Subject: Re: [linux-audio-dev] steinberg vst system link
From: Paul Davis (pbd_AT_Op.Net)
Date: Sat Feb 16 2002 - 16:10:19 EET
>Looks like it works by requiring a sound card with an ASIO driver,
>and using the digital audio in and out, somewhat as I had guessed.
>Then they can use up bit number 24 to send a bit of data with each
>sample. Or use up more bits of the audio data to send more networking
>data.
that much is obvious from the articles.
the question i have goes like this.
you've got 1 bit at 44.1->96kHz, providing total bandwidth of
43-93kB/sec. with that bandwidth, you've got to:
1) encode temporal position
2) transmit transport state changes
if its really going to work as well as they claim, the temporal
position has to be frame accurate, though presumably checking you are
in sync every so often will work OK. i can understand how you can
signify the passing of time with the 1 bit: if its set, time "ticked"
for that sample, if its not, then time was "stopped" for that
sample. this will let
the problem i see is that adding transport control messages
will destroy sample sync - you don't need to send start/stop messages,
because thats implicit in the "time-passing-bit", but i don't see how
you can reasonably locate and restart without problems. there has been
a significant discussion of this on the vst-plugins list (in the
context of regular VST), and we've also talked about it quite a bit on
jackit-devel. unless, of course, the "use more bits" option is
required for transport sync (i.e. more than one computer is running
transport-state-sensitive applications).
in addition, it appears to me that you still need up running word
clock to make sure the interfaces all see the same sample value at the
same time; yet steinberg don't mention this. this is either the usual
marketing BS, or they've figured some way around it.
so, i'm confused about exactly *what* they have done and *how* they
have done it.
--p
This archive was generated by hypermail 2b28 : Sat Feb 16 2002 - 16:04:44 EET