Multichannel software digital crossovers?

[EchiDna]

I've read the pc as a source threads and I feel that my ultimate aim would be to integrate a software based crossover into a "pc as a source" arrangement... for multichannel material.

I know I can use something like bruteFIR for crossovers for 3-way stereo, but what about 3-way speakers x 5, plus sub?
DRC would also be nice!

AFAIK there is nothing that can do this in either hardware or software right now, but you never know I could be wrong. Any/all input greatly appreciated!

TIA,

 

[MWP]

Yup, BruteFIR can do that.
All you need is a soundcard with enough outputs.

 

[EchiDna]

thanks MWP,

I have not played around with BruteFIR much, more with AsioXO...
do you have any reason for prefering Brute - other than it being open source?

I guess if I can figure out how to use a few m-audio revolutions (for starters anyway) for this purpose I'm on my way ;-)

cheers,

 

[MWP]

I have not played around with BruteFIR much, more with AsioXO...
do you have any reason for prefering Brute - other than it being open source?

Not really... i havnt used anything else.
For such a task as a dedicated audio processing box without a interface (montior, keyboard, etc), Linux is much better suited than windows.

I guess if I can figure out how to use a few m-audio revolutions (for starters anyway) for this purpose I'm on my way ;-)

You might have problems using Revo's (just as im finding out now) as they dont have a master clock sync option without lots of hardware hacking.
Other Envy24HT cards that have SPDIF in could manage it though.

 

[dwk123]

I know I can use something like bruteFIR for crossovers for 3-way stereo, but what about 3-way speakers x 5, plus sub?
DRC would also be nice!

AFAIK there is nothing that can do this in either hardware or software right now, but you never know I could be wrong. Any/all input greatly appreciated!

TIA,

Why not? Aside from the processing latency which might make HT integration a bit of a pain, BruteFir can do this right now with the appropriate hardware. An RME card with 2xADAT converters will give you 16 channels, and there's no reason to believe this won't work - Anders Torger is running Ambisonics with more than this many channels. You'd probably need a big-ish machine and have to run high-ish partition sizes to get it to work, but you could do it. Even 2xDelta 1010 or a single Lynx AES-16 (if/when the OSS drivers support it) will give you 16 channels.

If/when a decent multichannel convolution engine becomes available for Windows, the EMU 1820 or 1820M are good looking cards. 8 analog outs, and either 2x spdif + adat, or else 3xspdif for either 14 or 20 channels in a single interface, and the analog outs are *very* good for the price.

IMHO the biggest problem with this is cost - for good converters and amps, you're usually talking $200-$300/channel at least, and that assumes you're using digital volume control - I'm not even aware of a solution that will do 15+ channels in the analog domain. Also, don't forget the cables - even cheap-ish stuff adds up quick when you start looking at long runs for this many drivers. IMHO it would be questionable to put the type of effort this would require into the project and then use marginal gear - doing this to try to make crap gear sound better is a bad bad idea.
By the time you throw in the speaker drivers, the PC and the soundcard(s) it looks like $6k at a bare minimum, $7k-$10k is easily possible. To commit that to a system when it's still a bit unclear as to how good it will be/can be (since the 'early adopters' are still working through the stereo case) is risky to say the least.

 

[dwk123]

thanks MWP,

I have not played around with BruteFIR much, more with AsioXO...
do you have any reason for prefering Brute - other than it being open source?

Off the top of my head
a) source code is available
b) FIR filters (although DRC renders this point somewhat moot)
c) multichannel (at least I don't think AsioXO can handle multiple input channels)
d) Linux - here I agree with MWP that this system should be an appliance that sits there and does it's job without an interface. Much easier to do with Linux.

I guess if I can figure out how to use a few m-audio revolutions (for starters anyway) for this purpose I'm on my way ;-)

cheers,

As MWP points out - multiple revos won't work. Dont even bother trying it- the lack of clock sync and the hassle of getting Jack or BruteFIR to talk to multiple cards will doom you to failure. If you want to run 15+ channels and are *serious* about it (not just a lark) then pony up the bucks for a real card.

I'd *STRONGLY * *STRONGLY* suggest forgetting multichannel to start with and just get a stereo rig up and running. Once you have that down and are comfortable/confident in the approach, extending it to 5.1 won't be 'all that difficult' - mostly choosing the hardware and doing a LOT of measuring.

 

[EchiDna]

Thanks for the input guys, I'll go stereo first as suggested

if the pc is just going to sit there as a pass-through device running linux, it does seem to defeat one of it's main purposes for me, that is, working as a source that can also do crossovers and DRC. I also want to be able to play stuff via direct input via keyboard and monitor (or wireless PDA as in asioXO)....

now am I asking too much of Brute and it's family of plugins?

DWK123, your post has given me a bit of a reality check, so I'll aim for functionality in 3 way x two channel first, then work up to the rest!

I can just see a single piece of commercial software to do this coming out not too far down the track, because it really seems to be not too different from something a sound engineer could run for a small club...