7.0.12b2: Unprecise Tempo when running as MIDI clock slave

UHE is now closed. For Technical Support from Ableton, please go here: http://www.ableton.com/support
Locked
Crash
Posts: 805
Joined: Fri Sep 19, 2008 4:06 pm
Location: Nowhereland

7.0.12b2: Unprecise Tempo when running as MIDI clock slave

Post by Crash » Mon Nov 03, 2008 9:58 am

Alex wrote: Bug fixes:
...

When running as MIDI clock slave, Live would only display a rounded to the nearest integer tempo.
Fixed:

When running as MIDI clock slave, Live would only forward a rounded to the nearest integer tempo to external plugins.

When running as MIDI clock slave, Live would only play a rounded to the nearest integer tempo.


Not fixed:

When running as MIDI clock slave, Live would introduce changing MIDI clock Sync delay depending on the chosen tempo. The slower the tempo the higher the Sync delay.


Non Improvements:

When running as MIDI clock slave, Live would introduce very imprecise tempo calculations depending on the chosen tempo. The higher the tempo the worse the tempo changes.

Because of the considerable tempo changes that happen after the rounded integer playback fix audible phasing sounds can happen when two instances play the same audio clip. Before the rounded integer fix that phasing sound only happened at non integer Master tempi (like 110.5 BPM), now it happens all the time. All in all Live's tempo stability as a Slave is worse now than before because the rounding bug at least helped against that.
Last edited by Crash on Mon Nov 03, 2008 12:58 pm, edited 2 times in total.

Crash
Posts: 805
Joined: Fri Sep 19, 2008 4:06 pm
Location: Nowhereland

Post by Crash » Mon Nov 03, 2008 11:22 am

Here are some animated screenshots that demonstate just how really really bad Live's current MIDI Slave tempo calculation is compared to Kore 2 running on the very same system synced to the very same MIDI Master.

Both the Kore 2 Standalone and the Live instance run Kontakt 3.

111 BPM

While Kore 2's tempo stays withtin 0.05 (!) BPM of the Master, Live's tempo varries by upto 1.8 BPM.

Image


333 BPM

While Kore 2's tempo stays withtin 1 BPM of the Master, Live's tempo varries by over 15 (!) BPM.

Image

Also take notice of how Live's own Tempo display is often way off the tempo displayed by the Kontakt plugin that is running inside Live.
Last edited by Crash on Mon Nov 03, 2008 6:12 pm, edited 1 time in total.

Alex
Posts: 4006
Joined: Wed Apr 03, 2002 1:07 pm
Location: Ableton Headquarter

Post by Alex » Mon Nov 03, 2008 11:34 am

Hi Crash,

I moved your postings into that separated thread.

Could you let me some more infos about the setup.

1) Are you on Mac OS X or on Windows XP or Vista?
2) Is the MIDI clock master running on the same machine and your are using a virtual MIDI driver to MIDI-connect Live?

Regards
/Alex

Crash
Posts: 805
Joined: Fri Sep 19, 2008 4:06 pm
Location: Nowhereland

Post by Crash » Mon Nov 03, 2008 11:55 am

Hey Alex,

thanks for moving the post. I wasn't sure wether "Beta" reports should get their own threads or just reply to the announcement thread. I will leave the announcement thread alone in the future. :oops:

I just send an Email to beta@ableton with precise informations about my system. I'm on XP at the moment (when will Live's missing Vista compatibility be fixed?) and all MIDI Slaves were run by the very same MIDI Master, a Live 7.0.9 instance running on the very same PC. Both Kore and an instance of MidiOX proved the Master to be stable.

It doesn't matter wether virtual ports are used (MidiYoke, LoopBe) or physical ports (Kore 1, FF400). I reported that already in another thread that Dom "answered" to (as far as calling my reports fiction qualifies as anwering).

I can give you rendered audio examples that prove my reports about both the rounded integer playback of 7.0.10(/11b/12b1?) and the varrying tempo of 7.0.12b2.

dom
Posts: 936
Joined: Fri Jan 07, 2005 9:24 am
Location: Ableton Headquarters
Contact:

Re: 7.0.12b2: Unprecise Tempo when running as MIDI clock slave

Post by dom » Tue Nov 04, 2008 2:16 pm

Crash wrote: Fixed:

When running as MIDI clock slave, Live would only forward a rounded to the nearest integer tempo to external plugins.

When running as MIDI clock slave, Live would only play a rounded to the nearest integer tempo.
The first statement is right, just as Alex wrote in the announcement.

The second, however, was made up by yourself.
As i already mentioned in another thread: Live does not and was never playing using a rounded tempo.
It only was displaying a rounded tempo and was communicating this rounded tempo to plugins - the latter was obviously a bug and was fixed now.

As i already asked you several times: Please, with sugar on top, stop confusing people with cooked up statements.

This time you reached a new level by even making up bugfix statements that Alex never wrote.
That's just cheeky, bad for other people and even damaging your own credibility.
Crash wrote: Not fixed:

When running as MIDI clock slave, Live would introduce changing MIDI clock Sync delay depending on the chosen tempo. The slower the tempo the higher the Sync delay.




Non Improvements:

When running as MIDI clock slave, Live would introduce very imprecise tempo calculations depending on the chosen tempo. The higher the tempo the worse the tempo changes.
Can't confirm or deny this yet, but we will investigate.
Crash wrote: Because of the considerable tempo changes that happen after the rounded integer playback fix audible phasing sounds can happen when two instances play the same audio clip. Before the rounded integer fix that phasing sound only happened at non integer Master tempi (like 110.5 BPM), now it happens all the time. All in all Live's tempo stability as a Slave is worse now than before because the rounding bug at least helped against that.
It makes no sense to try judging midi clock sync using phase tests and we will not investigate this further.

Best,
Dom
ableton support team
support@ableton.com

Crash
Posts: 805
Joined: Fri Sep 19, 2008 4:06 pm
Location: Nowhereland

Re: 7.0.12b2: Unprecise Tempo when running as MIDI clock slave

Post by Crash » Tue Nov 04, 2008 6:09 pm

dom wrote:The first statement is right, just as Alex wrote in the announcement.
No, he did not. He just wrote that "display" was fixed, which in contrast to the real fix is a minor cosmetic one. I added the more important unannounced fixes that are related to the mentioned fix.
The second, however, was made up by yourself.
No, it is not, you are just too ignorant and stubborn to check the evidence and your own reality. It doesn't matter anymore, because the latest Beta shows a whole different playback behavior now.
As i already mentioned in another thread: Live does not and was never playing using a rounded tempo.
Whatever it does, the output of audio clips playing at non-integer slave tempi sounds very obviously different to integer tempi upto 7.0.12b1. With 12b2 the playback behavior changed! This is not a simple change of Live's tempo display. I still can prove my reports by audio examples.

You are either lying and discrediting me or lacking knowledge of what your own application is doing while ignoring the possibility and evidence of potential errors that are presented to you!
It only was displaying a rounded tempo and was communicating this rounded tempo to plugins - the latter was obviously a bug and was fixed now.
For such an "obvious" bug and your repeated bragging about how detailed your inhouse testing of the Midi engine was I wonder why you never noticed it yourself and had me do your work again?
As i already asked you several times: Please, with sugar on top, stop confusing people with cooked up statements.
No need for sugar, you keep thinking your software is running by the "specs" and ignore any evidence for potential errors as long as it is not at least blantantly easy and obvious to understand.
This time you reached a new level by even making up bugfix statements that Alex never wrote.
That's just cheeky, bad for other people and even damaging your own credibility.
Excuse me? I quoted Alex' announcement under his name and added my own findings under my own name. Alex moved the post into its own thread and even wrote me a PM about it. He did not mention anything about me making up statements in his name.

Like before you are beginning to spam my content and topic related thread with accusations. Please stay on topic. If you think I am wrong then argue on the basis of the material I provide (screenshots, rendered audio samples, Live sets, videos) or provide your own material.
Can't confirm or deny this yet, but we will investigate.
Nice, last time you were too busy with calling me things and ignored that part of my report completely. I guess alot of Ableton Live users will be very grateful if you finally fix Live's Midi Sync Slave behavior.
It makes no sense to try judging midi clock sync using phase tests and we will not investigate this further.
Could please someone with more sense take over here? Not only did I provide animated screenshots that prove Live 12b2's jumpy tempo response visually while also offering audio examples, but claiming that phase tests are senseless when trying to judge tempo(-stability) related issues is such a #!§$& statement that I can only repeat publically what I wrote in an email before: Shame on you, Ableton! :oops: :?

Besides that the aim of the tests was not phasing, but timely playback. The phasing happens on its own whenever Live behaves erratic. It is just one way to make the problems audible in order to understand them. But now you are gonna tell me that audio producers, musicians and engineers don't care for unexpected audible phasing sound!? Sorry, no bonus! :arrow:

dom
Posts: 936
Joined: Fri Jan 07, 2005 9:24 am
Location: Ableton Headquarters
Contact:

Post by dom » Tue Nov 04, 2008 9:31 pm

No offense, but a little knowledge is a dangerous thing... just because you think something is right, solely based on your observations does not make it right at all, regardless how often you argue about it.
Remeber the words "earth" and "flat"? Those guys were as sure in their observations as you are. But obviously this does not automatically mean they were right.

You're only guessing what's going on under the hood and base all your speculations on pure observations - but at the end of the day it sadly turns out that often you really don't have a clue about the technical reasons at all - only about the results you have found and some "dangerous half-knowledge" as the germans call it. That's why the majority of your ubergeek posts are edited multiple times after some days. Not only in our forum.


The only surprise for me and again reason for this reply:

It's amazing that you still don't seem to notice that in your case i'm absolutely not willing to discuss technical statements at all, do you?
Trust me, it's utterly fruitless. I only chime in to correct false statements, make my comments and that's it. Take it or leave it.

This forum is not about you or general justice, it is about our product and how we decide it to be.

You're invited to come up with your own software if you think you can do it better - that's actually not a bad thing and exactly what some guys over here did a few years ago!
But as long as you decide to stick to our product please be so kind and don't try to tell us how to do our job. Again, no offense, but it's just not your business and nobody forces you into using Live.

In short: We're happy about every observaton we can get as we're known for our motivation to improve Live as much as possible, but it's not your your job to come up with assumptions regarding the reasons for what you observe as wrong or buggy behaviour. Not satisfied with that? Go back two paragraphs and start again.



Thanks,
Dom
ableton support team
support@ableton.com

Locked