The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Ableton - sound quality DAW Software
Old 26th August 2018
  #301
It's not a hidden option. It's clear in the Live set up window. It's the first thing I did when launching Ableton Live for the first time. Click 'off' fades on clips.
It doesn't actually change the sound in the way being discussed here, but it softens the attack of all your audio files, which of course is particularly bad for drums
1
Share
Old 26th August 2018
  #302
Lives for gear
Quote:
Originally Posted by chrisso View Post
It's not a hidden option. It's clear in the Live set up window. It's the first thing I did when launching Ableton Live for the first time. Click 'off' fades on clips.
It doesn't actually change the sound in the way being discussed here, but it softens the attack of all your audio files, which of course is particularly bad for drums
The default setting these days is off. It never hurts to check though.

We can also note that having fades off means you may have to keep an ear on transitions, like for example loop end to loop start, so you don't get clicks. That's one thing fades may alleviate.
Old 26th August 2018
  #303
Lives for gear
 

Quote:
Originally Posted by parricide View Post
this thread is still going???
have we all agreed that 1+1=2 in all DAWs yet?
1+1=2 in all DAWs when it comes to audio.
With MIDI it is not as an easy count as long as you prefer the MIDI quantize option over a real-time MIDI performance.

Which DAW offers a better representation of real-time MIDI would make an interesting thread I think.
Old 26th August 2018
  #304
Lives for gear
 
Acid Mitch's Avatar
 

Quote:
Originally Posted by lllubi View Post
1+1=2 in all DAWs when it comes to audio.
Over in the electronic music and production forum there is a current thread where people are stating that some daws and audio editors ( Soundforge and I think logic) do not display precise values. This means that you can change volume on a channel and then not get back to the same value you started with, due to the gui not giving the same values as what is happening internally.
So no, 1+1 is not always equal to 2 in all daws.
Old 26th August 2018
  #305
Lives for gear
 
stinkyfingers's Avatar
 

the summing is always the same for DAWS (1+1=2), it's the input that can be different (1+0.9999 != 2) due to rounding or truncation.
Old 26th August 2018
  #306
Lives for gear
 
Acid Mitch's Avatar
 

Quote:
Originally Posted by stinkyfingers View Post
the summing is always the same for DAWS (1+1=2), it's the input that can be different (1+0.9999 != 2) due to rounding or truncation.
But when the daw displays 1 instead of 0.9999 and you try adding 1+1 you are really getting 1.9999 and not 2.
Old 26th August 2018
  #307
Lives for gear
 

Quote:
Originally Posted by Acid Mitch View Post
Over in the electronic music and production forum there is a current thread where people are stating that some daws and audio editors ( Soundforge and I think logic) do not display precise values. This means that you can change volume on a channel and then not get back to the same value you started with, due to the gui not giving the same values as what is happening internally.
So no, 1+1 is not always equal to 2 in all daws.
In the electronic music forum, people are more sensitive to differences when controlling DAWs. Due to controllers and hardware pieces involved in their workflows and all the headache which sometimes comes with it.

The MIDI jitter thread there might help more users to understand why they perceive differences during their real-time MIDI performances than all the guesswork on 1+1= 2 here.
Old 26th August 2018
  #308
Lives for gear
 
stinkyfingers's Avatar
 

Quote:
Originally Posted by Acid Mitch View Post
But when the daw displays 1 instead of 0.9999 and you try adding 1+1 you are really getting 1.9999 and not 2.
yeh, that's what i just wrote...
Old 26th August 2018
  #309
Lives for gear
 
Acid Mitch's Avatar
 

I’m not talkng about MIDI or using controllers or anything outside the box. I’m talking about using the internal volume controls to change volume of recorded audio. Some daws and audio editors don’t always give you an accurate representation of what is happening to the audio which leads to rounding errors when doing simple volume changes.
Old 26th August 2018
  #310
Lives for gear
 
Acid Mitch's Avatar
 

Quote:
Originally Posted by stinkyfingers View Post
yeh, that's what i just wrote...
Yes but that means that sometimes 1+1= 1.9999 and not 2. I thought that you were saying that 1+1 is always 2. Sorry if I misunderatood.
Old 26th August 2018
  #311
Lives for gear
 
stinkyfingers's Avatar
 

Quote:
Originally Posted by Acid Mitch View Post
Yes but that means that sometimes 1+1= 1.9999 and not 2. I thought that you were saying that 1+1 is always 2. Sorry if I misunderatood.
ok, I think we are talking about the same thing, just a different angle. You are talking about real vs GUI values...I was just talking about the real values.
Old 26th August 2018
  #312
The problem is everything influences everything....
For example, you can't easily test midi performance between daws because the midi device you use highly influences the results - the controller, the interface, the computer.

Likewise, if it's suggested everyone tests the same mono or stereo audio file, then presumably the DAW where the original file was created is going to have an effect.
There is no 'control' in any of these tests.
You need to import an audio file that wasn't created in any digital audio program. I don't know how you would do that?
Old 26th August 2018
  #313
Lives for gear
 

Quote:
Originally Posted by Acid Mitch View Post
I’m not talkng about MIDI or using controllers or anything outside the box. I’m talking about using the internal volume controls to change volume of recorded audio. Some daws and audio editors don’t always give you an accurate representation of what is happening to the audio which leads to rounding errors when doing simple volume changes.
Couldn´t find the thread you are speaking about.

A DAW should not calculate another value if you type in a value of 1 except you detected a bug under certain conditions.
Old 26th August 2018
  #314
Lives for gear
Quote:
Originally Posted by chrisso View Post
if it's suggested everyone tests the same mono or stereo audio file, then presumably the DAW where the original file was created is going to have an effect.
Please elaborate.

Quote:
Originally Posted by chrisso View Post
There is no 'control' in any of these tests.
If you're testing SUMMING then the source files doesn't matter as long as they are valid and relevant in content. The DAW used for the RECORDING should have zero effect on the outcome of the test as all DAWs would sum the same sources.

Your implications here would seem to enter into the magical. Which would seem unexpected coming from you. My apologies if I missed the detail that would render your reasoning here understandable.

Last edited by Mikael B; 26th August 2018 at 07:39 PM..
2
Share
Old 27th August 2018
  #315
Lives for gear
 
stinkyfingers's Avatar
 

Quote:
Originally Posted by lllubi View Post
A DAW should not calculate another value if you type in a value of 1 except you detected a bug under certain conditions.
if you actually tested some DAWs, you would not be saying this. I mean, of course they should not change the value you give them, but they do.
I said in another thread, but I know Logics faders are rounding/truncating values, I’m sure there are others...

edit:ok, the curiosity was killing me so i downloaded the live10 demo and the fader values are not really what they say....i just dragged the fader, though, because i can't figure out how to enter a text value in this cluster**** of a daw.

edit again: ok, i just figured out how to enter text values and they are also incorrect.

attached a pic showing a "-3" dB gain reduction in live 10...the "L" shows the value when entering via text, the "R" shows the value when i click/drag the fader. neither of those are "-3 dBFS" (looks like 24 bit to me, one is truncating, the other is rounding). with Reaper's faders i get exactly "-3 dBFS" when i enter that via text (64 bit resolution).
Attached Thumbnails
Ableton - sound quality-screen-shot-2018-08-27-8.59.02-am.png  

Last edited by stinkyfingers; 27th August 2018 at 02:55 PM..
2
Share
Old 27th August 2018
  #316
Quote:
Originally Posted by Mikael B View Post
Please elaborate.
If DAWs change audio files, then presumably you can't discount they record audio in different ways. Or they process software instruments in different ways.
Any unbiased clinical test needs a 'control'.

If you record the original file in Ableton Live (for example), then export it, then find it doesn't null in Logic or Cubase, one could then question if the original recording in Live had had some influence on the playback in Logic and Cubase, or the method of exporting, or the software instrument used to create the audio file.
Likewise any DAW as a starting point.
Essentially there is no 'control' in this proposed test. Someone is starting the process by recording in A DAW of some kind, while trying to prove DAWS sound different.

And you don't need to add in the snidey personal commentary (thanks). If you disagree it's sufficient to say so
Old 27th August 2018
  #317
Lives for gear
Quote:
Originally Posted by chrisso View Post
If DAWs change audio files, then presumably you can't discount they record audio in different ways. Or they process software instruments in different ways.
Any unbiased clinical test needs a 'control'.

If you record the original file in Ableton Live (for example), then export it, then find it doesn't null in Logic or Cubase, one could then question if the original recording in Live had had some influence on the playback in Logic and Cubase, or the method of exporting, or the software instrument used to create the audio file.
Likewise any DAW as a starting point.
Essentially there is no 'control' in this proposed test. Someone is starting the process by recording in A DAW of some kind, while trying to prove DAWS sound different.
If you read my previous (incomplete but IMO relevant) suggestions on how to make a proper DAW comparison you'd see I suggested some time ago to record in another DAW than those tested.

This is not important though. You need to separate testing recording from testing mixing. This is also what I suggested to RichardNolan. If you have one set of files that is the basis of the control. If you do need one file to compare with all you need to do is to do the summing manually. The DAW developers have in most cases already done this as part of their development. The first summing mission is not to decide which DAW does summing "right", but if there are differences between them. When you test summing any valid and content relevant audio files of the required quality are good enough.

It's also possible to record into multiple DAWs from the same incoming audio streams and investigate any recording differences, but that's another test.

Surely you understand these things? Feel free to show where I go wrong here. Separation is how you do tests and can evaluate results. I thought that was obvious.

Quote:
Originally Posted by chrisso View Post
And you don't need to add in the snidey personal commentary (thanks). If you disagree it's sufficient to say so.
That was meant as a compliment. I truly didn't see that kind of thinking coming from you. I still would like to believe you are reasonable and intelligent. Please continue with logical reasoning and forego assumptions. I'll continue to attempt the same.
Old 27th August 2018
  #318
Lives for gear
Recording differences are of course another kettle than summing differences as are playback differences. While we can possibly expect some level deviations between DAWs recording the same incoming signal, do we really assume there are "magical bits" applied by the DAW in the storage stage? If so, that's distortion. Hardly what you want in a DAW, even though if you have processing on the incoming signal that's intended distortion. Which DAWs apply their own processing without user control in the recording stage?

The DAW AFAIU never affect what's happening in the ADC stage. I don't see how that would even work. What's comes in is an audio stream already. At least Live doesn't apply any processing to this signal AFAIU before it's stored to disk, except for Delay Compensation which I realize might affect things though it shouldn't. I suppose other DAWs could apply processing before storage. This would be important to know when evaluating. As with anything there is no knowledge before you test if this works as assumed.

Personally I really think this is an aspect I'm totally uninterested in. If I like what I hear when playing back a recording I made, what is there to care about?
Old 27th August 2018
  #319
Lives for gear
Quote:
Originally Posted by Acid Mitch View Post
I’m not talkng about MIDI or using controllers or anything outside the box. I’m talking about using the internal volume controls to change volume of recorded audio. Some daws and audio editors don’t always give you an accurate representation of what is happening to the audio which leads to rounding errors when doing simple volume changes.
Do you know any producers/mixers that can hear a 0.05 dB difference? Or however large those truncating differences might be.

This is still a good point though as it would prohibit null tests working out with different fader positions. It also touches on real differences, which is, for example, mixer possibilities beyond Pan Law. If you can't do routing "X" in DAW "B" then good luck attempting to make it sound the same as DAW "A" that can (if you use this routing obviously). In order to make fair comparisons you have to use shared functionality types.

As much as that would suggest DAWs have a sound, keep in mind comparing native functionality is in it's own category of tests. I think no-one would dispute different devices and functions can be, you know, different.
Old 27th August 2018
  #320
Lives for gear
 

Quote:
Originally Posted by stinkyfingers View Post
if you actually tested some DAWs, you would not be saying this. I mean, of course they should not change the value you give them, but they do.
I said in another thread, but I know Logics faders are rounding/truncating values, I’m sure there are others...

edit:ok, the curiosity was killing me so i downloaded the live10 demo and the fader values are not really what they say....i just dragged the fader, though, because i can't figure out how to enter a text value in this cluster**** of a daw.

edit again: ok, i just figured out how to enter text values and they are also incorrect.

attached a pic showing a "-3" dB gain reduction in live 10...the "L" shows the value when entering via text, the "R" shows the value when i click/drag the fader. neither of those are "-3 dBFS" (looks like 24 bit to me, one is truncating, the other is rounding). with Reaper's faders i get exactly "-3 dBFS" when i enter that via text (64 bit resolution).

Can you post a link to the thread you are speaking about?
Would like to learn about the testing process and the tools involved.

Sounds like an easy way to test MIDI controllers and plugins for their internal resolution.
Old 27th August 2018
  #321
I don’t have a horse in this race, but yes, I saw the suggestion to record in a daw then test in other daws.
I was just suggesting if daws supposefly alter audio, it makes sense to find a neutral way of generating the audio file being used for the test.
It’s also a shame one has to tread on egg shells to make any kind of critique of the discussion.
And when people say they are surprised you are thinking and they hope you are reasonable and intelligent......
Sigh.
Old 27th August 2018
  #322
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by chrisso View Post
I was just suggesting if daws supposefly alter audio, it makes sense to find a neutral way of generating the audio file being used for the test.
As long as the same files are used, it makes no difference where the files come from. Those files are now the reference and what counts is that they are identical in all DAWs tested.

Alistair
Old 27th August 2018
  #323
Lives for gear
 
stinkyfingers's Avatar
 

Quote:
Originally Posted by lllubi View Post
Can you post a link to the thread you are speaking about?
Would like to learn about the testing process and the tools involved.

Sounds like an easy way to test MIDI controllers and plugins for their internal resolution.
i don't know of any threads off hand where this has been tested, but it has been mentioned in this thread as well as here > Does repeated destructive editing of lossless file degrade the file?

i've only tested Logic and Reaper (and briefly Live now), but i basically use a (64 bit .wav) file of DC @ 0 dBFS. i use that or a 1 sample pulse. gain adjustments are done by entering text. then there's a "base line" to compare fader moves with if testing that, too.
i measure the actual levels in Reaper with a (JSFX) meter i made. it's 64 bit float precision code/calculation, so it's pretty accurate as far as meter go. (up to 15 decimals)
i don't know of any other meters/method of finding "exact" levels, maybe with a command line tool or something ? i know you can get some info from SoX but not sure how to use it for much other than basic "stats" or SRC.

edit: "Bitter" is good for checking pug-ins if you haven't used it. it can at least tell you if something is truncating down to 24 or 16 bit or whatever...

also forgot to note that sometimes i alternate the DC +/- to make a 1 Hz square wave...that way it tests both "1" and "-1". that's some trivial ****, though. (test if it's truncating or rounding)

Last edited by stinkyfingers; 27th August 2018 at 06:51 PM..
Old 27th August 2018
  #324
Lives for gear
Quote:
Originally Posted by chrisso View Post
I don’t have a horse in this race, but yes, I saw the suggestion to record in a daw then test in other daws.
Good!

Quote:
Originally Posted by chrisso View Post
I was just suggesting if daws supposefly alter audio, it makes sense to find a neutral way of generating the audio file being used for the test.
It’s also a shame one has to tread on egg shells to make any kind of critique of the discussion.
It's sufficient to be coherent and argue logically. As I see it, it's not coherent to have a quite long discussion on summing and then read suggestions that the audio file itself could possibly interact with the DAW and therefore cause differences. I didn't see anyone suggesting this. While I'd agree interaction could occur, the source files still would have to be viewed as being neutral.

On the other hand without testing there can be no knowledge and one can also see this suggestion of yours as way to turn every stone in order to uncover the truth. I'd have to agree that to investigate and question also the obvious can be very useful. My biggest breakthroughs come from such thinking.

Quote:
Originally Posted by chrisso View Post
And when people say they are surprised you are thinking and they hope you are reasonable and intelligent......
Sigh.
Now, look! It's not at all far fetched to see your suggestion about audio file interaction as a suggestion within the magical realm. As I held you as being reasonable and intelligent from previous posts I found that part shocking.
Perhaps you were just speculating and this wasn't actually your suggestion but your interpretation of other posts that you were giving us. I didn't see that really so was very surprised you would suggest something like this. There is some value in this suggestion though, no matter how you meant. You added something here that made me at least to question an assumption of mine.

I do have to maintain that this file interaction would have to be tested separately though. Ideally you test one process step, not multiple. Not unless you have to.
Old 27th August 2018
  #325
Gear Guru
 

Quote:
Originally Posted by chrisso View Post
If DAWs change audio files, then presumably you can't discount they record audio in different ways. Or they process software instruments in different ways. Any unbiased clinical test needs a 'control'.
I have read plenty of tests where a digital signal is split, fed to different DAWs and the signals nulled. Since every DAW can only record an already-digitized signal, the 'quality' of the signal is dependent only upon the A to D hardware. Which is not "part" of a DAW.

If they did not null, one of the DAWs would be broken. It would be receiving "11101" but recording }11100" But this has been tested and no DAW does this. In any case, while it might be possible for a DAW to "fail" at recording a signal accurately, it is hardly possible for such an inaccuracy to be consistently "harsh", consistently "mellow" , consistently "woody" as to render a "characteristic" sound to the software.

Do you think it is remotely possible for a DAW to consistently record "wrong" and then consistently sum "wrong" in such a way as to exactly cancel out the original errors in every single test?


This is just the difference-hearers grasping at straws, trying to find some excuse why the DAWs sound different to them. ANY 'reason' other than the only one that is true, which is that they made them sound different. And that they themselves know which DAW they are listening to when they listen. So let's see, unfair testing and unblinded listening, and a total disregard for the known science... who gives a rat's ass what such people state as "facts"?

Quote:
If you record the original file in Ableton Live (for example), then export it, then find it doesn't null in Logic or Cubase, one could then question if the original recording in Live had had some influence on the playback in Logic and Cubase, or the method of exporting, or the software instrument used to create the audio file.
except that nobody has "found" that. In the first place they all null anyway. In the second place "exporting" a file does not change even one bit of the data. If it did, you could not download a software program and install it on your computer because if a single "1" or "0" was out of place, the software would not work.

This reminds me of one of the threads where someone posted nulling files labeled "Pro Tools" "Logic" "Cubase" and "Digital Performer" and claimed they still sounded different. So another member downloaded the files, renamed them W,X,Y and Z and put the back up and challenged the guy to identify them. The guy could not and so he claimed that "The Internet" had changed his files. The files nulled before and after the uploads. The only "change" was disguising the names of the DAWs. Depriving the difference-hearers of their precious placebos.

Quote:
Essentially there is no 'control' in this proposed test. Someone is starting the process by recording in A DAW of some kind, while trying to prove DAWS sound different.
you are incorrect in your assumption that you are the first person to think of this and that nobody has ever tested the "record side". I recall seeing several tests along these lines.
Old 27th August 2018
  #326
Lives for gear
 

Quote:
Originally Posted by stinkyfingers View Post
i don't know of any threads off hand where this has been tested, but it has been mentioned in this thread as well as here > Does repeated destructive editing of lossless file degrade the file?

i've only tested Logic and Reaper (and briefly Live now), but i basically use a (64 bit .wav) file of DC @ 0 dBFS. i use that or a 1 sample pulse. gain adjustments are done by entering text. then there's a "base line" to compare fader moves with if testing that, too.
i measure the actual levels in Reaper with a (JSFX) meter i made. it's 64 bit float precision code/calculation, so it's pretty accurate as far as meter go. (up to 15 decimals)
i don't know of any other meters/method of finding "exact" levels, maybe with a command line tool or something ? i know you can get some info from SoX but not sure how to use it for much other than basic "stats" or SRC.

edit: "Bitter" is good for checking pug-ins if you haven't used it. it can at least tell you if something is truncating down to 24 or 16 bit or whatever...

also forgot to note that sometimes i alternate the DC +/- to make a 1 Hz square wave...that way it tests both "1" and "-1". that's some trivial ****, though. (test if it's truncating or rounding)
Thanks for your detailed response

Live 10 shows now intersample peaks when you zoom in on a sample. But it reflects it only on the master channel for readouts. A readout of 0 dBFs on an individual channel might already clip the master on either the left or right channel.

Can you read out Ableton´s master with your meter? Does it reflect your findings?

Ableton´s GUI doesn't reflect LEDSs readouts in PUSH in the upper range if I remember correctly. The first straight number I was able to dial in on both was around -10 dBFs. Would be weird if lower values would read out correctly.
Old 27th August 2018
  #327
Lives for gear
Quote:
Originally Posted by joeq View Post
I have read plenty of tests where a digital signal is split, fed to different DAWs and the signals nulled.


you are incorrect in your assumption that you are the first person to think of this and that nobody has ever tested the "record side". I recall seeing several tests along these lines.
It would be handy if there was a link list to previous tests to refer new willing testers to. Hopefully with project files and audio sources. Maybe someone already made and published such a collection?

There's nothing wrong redoing tests that have already been done as DAWs are updated. How interesting this is is really to up to each person. Generally I find challenging my assumptions to be interesting and a great way to learn about my tools and digital audio, but you can't do that more than now and then. Except for Pro Tools I don't have a great interest in other DAWs besides Live. I'm testing Reaper at the moment though, just for fun.
Old 27th August 2018
  #328
Lives for gear
 

Tried to null out Ableton´s internal routing.

Even volume fader figures from 0 to -10 against even send levels null out to infinity in Span. As well as -15 and - 20 and -30.

Some figures null at around -150 dBFs others around - 170 dBFs.

In Live 8 results have been around the hearing level from my memory.

Would be interesting to see how PUSH scales.
Old 27th August 2018
  #329
Lives for gear
Quote:
Originally Posted by stinkyfingers View Post
the summing is always the same for DAWS (1+1=2), it's the input that can be different (1+0.9999 != 2) due to rounding or truncation.
And this rounding or truncation can creep up on you. For example if the source DAW has a tempo with three decimals and not ending in zero, then how can a DAW with two decimals in the tempo play back at the exact same tempo? Possibly it could round off what's displayed only and maintain original tempo. But if you need to set the tempo exactly and you can't, then tempo will be shifted to what you can set and for that to work you must use some kind of tempo shifting processing and once that kicks in there's likely no perfect null of that signal.

Does this have to translate to bad sound? I think not, but noticeable or not the sound will likely be different to some small degree. Certainly not to the degree that one could make generalized sound quality statements about specific DAWs, but nevertheless.
Old 27th August 2018
  #330
Lives for gear
 
stinkyfingers's Avatar
 

Quote:
Originally Posted by Mikael B View Post
And this rounding or truncation can creep up on you. For example if the source DAW has a tempo with three decimals and not ending in zero, then how can a DAW with two decimals play back at the exact same tempo? Possibly it could round off what's displayed only and maintain original tempo. But if you need to set the tempo exactly and you can't, then tempo will be shifted to what you can set and for that to work you must use some kind of tempo shifting processing and once that kicks in there's likely no perfect null of that signal.
i thought all DAWs had to quantize the bpm to the sample rate...because you can't have any values that do not land on a sample (...obviously.)
any time domain edits or values or bpms, etc are quantized to the sample rate unless interpolated/oversampled.

edit: ok, I get what you’re saying now, and its a little different than what I was thinking at first.
I haven’t tested/thought about time resolution too much in this aspect. I would imagine it could cause a problem but have no idea how much...

Quote:
Originally Posted by lllubi View Post
Can you read out Ableton´s master with your meter? Does it reflect your findings?
i just tested a few levels quickly (Live > soundflower > Reaper) and it seems to be the same whether it's realtime stream or a rendered file. it's not a conclusive test, but i didn't notice anything unusual. i did notice the master output will clip @ 0 dBFS...ie...the stream is only 24 bit from Live.

Last edited by stinkyfingers; 28th August 2018 at 12:19 AM..
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump