AAX Native voice usage on PT 11.1.3 HDX

[jafbox]

i've noticed that AAX native plug ins don't seem to take up voices on an HDX system like i previously understood. i previously thought that one AAX Native plug in would take up one voice, but on my system this is not the case. i can add numerous AAX native plug ins without changing the voice count! i can then save the session, power cycle and the voice count remains the same..

is this a bug? I'm certainly not complaining either way :)

PT11.1.3 HDX1
2014 12 core mac pro os 10.9.2

[Drew Mazurek]

I think it depends on where they are. If they're on the first Insert they likely don't use a voice. But what if you put them on an Aux Input? What if there's a DSP plug in front of it?

I'm not an HDX user so I have no direct knowledge, sorry.

I'm sure someone that does will chime in.

[Brandonx1]

Avid made a pm improvement with regard to voice stealing in 11.1.3 and up. You get 64 free native jumps now. So you will eventually have voice stealing issues but avid greatly I proven the problem. Big thank you to avid

[jafbox]

Wow, this is fantastic news!! I'm surprised it wasn't in any of the docs or spoken about. Huge bonus for my workflow :)

[Frank Kruse]

Originally Posted by Brandonx1

Avid made a pm improvement with regard to voice stealing in 11.1.3 and up. You get 64 free native jumps now. So you will eventually have voice stealing issues but avid greatly I proven the problem. Big thank you to avid

Wow, that's interesting info. So now we can actually play big sessions that run fine under HDn also on HDX1? At least it's more likely they will play.

Ironically that often wasn't possible and you needed an HDX2 to playback big HDn sessions.

Thanks for the info!

Frank.

[loud neighbor]

Found this thread doing a search on HDX voice count and wanted to reopen this discussion for my own enlightenment.

I have no experience with HDN - I've got an HDX rig - which was HDX1 until earlier this month when I had to purchase a second HDX card to work on a huge session. It bothers me that I will NEVER use the second card for any sort of DSP (mixer / plugins) but AVID requires it to access another 256 voices. Surprised to learn about the "64 voices for free" situation that was quietly brought in at 11.1.3. What's going on here? Does AVID needlessly cap voices at 256 for HDX? Couldn't we get more since the DSP isn't being used for plugins? (and never will be since AAX DSP is a joke)

What am I not understanding about voices?

[Frank Kruse]

Originally Posted by loud neighbor

Found this thread doing a search on HDX voice count and wanted to reopen this discussion for my own enlightenment.

I have no experience with HDN - I've got an HDX rig - which was HDX1 until earlier this month when I had to purchase a second HDX card to work on a huge session. It bothers me that I will NEVER use the second card for any sort of DSP (mixer / plugins) but AVID requires it to access another 256 voices. Surprised to learn about the "64 voices for free" situation that was quietly brought in at 11.1.3. What's going on here? Does AVID needlessly cap voices at 256 for HDX? Couldn't we get more since the DSP isn't being used for plugins? (and never will be since AAX DSP is a joke)

What am I not understanding about voices?

HDX doesn't exactly cap voices. It uses the processing power differently. The problem is that complex sessions with a ton of surround busses etc. will eat chips away for the mixer. Also having DSP and native plugs in a channel will eat an extra voice as well.

As a result it's quite often not possible to run HDn sessions on HDX1 or even HDX2. Not so much in straight stereo sessions but when working in surround the cards quickly fill up with mixer-instances that for some strange reason only use 47% of a chip and the rest remains unused. Maybe because of RAM on the cards. Not sure. Maybe it's a bug. Having said that, in many situations HDX2 can play much more tracks than HDn but not always. It really depends on how you work.

If you used TDM back then it's the exact same thing that happens only on a different scale. You will always get your 256 voices per card but they do not necessarily translate to track-count depending on your session and plugin setup.

This doesn't happen on HDn.

More info here:

http://duc.avid.com/showthread.php?t=309994

frank.

[loud neighbor]

Originally Posted by Frank Kruse

HDX doesn't exactly cap voices. It uses the processing power differently. The problem is that complex sessions with a ton of surround busses etc. will eat chips away for the mixer. Also having DSP and native plugs in a channel will eat an extra voice as well.

As a result it's quite often not possible to run HDn sessions on HDX1 or even HDX2. Not so much in straight stereo sessions but when working in surround the cards quickly fill up with mixer-instances that for some strange reason only use 47% of a chip and the rest remains unused. Maybe because of RAM on the cards. Not sure. Maybe it's a bug. Having said that, in many situations HDX2 can play much more tracks than HDn but not always. It really depends on how you work.

If you used TDM back then it's the exact same thing that happens only on a different scale. You will always get your 256 voices per card but they do not necessarily translate to track-count depending on your session and plugin setup.

This doesn't happen on HDn.

More info here:

http://duc.avid.com/showthread.php?t=309994

frank.

Thanks for the info Frank. I guess I don't understand why 256 is the magic number for HDX. There must be a way to reallocate the processing power in the HDX card to free up additional voices in situations where more tracks are needed. I have never maxed out the DSP in my HDX1 system - only needed an additional HDX card to increase the voice (in my case, track) count.

I'm sure I'm not fully comprehending the allocation of processing but since HDN can do 256, it seems ridiculous that HDX can't be customized (playback engine setting?) to do more - especially in a stereo mix setup.

[Craig F]

on TDM/TDM2 the Voices and Processing was handle by the same chips
on HDX cards Voices and DSP are separate/different chips now

[loud neighbor]

Originally Posted by Craig F

on TDM/TDM2 the Voices and Processing was handle by the same chips
on HDX cards Voices and DSP are separate/different chips now

Thanks Craig. I remember reading about that when HDX came out. Makes sense for stability but in reality it is a painful limitation since DSP plugins are so scant and most of the DSP chips are unused in my experience.

Regardless of dedicated chips, why lose the option to manually allocate voices in HDX like we could with TDM? Furthermore, if audio is playing non-simultaneously on separate tracks and those tracks share the same output / routing, why does track count even matter? With "dynamic" voice allocation, it would imply that the software is figuring that out in the background. Since it isn't, we should be able to manually dictate voice allocation in order to preserve voices while maintaining a more organized session.

Between the chips on the HDX card and the computer's host processing there seems to be a lot of wasted power. AVID: give users the option to access all of the resources they have at their disposal! Voices are ALWAYS the first thing to run out in my experience. Even with HDX2!! Silly.