didn't zero anymore) and repeated everything: again, total esoteric meditative silence.

I started this because I was interested in the claim of the SAWfolks that their DAW sounded so much better then all the other. Well, with mixing it didn't, with pluginhandling it didn't, so what's left of this claim is the quality of the onboard EQ.

So I took the pluginversion of the internal SAWeq from JMLlabs and put it in Cubase. Did the same phase inverting trick, imported the result in SAW and I could zero the mix again by setting SAW's onboard EQ at the same settings I did with the JMLplugin.

Maybe SAW is very good coded, maybe the eq sounds pretty good, but its claim of beeing sonically superior is BS.

Just as BS it is to claim Nuendo as a better sounding tool than PT and viceversa.

...that NOBODY did this trick to confirm the (false) claim of the (any) coder is scary in the "do people really need to be this ignorant" way.

So its all up to proper eq'ing, dynamics, room and automation that makes a great mix. On which platform you do it doesn't seem to make any difference (if you have the same plugins ofcourse=).

Bye,

Nick Mulder
-------

KJ
----------------
Kyro Studios

Hey thanks for posting that. That's very interesting - especially since I've been hoping for a solid reason to ditch SX.

Would be good to double check that person's test and compare results. The test in itself is perfect.

If there are any inconsistencies between mixdowns made on two different platforms, then phase inverting one and mixing it with the other (both at equal volumes) will reveal all.

The only troubling omission is dithering!! No mention of it probably means done without dithering. Would be interesting to run the same test to compare different dithering algorythms - for example Waves IDR vs. Apogee's UV22. But you'd only be able to hear in what ways they differ, not which is "better".

As for SawStudio - I'd probably switch over to that host app if only:
* it didn't have the most garrish GUI ever!
* it was affordable

Imagine Sawstudio with the look of MOTU's Digital Performer... no sickening color diseases, no ridiculous fonts.

You can customize the look of SAWstudio...

http://www.sawstudio.com/SAWStudioShades.htm

I recently saw one that was aimed to look exactly like cubase, for example, and it was kinda close - surely, it was easier on the eyes then the above examples.

Arrowfan wrote: If there are any inconsistencies between mixdowns made on two different platforms, then phase inverting one and mixing it with the other (both at equal volumes) will reveal all.

This so called 'polarity test' is a red herring. It only provides a false testing methodology for people insecure in there own listening abilities and judgement.

How is one to perceive subjective audio quality by listening to SILENCE!? Almost any such polarity test, analog or digital, will result in simular levels of silence (gross device deviations not withstanding). A proper method would be to level match and blind a/b two audio devices in a calibrated listing environment.

I can share that my experience with comparing DAW audio quality has shown that sonic differences do exist between them. They have enough of a sonic signature that test participants could eventually pick which DAW they were listening to, blind.

If two tracks null, then there is no difference. If they do NOT null, then you need to use your ears to decide which one is better.

If you're hearing a difference between two tracks that will null when inverted ... :?

I believe different programs sound different. I have not done scientific testing to confirm that.

This so called 'polarity test' is a red herring. It only provides a false testing methodology for people insecure in there own listening abilities and judgement.

Zilla- you must have misunderstood the point here. IF there is silence, pure silence, it means the two mixdowns are identical. It means that both software apps performed exactly the same processes in exactly the same way. There is nothing "false" about it whatsoever.

If there is NOT silence... then people can debate about which is "better" while the rest of us dine on red herring.

This is an interesting post - one that always sparks debate.

First, let me come right out and state that I agree with Zilla on this. Furthermore, I don't believe the outcome of the above "test" as I've done similar tests and gotten completely different results. The test that I did was simple - record the exact same stereo wave directly into the DAWs (Sequoia and Cubase SX2). Both DAWs recorded a simple left and right channel panned as hard left and right as the DAW would allow (later, I also repeated the test with both tracks mono-ed out).

BTW - the signal was from the same converter going into two different PCs.

First, for the subjective -

Cubase sounded heavier or weightier in the bass region and somewhat rolled-off in the higher frequencies - also the width was, to me, less than desirable.

Sequoia lacked any of the extraneous weightiness of the Cubase signal and seemed to have further extension into both the lower and higher frequencies. The width of Sequoia seemed to be more appropriate to what I actually recorded.

Second - for the objective -

Doing a direct comparison between the waves (using both FFT analysis and the mono with phase reverse test) showed that the two recordings were different. The surprising fact was just how different they were. I would have expected minor differences but I simply got more than I expected.

I did this test after a lengthy discussion with fellow poster David Spearrit. I never published the results since the discussion got WAY out of hand and people started posting details pertaining to chips and instruction calls that was WAAAYYY over most peoples' heads.

BTW - no dithering was performed - everything was recorded at 24 bit 44.1 kHz.

In my personal experience, I have found certain DAWs to possess a linear, truthful quality and others to possess a sound which lends itself to specific types of music. In the linear camp, I have found:

Sequoia (and Samplitude)
Pyramix
Wavelab
Sonic
Sadie

Sorry Mac-ites, I haven't had the opportunity to play around much with Logic or PT.

J.

cucco wrote:

This is an interesting post - one that always sparks debate.

First, let me come right out and state that I agree with Zilla on this. Furthermore, I don't believe the outcome of the above "test" as I've done similar tests and gotten completely different results. The test that I did was simple - record the exact same stereo wave directly into the DAWs (Sequoia and Cubase SX2). Both DAWs recorded a simple left and right channel panned as hard left and right as the DAW would allow (later, I also repeated the test with both tracks mono-ed out).

BTW - the signal was from the same converter going into two different PCs.

First, for the subjective -

Cubase sounded heavier or weightier in the bass region and somewhat rolled-off in the higher frequencies - also the width was, to me, less than desirable.

Sequoia lacked any of the extraneous weightiness of the Cubase signal and seemed to have further extension into both the lower and higher frequencies. The width of Sequoia seemed to be more appropriate to what I actually recorded.

Second - for the objective -

Doing a direct comparison between the waves (using both FFT analysis and the mono with phase reverse test) showed that the two recordings were different. The surprising fact was just how different they were. I would have expected minor differences but I simply got more than I expected.

I did this test after a lengthy discussion with fellow poster David Spearrit. I never published the results since the discussion got WAY out of hand and people started posting details pertaining to chips and instruction calls that was WAAAYYY over most peoples' heads.

BTW - no dithering was performed - everything was recorded at 24 bit 44.1 kHz.

In my personal experience, I have found certain DAWs to possess a linear, truthful quality and others to possess a sound which lends itself to specific types of music. In the linear camp, I have found:

Sequoia (and Samplitude)
Pyramix
Wavelab
Sonic
Sadie

Sorry Mac-ites, I haven't had the opportunity to play around much with Logic or PT.

J.

OH WOW !!! finally someone seems to have EARS, and is not only reading data sheeds or going into mathematical micro sience...

yes there are MASSIVE differences in sound quality between software products, between all of them!!

I don't understand the issue folks have with files that null. Digitally reproduced music is a digital record of a waveform. That digital representation will be decoded back to an analog waveform.

If two files contain the exact same data and are played back through the same reproduction system, then the resulting analog forms will be identical. If the files do NOT contain the exact same data, they will not NULL when inverted and will not produce the same audio waveform.

If two DAWs create files with the exact same data, then it is no different than making a simple copy of one files. If the data is the same, and the playback setup is the same, any audible differences are happening between your ears, not outside of them. FWIW, I would NOT expect different DAWs to create identical data. I have Vegas, Audition, and Samplitude - and they do not sound the same.

I can see two possible places where there could be misunderstanding here.

1) The same data played through different software can sound different.

I would not refute this for a second. Clearly, programs will use different algorithms to decode the signal which will result in audible differences.

2) A visual flat-line is being interpreted as a NULL, but zooming in would reveal differences.

That means the files aren't the same - they just aren't being looked at closely enough and will sound different.

zemlin wrote: I don't understand the issue folks have with files that null....If two files contain the exact same data and are played back through the same reproduction system, then the resulting analog forms will be identical.

Yes, that is the theory. The issue that I have with null tests (and even data-verification tests, for that matter) is that they often do not correlate with what I am hearing in the real world. The tests may tell me that a file is bit for bit perfect, which I am totally willing to stipulate. But then why is it that I can get a number of trained listeners to accurately destinguish/identify two bit-for-bit perfect files, blind? Apparently something other that just data content is relevant to audio playback quality. Therefore the null test, while being a good indication of an error free copy, does not tell the whole story: it does no "reveal all" as stated in an earlier post.

zemlin wrote: A visual flat-line is being interpreted as a NULL, but zooming in would reveal differences.... they just aren't being looked at closely enough and will sound different.

This is the other big beef I have with this test. Are people really being careful? Most people can't even set up two speakers for stereo correctly. Do all software programs actually display that last bit in the waveform?

The main point I would like to get across is that no single test "reveals all". Even a battery of tests will hardly tell you whether a device has high fidelity. The ear must be the final judge, and the ear cannot judge by listening to silence.

Zilla wrote: [quote=zemlin]I don't understand the issue folks have with files that null....If two files contain the exact same data and are played back through the same reproduction system, then the resulting analog forms will be identical.

Yes, that is the theory. The issue that I have with null tests (and even data-verification tests, for that matter) is that they often do not correlate with what I am hearing in the real world. The tests may tell me that a file is bit for bit perfect, which I am totally willing to stipulate. But then why is it that I can get a number of trained listeners to accurately destinguish/identify two bit-for-bit perfect files, blind? Apparently something other that just data content is relevant to audio playback quality. Therefore the null test, while being a good indication of an error free copy, does not tell the whole story: it does no "reveal all" as stated in an earlier post.

zemlin wrote: A visual flat-line is being interpreted as a NULL, but zooming in would reveal differences.... they just aren't being looked at closely enough and will sound different.

This is the other big beef I have with this test. Are people really being careful? Most people can't even set up two speakers for stereo correctly. Do all software programs actually display that last bit in the waveform?

The main point I would like to get across is that no single test "reveals all". Even a battery of tests will hardly tell you whether a device has high fidelity. The ear must be the final judge, and the ear cannot judge by listening to silence.

Are the trained listeners listening to the various Software packages through the same soundcard? Since bit for bit identical is stipulated my thought is that the only conversion remaining (assuming no dithering as mentioned above) to affect the sound is the soundcard's conversion back to an analog waveform. In which case, its a soundcard issue more than a software issue I would think.

axel wrote: .. there are MASSIVE differences in sound quality between software products, between all of them!!

While I have experienced quality differences between various DAWs, I don't know that I would go as far as to say they have "MASSIVE" differences. Enough to form a preference; yes. But not enough to say that the sky is falling

hi Zilla,

fair enough... i am a sonic maniac so to me it does!!

to some others don't. there are mp3s spreading the entire world (and listeners ears) with essentielly music, but to me the sonic aspect is lost.

TheArchitect wrote: Are the trained listeners listening to the various Software packages through the same soundcard?

Yes.

TheArchitect wrote: ... my thought is that the only conversion remaining...to affect the sound is the soundcard's conversion back to an analog waveform. In which case, its a soundcard issue more than a software issue I would think.

Possibly, but I think that there is even more going on than just that. Hard drive performance/formating/buffering, IDE/SCSI bus mastering, cable impeadances, software algorithims, cpu interrupts, OS services & processes, fsb speed, pci bus traffic, card drivers, card hardware...and so on. Our audio data must traverse a very complex system, indeed. Who knows how all those elements might influence the final, reconstructed sound?

Zilla wrote:
Possibly, but I think that there is even more going on than just that. Hard drive performance/formating/buffering, IDE/SCSI bus mastering, cable impeadances, software algorithims, cpu interrupts, OS services & processes, fsb speed, pci bus traffic, card drivers, card hardware...and so on. Our audio data must traverse a very complex system, indeed. Who knows how all those elements might influence the final, reconstructed sound?

I do computers for a living and I will agree cable impedance anywhere the audio is in an analog form and software algorithms could have influence but based on the above scenario cables would be the same. but I suppose there could be something in how the tracks are summed to the stereo output.

The rest of the stuff you listed however are of a 'it works or it doesn't' nature. If those items aren't working properly you hear it in the form of clicks, pops and dropouts not in an altered frequency response.

All that said I'm not sure the variation in sound is significant enough that it is all that important. Variation in head alignment, gauss strentgh, variation in tape formula etc generally meant that no two analog machines sounded the same let alone the tone stacks in 2 different mixers influencing sound or the variations in different monitors. I would bet that that the variations from the analog realm added up to a lot more variation than we see from one app to another.

Well,

here is my 2 cents,

we haven't done the A/B test but we do work with
three high profile producers who have.

Interestingly, all three came to the exact same conclusion
that Samplitude and Sequoia induce no phase shift during a
mix inside the box resulting in better stereo image, sound
clarity and overall better sound quality than the other
leading two DAWs.

Guy Cefalu
Sonica Audio Labs

TheArchitect wrote: ...I'm not sure the variation in sound is significant enough that it is all that important....variations from the analog realm added up to a lot more variation than we see from one app to another.

That is a pretty healthy way to look at it (assuming that conclusion is based on actual listening, and not only meditation). Also, you have made a good point about variations in all types of equipment. If something in the signal chain has changed, one is likely to have a change in audio quality. Therefore it is wise to always evaluate such changes. Sometimes its significant, other times it is no big deal.