HDX cards, AAX plugins and voices

[John_Toolbox]

Yeah, there are 50 jillion posts about HD10. I'm not sure if this has been answered in any of them, but I'm curious about something:

How will HDX cards handle voices when using an AAX plugin that is hosted on the host CPU vs. the DSP on the HDX card? Will it reduce available voices the same way an RTAS plugin currently does with TDM systems? Essentially, with an HDX card, is the session going to be based around the architecture of the card, where as the card is a 256 voice mixer that will need to use up i/o to communicate with the host CPU?

[drumster]

I read a post from someone at AES who saw a demo of the new system and apparently there are some rules for burning voices in and out of the DSP card. In that way I guess it bears some similarities to TDM, but there are so many many more available voices now that it's probably a non issue.

[MixerGuy]

what is the definitive answer for this?

just curious

[Tobias Eichelberg]

Yes, it's the same way as with TDM systems. AAX Native plug-ins on aux channels take one voice per channel, putting HEAT to pre on audio tracks that have AAX Native plug-ins active will cost an extra voice per channel.

Toby

[M-ManLA]

It isn't the same. The FPGA chip takes the mixing & voices chores, and should use no extra DSPs for more voices. I believe an Avid rep cleared this up already. I'll have to find the post somewhere on here.

[Firechild]

Originally Posted by M-ManLA

It isn't the same. The FPGA chip takes the mixing & voices chores, and should use no extra DSPs for more voices. I believe an Avid rep cleared this up already. I'll have to find the post somewhere on here.

The question was not about DSP but voice count.

[Brandonx1]

HDX does use voices to move in and out of the dsp mixer for native plug ins. It's a little different then tdm tho. Tdm uses 2 voices per trip. One voice to get there and one voice to get back. HDX uses 1 voice per trip. So HDX is uses 1/2 the voices that tdm does. Another really important difference is HDX is designed to run at a 64 sample buffer. So latency when moving to out of the dsp mixer to the CPU and back is radically reduced. In short, HDX is realy comfortable using native plug ins on auxs! Great improvement over tdm. HDX really is a hybrid system of dsp and CPU power.

[NewdestinyX]

Originally Posted by Brandonx1

HDX does use voices to move in and out of the dsp mixer for native plug ins. It's a little different then tdm tho. Tdm uses 2 voices per trip. One voice to get there and one voice to get back. HDX uses 1 voice per trip. So HDX is uses 1/2 the voices that tdm does. Another really important difference is HDX is designed to run at a 64 sample buffer. So latency when moving to out of the dsp mixer to the CPU and back is radically reduced. In short, HDX is realy comfortable using native plug ins on auxs! Great improvement over tdm. HDX really is a hybrid system of dsp and CPU power.

Yes. A great improvement over TDM. That loss of two voices per RTAS plug used to really level us HD1 owners. Though HDX is not in my future anytime soon because ya really don't need anything more than native, even at the Pro studio level, it's still nice to know that Avid fixed that particular little logjam in their newest flagship HD system.

[JFreak]

"voice" is an i/o point of a mixer.

They just pulled that term out of their hats, because they think we're idiots. Apparently it's nicer to talk about "voices" than "mixer i/o points" -- and as long as you have 100% DSP mixer or 100% Native mixer, it does not matter.

But if you mix the two, you need to talk about mixer i/o points and understand the concept.

That said,

HDX mixer eats a voice every time signal goes in and out of the DSP.
TDM mixer eats a voice every time signal goes in and out of the DSP.

Period.

[JFreak]

Originally Posted by Brandonx1

HDX is designed to run at a 64 sample buffer.

So is HD, so what's the point?