2021.6 Instrument Track DSP mode.

[Marsdy]

It’s kind of frustrating that Instrument Tracks containing a VI plug-in can’t be put into DSP mode. In HDX “classic” mode it worked. I don’t see the difference between streaming audio from a virtual instrument into the HDX mixer from streaming an audio audio track into DSP. Both originate natively. Does anyone know why this is the case?

The odd thing is that Instrument Tracks with returns from multiple output VIs, specifically VE Pro, can be put into DSP mode which is kind of interesting!

So could the first plug-in in the chain, as in a virtual instrument, get DSP safe mode for VIs in future?

Glad to see that Instrument Tracks with a physical hardware input can also be put into DSP mode for low latency input from hardware synths etc.

[LukeHoward]

Originally Posted by Marsdy

It’s kind of frustrating that Instrument Tracks containing a VI plug-in can’t be put into DSP mode. In HDX “classic” mode it worked. I don’t see the difference between streaming audio from a virtual instrument into the HDX mixer from streaming an audio audio track into DSP. Both originate natively. Does anyone know why this is the case?

I think it helps to see DSP Mode as a way to reduce latency when monitoring from hardware inputs (by providing a DSP mixer path for those inputs and their dependents), rather than a way of offloading processing to DSPs, even though that is possible with audio tracks.

If you want the old behaviour you can always turn Hybrid Engine off

[Marsdy]

Originally Posted by LukeHoward

I think it helps to see DSP Mode as a way to reduce latency when monitoring from hardware inputs (by providing a DSP mixer path for those inputs and their dependents), rather than a way of offloading processing to DSPs, even though that is possible with audio tracks.

If you want the old behaviour you can always turn Hybrid Engine off

That’s the thing though, I want a low latency monitoring path for VIs at high buffer settings, not just for hardware inputs! It’s not about offloading DSP

For me the problem with the old behaviour is the 256 voice limit with a large composing template!

[albee1952]

Originally Posted by Marsdy

That’s the thing though, I want a low latency monitoring path for VIs at high buffer settings, not just for hardware inputs! It’s not about offloading DSP

For me the problem with the old behaviour is the 256 voice limit with a large composing template!

This can't happen unless companies start making DSP Virtual Instruments

[Marsdy]

Originally Posted by albee1952

This can't happen unless companies start making DSP Virtual Instruments

Why not though? Let’s say you were punching in on an audio track that was in DSP mode, (not even tried it yet but I presume you can!) Playback of audio is coming from the native world and streaming into DSP until you punch in and then you monitor the track input. In other words, the audio playback is originating natively and being piped to the DSP until you punch in.

So why can’t a native VI pipe to DSP in the same way an audio track can? Separate/individual outputs of a VI can go DSP so what is the distinction between those and the audio coming out of the VI on the track it’s actually instantiated on.

[LukeHoward]

Originally Posted by Marsdy

Why not though? Let’s say you were punching in on an audio track that was in DSP mode, (not even tried it yet but I presume you can!) Playback of audio is coming from the native world and streaming into DSP until you punch in and then you monitor the track input. In other words, the audio playback is originating natively and being piped to the DSP until you punch in.

The existing recorded audio (or indeed MIDI) isn't the issue, because Pro Tools can read that as far ahead of time as it needs to. Instead, it is the punching in. With a VI you are always subject to the host buffer size; with an audio track in input/record fed by hardware, you can short cut through the DSP mixer so it never hits the CPU.

[The Weed]

This thread may help, see post 8.

[Marsdy]

Originally Posted by LukeHoward

The existing recorded audio (or indeed MIDI) isn't the issue, because Pro Tools can read that as far ahead of time as it needs to. Instead, it is the punching in. With a VI you are always subject to the host buffer size; with an audio track in input/record fed by hardware, you can short cut through the DSP mixer so it never hits the CPU.

I understand that. What I don’t get is that an Instrument Track that is getting its input from a native VI can actually be put in DSP mode. I can record enable that track (or select it if MIDI thru is set to last selected) and play the native VI, even though the track is in DSP mode!