[RDD] More info on the Segue delay
small.net.nz at gmail.com
Mon Sep 28 06:38:36 EDT 2009
That's the only thing I've been waiting patiently for with RD too, the gaps in audio starting to disappear. I think if I remember correctly it required a re-work of the core audio engine in RD and was roughly mentioned being something for version 2.0? Or something to do with the way the TCP/IP between RDAirPlay and CAED or something handled acknowledgements for everything? It'll be in the archives I guess. I haven't used RD in a quite a while now (I went back to Windows, no gaps with segues and wave's even on 400Mhz machines) but would certainly would throw RD back on once that's sorted.
----- Original Message -----
To: rivendell-dev at lists.rivendellaudio.org
Sent: Monday, September 28, 2009 6:24 PM
Subject: [RDD] More info on the Segue delay
I know your working on a fix to eliminate the segue delay gap in RD.! I'd mentioned that a good way to demonstrate the gap was to have a five second long 400Hz tone cart (with CUT START at beginning of audio, SEGUE START at end of audio), enter 10 in a row in the RDAirplay main log, and you can hear the gap in the tone on playback.
I've since noticed that the delay seems larger still when RD is busy starting multiple events. Not sure if your aware of this, and whether it's related. For example, take a music cart and move the SEGUE START marker to the same position as CUT START, place 5 copies of the cart in the main log and start the first item. You can not only hear multiple copies of the same song running out of sync, you can see it on the time elapsed counter of the first 4 carts, with the first cart being up to a couple hundred ms different position to the last cart.
A similar thing happen if you put the same cart once in a log with a hard timed start, and load the same log into RD Main, Aux1 & Aux2. When it hits the hard time marker, all three logs are not playing the cart in perfect sync.
RD 1.5.1, Pentium 4 2.8 GHz, Opensuse 10.2, Alsa, audio files on the same pc, all outputs set to same card/port.
I'm not sure what your RD fix involves, but I can vouch for the concept of pre-cuing (loading and pausing) each file in the background ~one second before the end of the previous item, and instructing the application to start playing the cart at a set system time; 05:24:31.235658 etc. I've still got a couple of really old slow pc here; one is a Pentium 233 MHz, Win2k, 96mb ram, old ISA sound card, taking audio from a file server via a 10 mbit/s network interface card. The automation software running on it (which background pre-loads and cues) plays files perfectly. No gaps between sequenced tones, and 4 songs (the most it can play at once) in sync when moving segue start markers to 0ms.
Got to say it's not exactly a fast pc to load up the operating system, the application and a log though. Takes forever... Anyway, trying both the above tests on an ancient pc would be a fair test for your RD fix. Hope this helps. Thanks for your effort to improve this.
Rivendell-dev mailing list
Rivendell-dev at lists.rivendellaudio.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Rivendell-dev