Samplitude sound quality

[drpenguen]

i am with you actually but to make it clear; and some other things than i am not so with you

Originally Posted by laser

Read carefully my friend:
Originally Posted by Bob Olhsson
Samplitude is virtually the only DAW that defaults to properly dithering the audio feeding the interface. (all D to A chips have 24 bit inputs.)
I've found that adding 24 bit dither to a Pro Tools LE mix bus eliminates MOST OF THE DIFFERENCES in sound. Unfortunately this requires a third party plug-in such as Ozone or a Waves L1.

if you use a different dithering even a basic stereo source will not NULL...

samp has pow-r dithering and and its available to buy.
comparing paw-r to L1 is rather insulting too

anyway;

dithering has got nothing to do with the actual audio engine.
discussing specific futures of softwares is a different matter.

i have performed NULL test too.
with different variables. same plugs with same settings etc
and do simply null.
BUT
when automation involved. it was different story.
it doesnt even null inside the same sequencer.
every pass/BTD/RTD/export is a different result when automation involved


basically on automation NULL value changes through out the song...
from perfect NULL to very audible...

EDIT:

personally i dont use any automation on mixing stage. any automation i make while i working on a new track i bounce it to disc.
and delete the automation. but that suits my work flow. not for everybody.
when i am ready to do the final mix all my automation are already printed.
only minor level automatiom. but i rather do that on the OTB. not ITB
so if i take those files and mix them in any DAW. and copy the same plugs to any other DAW.
they will all sound the same.

people with heavy automation, may claim their sequencer sounds better than others... and i dont think any one can
stop this... since same track doesnt even null inside the same sequencer with automation.!!!

[work2do]

Originally Posted by drpenguen

i have performed NULL test too.
with different variables. same plugs with same settings etc
and do simply null.

Lynn is correct, you're not. You just say you've done this and that with no proof. I'll trust my ears and the pro's who have actually done extensive real world documented testing like Lynn.

Originally Posted by Lynn Fuston
Let me be very specific though. I'm not saying all DAWs sound alike. All my testing proved was that summing (of 22 channels) ONLY, when done correctly, is the same on the DAWs tested.

Throw plugins (and fader movements, panning, eq'ing) into the mix and all bets are off.

As soon as someone figures out a way to compare two things with a near-infinite number of variables, then we'll be well on our way to a way to devising a (definitive) test.

One thing I discovered in my testing is that a single variation, such as a value difference of .1 dB on one fader OR a value difference of 1 click on a pan setting OR moving ONE file of the 22 by ONE SAMPLE can cause an enormous chan

When discussing a DAWs engine you have to include all the basic functions and features we use on a daily basis as part of the "sound". What good is it if it nulls but then as soon as u put on plugins, automation, and dither it sounds like crap? That is why u can say one daw "sounds better" than another when it's used in the real world.

[laser]

Originally Posted by drpenguen

anyway;

dithering has got nothing to do with the actual audio engine.
discussing specific futures of softwares is a different matter.

You're joking, right?

Laser

[drpenguen]

Originally Posted by laser

You're joking, right?

Laser

not really

if you do SRC etc yes it matters...
and as Bob Olhsso commented, when used dithering it made it closer... but he used different dithering.

am i misiing something

EDIT :
on SRC result page, i think Logic was the cleanest...

[laser]

Originally Posted by drpenguen

not really

if you do SRC etc yes it matters...
and as Bob Olhsso commented, when used dithering it made it closer... but he used different dithering.

am i misiing something

Yes, quite a bit, actually.

Dithering has to do with quantization and not SRC. Bob's point is that all DAWs have to, at some point, go into 24-bit D-A converters. How and how well the program dithers has a significant impact on the sound of a DAW. It is an integral part of the sound engine.

Laser

[drpenguen]

Originally Posted by laser

Yes, quite a bit, actually.

Dithering has to do with quantization and not SRC. Bob's point is that all DAWs have to, at some point, go into 24-bit D-A converters. How and how well the program dithers has a significant impact on the sound of a DAW. It is an integral part of the sound engine.

Laser

[EDIT : i was refering to this actually
SRC Comparisons
EDIT end]


definition of dithering you are right. i was assuming
32/96 to 16/44.1 staff. apologies...
poor english doesnt help that much sometiimes

anyway,

we are doing everything ITB i assume

i mean on my null test
i took 24/44.1 alredy recorded files imported to 24/44.1 project files,
and exported them 24/44.1 again... and NULL on all

i am little lost here,
re dithering and using different dithering options
i have already mentioned. youi can use paw-r dithering on protools too

[Leon EA]

Originally Posted by laser

Yes, quite a bit, actually.

Dithering has to do with quantization and not SRC. Bob's point is that all DAWs have to, at some point, go into 24-bit D-A converters. How and how well the program dithers has a significant impact on the sound of a DAW. It is an integral part of the sound engine.

Laser

Dithering is also required if you bounce to a 16 bit or 24 bit file, not just when hitting the converters. But since it is done post master fader that kind of is included in your statement.
Issue here is that if you bounce processed files to save up CPU and mix those again, you have the choice between added dithering noise (multiple dithering, which shoudl be avoided), or using truncated files. Even though at 24 bit I preferred the 24 bit bounces that need to be mixed instead of multi dithering noise.

I do think dithering does not belong in the "DAW sound" discussion, but since there is a big difference between dithering algorithms, it is surely a valid question to value the "sound excellence" of a DAW with its capability of loading the best dithering algorithms. POW-R is considered the best, but AFAIK it is not available as a separate VST plugin.

I don't have time at all, but I cannot imagine that, unlike DAW summing, different EQ algorithms are nulling. I will test that in the near future.
Also, I wonder what DAW's get different results with automation bounces. Without 3rd party plugins or instruments (to avoid unsynced osc etc) I can't imagine pan and level automation doesn't null.

[Tha Wiizard]

F**K all these tests, Samplitude sounds the best, period.thumbsup

[monomer]

Originally Posted by laser

You're joking, right?

Laser

Nope, he's perfectly right.
Once you get to know how the audio engine performs you can go on and start researching the dithering.

[monomer]

Originally Posted by laser

Yes, quite a bit, actually.

Dithering has to do with quantization and not SRC. Bob's point is that all DAWs have to, at some point, go into 24-bit D-A converters. How and how well the program dithers has a significant impact on the sound of a DAW. It is an integral part of the sound engine.

Laser

Well, since the dithering on 24-bits convertors is @ -136dBfs and the self noise of most convertors is above that, how would different dither make a significant impact on the sound?
And i mean after you have set up a proper gain structure where -136dBfs is not perceivable by the human ear...

[White Falcon]

F**K all these tests, Cubase 5 sounds the best, period. I'm serious!

[monomer]

Starting to wonder about those EQ's..
Did anyone do some null tests with noise or other full bandwidth material?

[drpenguen]

Originally Posted by monomer

Starting to wonder about those EQ's..
Did anyone do some null tests with noise or other full bandwidth material?

i did it last week

sonnox eq vs PSP vs URS etc


i used white noise as well as commercial music files...

[drpenguen]

Originally Posted by Leon EA

Dithering is also required if you bounce to a 16 bit or 24 bit file, not just when hitting the converters. But since it is done post master fader that kind of is included in your statement.
Issue here is that if you bounce processed files to save up CPU and mix those again, you have the choice between added dithering noise (multiple dithering, which shoudl be avoided), or using truncated files. Even though at 24 bit I preferred the 24 bit bounces that need to be mixed instead of multi dithering noise.

I do think dithering does not belong in the "DAW sound" discussion, but since there is a big difference between dithering algorithms, it is surely a valid question to value the "sound excellence" of a DAW with its capability of loading the best dithering algorithms. POW-R is considered the best, but AFAIK it is not available as a separate VST plugin.

I don't have time at all, but I cannot imagine that, unlike DAW summing, different EQ algorithms are nulling. I will test that in the near future.
Also, I wonder what DAW's get different results with automation bounces. Without 3rd party plugins or instruments (to avoid unsynced osc etc) I can't imagine pan and level automation doesn't null.

you explained it all much better than me. thank you

pow -r is available separately to buy, but i dont know if it comes in vst format. i have to find the link for it. but i think it was around 300USD

Originally Posted by Leon EA

Also, I wonder what DAW's get different results with automation bounces. Without 3rd party plugins or instruments (to avoid unsynced osc etc) I can't imagine pan and level automation doesn't null.

well. this goes to protools.
on my humble basic test of RTD and BTD options for the final mix down. if you have automation they never null each-other.
not only RTD vs BTD, very pass of RTD produces different final mix. as well as every BTD produces different mix. and none of them nulls.

if you have no automation. there is no such problem.

very basic to test...
BTD your any protools session twice which includes some automation
and try null them. null value varies through out the song.
more the automation more the variation of NULL.
strange. but this is the case... when you record to disc (RTD)
using bus 31-32 etc... every thing is realtime. but the result is different every time.

i havent tried this with any other sequencer.

[drpenguen]

Originally Posted by White Falcon

F**K all these tests, Cubase 5 sounds the best, period. I'm serious!

e. i use cubase my self too. but i mix in protools as its mixer is more user friendly and has more analog GUI/

i am waiting for couple of bugs to be fxed. before i move to 5.
seems like everybody is so happy with 5

cubase rulls

[White Falcon]

Yeah, those new VST3s really blends with your audio better. I moved from SX3, and the plug-ins always sounded as if they were "on top" of the audio, now (Cubase 5) i get the smooth "blend" i get on my analogue console. With C5, i can finally challange my analogue mixes thumbsup All my freeware plugs integrates better (soundwise) to. And REVerence beats my Lexicons!

Best regards!

[drpenguen]

Originally Posted by White Falcon

And REVerence beats my Lexicons!

Best regards!

wow, is it really that good...

i have tried almost all vst reverbs ... nothing came close to lexicon90 so far...
there was no 3D on any of them...

[White Falcon]

Yes, i think so. But i put a de-esser before and light compression after the reverb and that makes it very smooth and dense. If you listen close, off course its no Bricasti, but still it's the best software reverb to date imho.

Best regards!

[laser]

Originally Posted by monomer

Well, since the dithering on 24-bits convertors is @ -136dBfs and the self noise of most convertors is above that, how would different dither make a significant impact on the sound?
And i mean after you have set up a proper gain structure where -136dBfs is not perceivable by the human ear...

If I can hear the difference with my ears--it's significant.

If it is important enough for Bob Olhsson to notice it and put in a 3rd party plug to compensate, it's significant.

He's the professional--you're the amateur. His ears are golden, yours are obvious lacking. Perhaps you should consider getting into a different line of work, like computer salesman?

By your logic, why dither at all?

I think I figured out why the crap I'm hearing on the radio doesn't measure up to those done 20, 30, 40 years ago.

Laser

[laser]

Originally Posted by Leon EA

(multiple dithering, which shoudl be avoided), or using truncated files.

With monomer's logic, it shouldn't matter. Before you even consider putting it on a 16-bit CD, you are converting it to 24-bit because you are monitoring.

Maybe the sound difference we hear is on our monitors and not on the finished product. This may be a more reasonable discussion.

Originally Posted by Leon EA

but since there is a big difference between dithering algorithms, it is surely a valid question to value the "sound excellence" of a DAW with its capability of loading the best dithering algorithms. .

We know that summing 22 channels on most DAWs null (remember, they did not all null on his tests). This is well accepted to those who know how meticulous Lynn Fuston is in his tests. But, this is not the discussion. The DAW incorporates much more than this. That is my point.


Laser

[bhodgson0739]

Originally Posted by nightscope

q: How many gearslutz members does it take to do a null test?

A: ??

Ns

hahahahahahahahahaha!!!

[bhodgson0739]

Jesus, people. Step aside and let a REAL pro tell you what the deal is...

If the gain staging in the dither is +1-(xy)^2 then the null of a noisy white neighbor will be +1x-2. This is OBVIOUSLY apparent when Cubase, Samplitude and Nuendo share converting algorithmic flortisheminangs and tells you that the square of the dither to 16bit means that 24bit 96k in all programs is 2x-34r^2 times the monitoring section was never NULLED to begin with. We can further test this by inserting a shitty EQ plug on the master bus, duplicating the file, reversing it, BUT also deleting both tracks. This should prove that you can't hear any sound on playback therefore meaning that OBVIOUSLY something happened.

[monomer]

Originally Posted by laser

He's the professional--you're the amateur. His ears are golden, yours are obvious lacking.

He's making money by telling people things they shuld believe.
I don't..

Perhaps you should consider getting into a different line of work, like computer salesman?

How do you know what line of work i'm in?
Maybe i am already a computer salesman?

By your logic, why dither at all?

In the situation of a converter with a higher selfnoise than its LSB, why dither indeed....

But if you think it will help you mix better then you shouldnt stop doing things your way.

[monomer]

Originally Posted by drpenguen

i did it last week

sonnox eq vs PSP vs URS etc


i used white noise as well as commercial music files...

Cool,.,.,. and.,,. Wow.... !!!

[obostic]

Originally Posted by Mirage

I just love you guys! Another heavy thread to discuss.... reminds me of converters or cables threads

However, one day i have decided to try several softs to "see" the differance using Motu MK828mk2 and Digi 002BLA. I haven't done the test with my Orpheus, cos i guess, i don't bother.

I have recorded Ac Gtr and also lined in my friend synth to see if there where any sound differance in software...

All friends (therey are just music lovers or musicians, no studio owners) i have shown the files for, have notice differance in sound. A friend who was with me while rec, notice the differance immediately!

I used Cubase SX2 (my first soft), PTLE7, Samplitude 10, Cakewalk Guitar Tracks Pro 3 (Worst soft playback i have ever heard!)

My thoughts are:
Cubase=Punchy, Compressed and coloured sound
PTLE=Cloloured with some PT signature; i guess
Samplitude=Transparent, detailed in all freq, most true to the source
Cakewalk Guitar Tracks Pro 3=is just a joke (the resolution was reminding me of casette tape)

I guess my filosofi is that I just want to hear my Ac Gtr from a monitor to sound so close as possible i hear in real world. Have no problem to learn new soft as far it is superior.

My friends and i have notice the differance even on non studio setup. I mean home PC/soundcard and some non pro speakers.

I must say that i'm shocked that there are many studio "engineers" here that don't hear the differance.

....please don't say that soft doesn't matter...



Good luck

Its quite clear that some people can't hear as well as others. It may be because of how they listen. Some people don't even know that hats will rims effect what you hear, turning your head to one side, etc.

What's really amazing it that this NULL test crap started with a Protools debate years ago about moving the faders. Now, here we are years later with the same crap about DAWs.

But, I personally doubt very serious if top engineers can't hear the difference between DAWs, they simply know how to compensate for any anomalies found in the DAW of their choosing. I'll say this you better believe what top engineers tell you publicly is miles different for what is done privately. When I read threads of this nature, you quickly assess what keeps the food chain divided.

[monomer]

Originally Posted by obostic

What's really amazing it that this NULL test crap started with a Protools debate years ago about moving the faders. Now, here we are years later with the same crap about DAWs.

I think that you don't understand what a null test in the digital domain actually means...
It is an absolute measurement of difference between two files.
If two files null completely it means there is no difference in result whatsoever. The samples are all exactly the same.
This is an amazing test when it comes to digital audio and it is very significant.
It realy does prove or disprove that there is a difference between two processes.

Furthermore, the resulting data will be exactly the difference between the two files.

[obostic]

Originally Posted by monomer

I think that you don't understand what a null test in the digital domain actually means...
It is an absolute measurement of difference between two files.
If two files null completely it means there is no difference in result whatsoever. The samples are all exactly the same.
This is an amazing test when it comes to digital audio and it is very significant.
It realy does prove or disprove that there is a difference between two processes.

Furthermore, the resulting data will be exactly the difference between the two files.

I know exactly what it means ABSOLUTELY NOTHING!

[joeq]

Originally Posted by laser

And also read from the person who has done the definitive null tests for DAWs:

The large differences Lynn was referring to were differences in the breaking of the null.. Who can hear >100 to >99? If you get up to answer the phone, and someone reaches over and moves one pan knob one tick, you are not going to come back and say "hey, who screwed up my mix?"

On the other hand, if you were performing a Null Test instead of a mix, your null would be totally broken - your test utterly ruined.

This speaks to the sensitivity of the null test itself. Not the audibility of one tick of the pan knob.

Even if dozens of such minute anomalies can accumulate into an audible difference, it doesn't mean one version will consistently end up sounding "punchier" or "wider" or "deeper" - the My DAW Sounds Best scenario that fills these threads.

Besides, such differences in the response of the controls, if they are audible, are audible to the engineer as he is mixing; they are not some nasty surprise that is sprung on him when the mix is played back. He is mixing INTO the DAW.

If he puts the pan knob at >99 and doesn't like it, he can put it back to >100.

To do a meaningful test, you need more data points. Four bands taking 15 frequency points in each band at every combination between the four bands (15^4) at several Q positions for each...

The possible Nulling of All Digital EQs is a side issue to the 'sound quality' of a DAW anyway. The 'better sounding' plug-in can simply be purchased and added to most any DAW. Many high quality dithers are also available for purchase cross-platform. (and a recent thread produced a number of users who prefer truncation to dither!)


If you can add the 'good' plugs and the 'good' dither to your 'bad' DAW, and the summing nulls, what is left that makes it sound "bad"?

This is audio, not the calculation of a bank account.

Actually, it is very much like the calculation of a bank account. If it wasn't, no nulls would ever occur. What incredible odds against getting a string of zeroes "by chance"! The differences are comparable to the way one bank might have a policy of rounding a half a penny up and another bank might round it down, except with money, more is always better.

Rounding in either direction is strictly speaking, an error. So which error "sounds" best? People want there to be magic and mystery in there somewhere, but computers are very deterministic. The magic and the mystery have to come from the person operating the DAW making the musical and sonic decisions.


For that one penny a year, I might as well go to the bank with the prettiest tellers.

[monomer]

Originally Posted by obostic

I know exactly what it means ABSOLUTELY NOTHING!

But can you explain why?

[monomer]

Originally Posted by joeq

On the other hand, if you were performing a Null Test instead of a mix, your null would be totally broken - your test utterly ruined.

This speaks to the sensitivity of the null test itself. Not the audibility of one tick of the pan knob.

There is just one situation where i would disagree.
In the case that you realy get NULL from the null test...
Then there realy is no difference between the subjects being tested and the null test would be meaningfull.
Hence the name of the test.
You test for the NULL condition.
Candidates either pass or fail the test.
No other conclusions should be drawn from this test directly besides deciding if there is a difference or not.