31 Oct
2002
31 Oct
'02
10:52 p.m.
linux-audio-dev'ers,
I am working in an Operations Center where we want to record 16-20 separate
analog audio voice nets/channels along with a few ethernet-based
display-data(not video) streams and play them all back individually,
selectively or all together for training purposes.
The audio nets are better than phone-quality, but not hi-fidelity and 22Khz
sampling should be sufficient. There are many turnkey voice-loggers and
pc-based "studio" applications but I need the audio recordings
closely-coupled with my own recordings of display-net data. I won't be doing
a lot of audio processing and would consider a higher rate system with
signal processing on board if it just works multi-channels well under Linux.
I suppose a multi-channel digital audio I/O board would be just fine as long
as I can get 16-20 analog channels converted up front. I am proposing
developing this application under Linux but need suggestions on the most
viable multi-channel audio in/out boards for use under Linux.
Any suggestions ?
Thanks !
Dave.
1 Nov
1 Nov
8:45 a.m.
linux-audio-dev'ers,
I am working in an Operations Center where we want to record 16-20 separate
analog audio voice nets/channels along with a few ethernet-based
display-data(not video) streams and play them all back individually,
selectively or all together for training purposes.
The audio nets are better than phone-quality, but not hi-fidelity and 22Khz
sampling should be sufficient. There are many turnkey voice-loggers and
pc-based "studio" applications but I need the audio recordings
closely-coupled with my own recordings of display-net data. I won't be doing
a lot of audio processing and would consider a higher rate system with
signal processing on board if it just works multi-channels well under Linux.
I suppose a multi-channel digital audio I/O board would be just fine as long
as I can get 16-20 analog channels converted up front. I am proposing
developing this application under Linux but need suggestions on the most
viable multi-channel audio in/out boards for use under Linux.
the obvious suggestion is the RME Hammerfall, which handles up to 24
channels of ADAT audio in and out. its well supported under Linux (i
wrote the driver). there are no boards/packages with analog i/o that i
know of that support more than 10 channels of analog i/o (the midiman
delta does 10), so you're better off buying external converters. the
problem you face here is that most multichannel converters (such as
the tango24, 8 in/8 out) are aimed at the audio market and are
overkill for what you're planning to do. you'd be paying about $700
per 8 channels. fostex makes a slightly cheap 8 channel converter, i
think.
--p
12:02 p.m.
On Fri, Nov 01, 2002 at 07:45:16AM -0500, Paul Davis wrote:
know of that support more than 10 channels of analog
i/o (the midiman
delta does 10),
just to clarify: the delta 1010 has 8 analog i/o.
The other 2 channels are S/PDIF.
The 1010 has word-clock sync so you *should* be able
to run 2 of them together for 16 analog I/O, or 3 for 24;
i know people do it regularly under Windows, but I don't know what's
involved in doing that on Linux. Probably some
black magic in .asoundrc.
But for the price of a pair of 1010s you could just as easily get
a Hammerfall and some cheapish converters.
Creamware's A16 Ultra might do the job, if I read the specs
right: looks like it has 16 channels of I/O, 2 ADAT connectors
(8 channels each), and price of US $1k.
Or a pair of Alesis AI3, which would definitely do the trick
and sell for US $400 each.
--PW
--
Paul Winkler
http://www.slinkp.com
"Welcome to Muppet Labs, where the future is made - today!"
1:02 p.m.
Paul Winkler wrote:
The 1010 has word-clock sync so you *should* be able
to run 2 of them together for 16 analog I/O, or 3 for 24;
i know people do it regularly under Windows, but I don't know what's
involved in doing that on Linux. Probably some
black magic in .asoundrc.
either that, or your alsa code opens both (all three, etc) and
links them like capture and playback on one card are linked. last
time this came up on alsa-devel it was claimed to work but no-
one seems to have done it yet. m-man itself seems to claim you
can link all of their ice1712-based cards this way.
paul d was a bit worried there may be a slight offset between
starting the cards (physically, they are started sequentially)
but that would apply to the windows drivers as well.
tim
1:18 p.m.
I think someone on one of the lists (ardour-dev, maybe) managed to do
this. There was a lot of messing around, however,
and much as I like M-Audio cards I wouldn't recommend this approach to
someone who just need a lot of analog I/O.
OTOH, doesn't the Hammerfall DSP have a multi-channel analog interface
as one of its options?
-dgm
Tim Goetze wrote:
Paul Winkler wrote:
The 1010 has word-clock sync so you *should* be
able
to run 2 of them together for 16 analog I/O, or 3 for 24;
i know people do it regularly under Windows, but I don't know what's
involved in doing that on Linux. Probably some
black magic in .asoundrc.
either that, or your alsa code opens both (all three, etc) and
links them like capture and playback on one card are linked. last
time this came up on alsa-devel it was claimed to work but no-
one seems to have done it yet. m-man itself seems to claim you
can link all of their ice1712-based cards this way.
paul d was a bit worried there may be a slight offset between
starting the cards (physically, they are started sequentially)
but that would apply to the windows drivers as well.
tim
1:58 p.m.
OTOH, doesn't the Hammerfall DSP have a
multi-channel analog interface
as one of its options?
you can get the interface card plus either:
multiface I/O: 8 analog in+out, plus 8 (or is it 16) digital in+out
digiface + external converters: up to 26 in+out
RME make external converters, as do several other companies. There
doesn't seem to much market interest in more than 8 channels per box,
though creamware and i think rme/steinberg now do sell such things. i
talked with frontier designs, who make the tango24, about doing this,
and they weren't interested.
the final related hardware, of course, are the new MOTU units, which
pack loads and loads of analog I/O into a 2U device, and then ship the
digital results in a format that nobody else can use :(
--p
2:43 p.m.
And the DigiFace/MultiFace are very nice products. It was my understanding
from one of RME's vendors that at least under ASIO both the DigiFace and
MultiFace support multiple breakout boxes with a single PCI or cardbus
adapter card. I was keeping this in the back of my mind for the day when I
can dump Pro Tools completely. Does this work under Linux or has anyone
tried?
These RME products use an RME proprietary interface that looks and smells
like 1394, uses 1394 cables, but isn't 1394, so you can't hang a 1394 audio
drive on it like I use on my systems. (I do not know if there is any danger
in making a mistake when you plug the wrong device into their cable...)
All of this seems pretty high end for what I thought the requestor was
originally looking for at 22KHz, but I know of no answer for that market
specifically. I think the AI-3 or one of the other devices similar to that
(8 I/O delivered over ADAT) tied to a Hammerfall is a probably a good
solution cost wise and would be well supported by Linux. Additionally it
would require no .asoundrc files and no word clock stuff, so it's pretty
straight forward. He can probably cut a deal on 3 AI-3's and get them for
$300-$350 each somewhere. With a $500 Hammerfall he'd be in for $1500 for 24
channels, or about $65/channel. Seems pretty reasonable for this
application.
I wonder if there's a business to be made in developing hardware for apps
like this? I'm looking for a good business idea like that.
With best regards,
Mark
-----Original Message-----
From: linux-audio-dev-admin(a)music.columbia.edu
[mailto:linux-audio-dev-admin@music.columbia.edu]On Behalf Of Paul Davis
Sent: Friday, November 01, 2002 9:58 AM
To: linux-audio-dev(a)music.columbia.edu
Subject: Re: [linux-audio-dev] Viable multi-channel audio I/O boards for
Linux ?
OTOH, doesn't the Hammerfall DSP have a
multi-channel analog interface
as one of its options?
you can get the interface card plus either:
multiface I/O: 8 analog in+out, plus 8 (or is it 16) digital in+out
digiface + external converters: up to 26 in+out
RME make external converters, as do several other companies. There
doesn't seem to much market interest in more than 8 channels per box,
though creamware and i think rme/steinberg now do sell such things. i
talked with frontier designs, who make the tango24, about doing this,
and they weren't interested.
the final related hardware, of course, are the new MOTU units, which
pack loads and loads of analog I/O into a 2U device, and then ship the
digital results in a format that nobody else can use :(
--p
2:57 p.m.
And the DigiFace/MultiFace are very nice products. It
was my understanding
from one of RME's vendors that at least under ASIO both the DigiFace and
MultiFace support multiple breakout boxes with a single PCI or cardbus
adapter card. I was keeping this in the back of my mind for the day when I
can dump Pro Tools completely. Does this work under Linux or has anyone
tried?
its simply false. there is a single 1394 connector on the back of the
PCI card and a single connector on the cardbus dongle, so you can only
connect a single breakout box to one interface unit. whoever told you
this either knows something that i don't or they are misinformed/confused.
this is the main area i am aware of where the MOTU units look more
appealing than the RME ones. SMPTE sync is the other one.
These RME products use an RME proprietary interface
that looks and smells
like 1394, uses 1394 cables, but isn't 1394, so you can't hang a 1394 audio
drive on it like I use on my systems. (I do not know if there is any danger
in making a mistake when you plug the wrong device into their cable...)
it is 1394. it just uses a proprietary mid-level protocol for the
communication. you don't have any access to the 1394
firmware/controller on the card, but if you did, you could write a
driver for it and use it as a general 1394 interface. alas, RME saw
the (commercial) danger of putting everything in the external box and
having you use a $50 1394 interface. as a result, we get this from
them and for the same reasons, the MOTU stuff from MOTU.
All of this seems pretty high end for what I thought
the requestor was
originally looking for at 22KHz, but I know of no answer for that market
specifically. I think the AI-3 or one of the other devices similar to that
that's true, i forgot the AI-3 units. they are a good deal for this
kind of thing.
I wonder if there's a business to be made in
developing hardware for apps
like this? I'm looking for a good business idea like that.
given the tiny size of the computer audio market as it is, it seems
unlikely to me. what i think would be a great success is a properly
priced 24 channel A/D-D/A converter. +10/-4 switchable balanced
inputs, ADAT I/O for the digital side or some 1394 standard once it
emerges as an option. that would be great, and would have an instant
home in many studios of many different sizes. 16 is enough for many
people though, and the creamware and rme solutions might have this all
sewn up.
--p
3:40 p.m.
Paul,
You are correct! I'm sorry. I went back and looked at the email I
received. He actually said you could buy two PCI cards and link them under
ASIO. My mistake. I'm was not at all attracted to that solution for the
obvious reasons of cost, syncing, heat and interrupts on more cards. Yuck.
However, if you're correct and RME is just moving standard 1394 packets,
then even with their own protocol you could just use a standard 1394 hub
between a single adapter card and two breakout boxes. That could work and
would only run about $50. It would require that RME be able to keep track of
multiple boards on a single 1394 bus, but that's no big deal really. I
should ask Matthias at RME. Worth looking into for this application for
sure. (P.S. - I do 1394 for a living...)
I'll have to look in my notes about the competitors to the AI-3. There
are a couple now that are lower priced I think. RME supports 96K across
pairs of ADAT lines. I wonder if they support 16 22K channels across a
single channel? ;-)
Mark
-----Original Message-----
From: linux-audio-dev-admin(a)music.columbia.edu
[mailto:linux-audio-dev-admin@music.columbia.edu]On Behalf Of Paul Davis
Sent: Friday, November 01, 2002 10:57 AM
To: linux-audio-dev(a)music.columbia.edu
Subject: Re: [linux-audio-dev] Viable multi-channel audio I/O boards for
Linux ?
And the DigiFace/MultiFace are very nice products. It
was my understanding
from one of RME's vendors that at least under ASIO both the DigiFace and
MultiFace support multiple breakout boxes with a single PCI or cardbus
adapter card. I was keeping this in the back of my mind for the day when I
can dump Pro Tools completely. Does this work under Linux or has anyone
tried?
its simply false. there is a single 1394 connector on the back of the
PCI card and a single connector on the cardbus dongle, so you can only
connect a single breakout box to one interface unit. whoever told you
this either knows something that i don't or they are misinformed/confused.
this is the main area i am aware of where the MOTU units look more
appealing than the RME ones. SMPTE sync is the other one.
These RME products use an RME proprietary interface
that looks and smells
like 1394, uses 1394 cables, but isn't 1394, so you can't hang a 1394 audio
drive on it like I use on my systems. (I do not know if there is any danger
in making a mistake when you plug the wrong device into their cable...)
it is 1394. it just uses a proprietary mid-level protocol for the
communication. you don't have any access to the 1394
firmware/controller on the card, but if you did, you could write a
driver for it and use it as a general 1394 interface. alas, RME saw
the (commercial) danger of putting everything in the external box and
having you use a $50 1394 interface. as a result, we get this from
them and for the same reasons, the MOTU stuff from MOTU.
All of this seems pretty high end for what I thought
the requestor was
originally looking for at 22KHz, but I know of no answer for that market
specifically. I think the AI-3 or one of the other devices similar to that
that's true, i forgot the AI-3 units. they are a good deal for this
kind of thing.
I wonder if there's a business to be made in
developing hardware for apps
like this? I'm looking for a good business idea like that.
given the tiny size of the computer audio market as it is, it seems
unlikely to me. what i think would be a great success is a properly
priced 24 channel A/D-D/A converter. +10/-4 switchable balanced
inputs, ADAT I/O for the digital side or some 1394 standard once it
emerges as an option. that would be great, and would have an instant
home in many studios of many different sizes. 16 is enough for many
people though, and the creamware and rme solutions might have this all
sewn up.
--p
8077
days inactive
8078
days old
linux-audio-dev@lists.linuxaudio.org
8 comments
6 participants
participants (6)
-
David Gerard Matthews -
Greene, David L. -
Mark Knecht -
Paul Davis -
Paul Winkler -
Tim Goetze