Strange hiccup after 25 seconds of recording

[kidbeyond]

Kid Beyond here, with a strange little bug.

My setup: Live 5.0.2; Mac OSX 10.4.3; 1.67 GHz PowerBook G4 w/2 GB RAM; MOTU Traveler.

I was working with a 79-BPM song in which I record an 8-bar loop, at 1/4-note-level quantization. But the loop kept coming out either longer or shorter than 8 bars -- 7.3, 8.1, never exactly 8 bars. This is an ALS I've been working with since Live 3, and I'd never had this problem in the song.

Looking into it more, I noticed the strangest thing: when recording a long loop, Live seems to "hiccup" -- or temporarily freeze up -- at 24.75 seconds or so. I took a little video capture of the Arrangement Position field to show you what I mean:

http://kidbeyond.com/livehiccup.mov

You'll see the numbers chugging along regularly -- until 8.4.2, when there's this odd pause. That's the hiccup.

Let's see if any of y'all can reproduce this:

1) Open up a New Live Set.
2) Set the tempo to 79 BPM.
3) Arm Track 1 for recording.
4) Click the record button on any slot in Track 1.
5) Watch the Arrangement Position field for about 25 seconds;
you should see the hiccup at 8.4.2 or so.

At different BPMs, the hiccup happens after a different number of bars, but always at 24.75 seconds or so. It just so happens that at 79 BPM, that's around 8 bars. At 120 BPM, the hiccup happens after 12 bars (13.1.1 or so). At 45 BPM, it happens after 4.5 bars (5.3.1 or so).

I tried this on both my MOTU Traveler and the PowerBook's built-in audio. I tried jacking up the Buffer Size way high on both. I tried changing the Global Quantization, the Launch Quantization, the Sample Rate, and so on. The hiccup does not go away.

The hiccup does not seem to happen at any time after that 25-second mark. It passes through 50, 75, 100 seconds of recording just fine.

Any thoughts? Anyone else seeing this besides me?

Sending this to the Abe support email as well...


Thanks in advance,

[gomi]

Kid Beyond here, with a strange little bug.

My setup: Live 5.0.2; Mac OSX 10.4.3; 1.67 GHz PowerBook G4 w/2 GB RAM; MOTU Traveler.

I can confirm this with my Mac Mini 1.42 + edirol fa-66 and my 1.5 powebook.

this is repeatable everytime.

[kidbeyond]

I got a response from Ableton's Don Wilms, with a
few questions. I sent him the following:

It seems that audio is affected at the same time
as the hiccup -- specifically, the ability to quantize
correctly. The little freeze-up happens, and by the
time it unfreezes, we're already into the next bar.

Here are two movies to illustrate this:

http://kidbeyond.com/abetests/live-78bp ... hiccup.mov
http://kidbeyond.com/abetests/live-78bp ... hiccup.mov

Here's what you can do to (hopefully) reproduce
them:

=-=-=-=-=-=-=-=-=-=

78 BPM -- 4 BARS -- NO HICCUP

1) Open up a New Live Set.
2) Set the tempo to 78 BPM.
3) Set Global Quantization to 1/4.
4) Arm Track 1 for recording.
5) Click the record button on any slot in Track 1.
6) Watch Track 1's Status Display...
7) Right after it gets to 3.3.x, hit the Clip Stop Button.

Quantization happens correctly; you get a nice loop
of length 4.0.0 (as you can see in the bottom right
of the movie).

=-=-=-=-=-=-=-=-=-=

78 BPM -- 8 BARS -- HICCUP

1) Open up a New Live Set.
2) Set the tempo to 78 BPM.
3) Set Global Quantization to 1/4.
4) Arm Track 1 for recording.
5) Click the record button on any slot in Track 1.
6) Watch Track 1's Status Display...
7) Right after it gets to 7.3.x, hit the Clip Stop Button.

The hiccup occurs. By the time Live unfreezes, the
Track Status Display is at 8.0.x, and we get an
"uneven" loop of length 8.1.0 (see bottom right again).

=-=-=-=-=-=-=-=-=-=

In Prefs / Default / Launch, Quantization is set to
Global.

As I said in my previous email, it's not 8-bar loops
that seem to be the problem, it's that the hiccup
always occurs at around 24.75 seconds of recording --
which, at 78 BPM, is around 8 bars.

Hope this helps...