Does Pro tools HDX add any latency to VI tracks?

[TNM]

Hi, I realise Pro tools HDX adds latency if one is putting native effects in the path of a monitoring input channel.

But what about just plain old native Vi channels.

I am guessing and *hoping*, that I choose my normal buffer size (which obviously does not affect DSP/HDX input channels),

and it's THAT buffer which determines what my VI live playing latency would be? Or does HDX add extra buffers for that too?

Finally, how does the HDX driver that talks directly to PT, rather than core audio, *behave* at say 64 or 128 samples when playing a VI live? Is 64 the lowest? is the performance the same as core audio in that regard?

Cheers and thanks very much

[YYR123]

Originally Posted by TNM

.....and it's THAT buffer which determines what my VI live playing latency would be? Or does HDX add extra buffers for that too?

So according to the manual yes it does affect the live playing (as you say) Latency. - So higher buffer, higher midi latency. (The manual says all PT versions are subject to this clause (DSP as well)

In my rudimentary and superficial tests (I am super busy with other projects)
I have found that (at least) XPand2 does not add anything latency wise (no delay reported in ADC mix view)

So you can run the buffer down as far as you normally would like to go, and it feels normal. No other work around needed

This is obviously on a track that I have halfheartedly begun (a one off, wedding song) - so the question that remains is at the end of a smaller Audio session

[TNM]

Originally Posted by YYR123

So according to the manual yes it does affect the live playing (as you say) Latency. - So higher buffer, higher midi latency. (The manual says all PT versions are subject to this clause (DSP as well)

In my rudimentary and superficial tests (I am super busy with other projects)
I have found that (at least) XPand2 does not add anything latency wise (no delay reported in ADC mix view)

So you can run the buffer down as far as you normally would like to go, and it feels normal. No other work around needed

This is obviously on a track that I have halfheartedly begun (a one off, wedding song) - so the question that remains is at the end of a smaller Audio session

Right, it makes sense that the buffer chosen affects the latency, like native, but i just meant i didn't know if hdx doubled it cause it's a native Vi, not an AAXDSP (then again, i don't know of any AAX DSP VI's, that would be wonderful if they existed.. guaranteed polyphony per chip, no native usage. That would be an amazing use of DSP chips!)

So basically, that sounds good, hdx is the driver, and i can choose the lowest buffer my system performs OK at, and the vi will play live at that buffer. If no extra latency is being reported, and extra latency IS reported if a native effect is added to an input track, then it would stand to reason that you are correct and a VI on a record armed track in HDX, is whatever the chosen buffer is, and whatever milliseconds output HDX is at that chosen buffer. I presume the driver would be decent.

Cool.

Is the lowest 64 by the way?

[BScout]

Your question is a little confusing as we're looking at instruments vs efx. The buffer setting is the one that affects VI (and native efx.) The HDX stuff (DSP) runs separately unless you make that pass back to the native system and back to the DSP mixer. An HDX system runs its mixer on-card/DSP. A HD system running native, runs on the host processor. When you throw those two systems together, you have to start looking at the delay comp and plugin arrangement means alot.

However, at the end of the day, the HDX DSP plugs aren't adding additional delay vs VI channels. The VI channel already has a delay to pass back to the DSP mixer (not the buffer setting).

[TNM]

Originally Posted by BScout

Your question is a little confusing as we're looking at instruments vs efx. The buffer setting is the one that affects VI (and native efx.) The HDX stuff (DSP) runs separately unless you make that pass back to the native system and back to the DSP mixer. An HDX system runs its mixer on-card/DSP. A HD system running native, runs on the host processor. When you throw those two systems together, you have to start looking at the delay comp and plugin arrangement means alot.

However, at the end of the day, the HDX DSP plugs aren't adding additional delay vs VI channels. The VI channel already has a delay to pass back to the DSP mixer (not the buffer setting).

Ok thank you, now I am confused myself.. LOL. Please forgive me as i put 2 and 2 together.

Sorry i confused you, this is how i understood it.

1)HDX aax dsp plugins, do not add any delay to the monitoring path, other than whatever the plugin has itself. For example, many avid plugins add 10 samples. This will be reflected in the PT mixer.
AAX DSP plugins are completely unaffected by chosen buffer in preferences, they are affected only by a plugin's own delay and the latency of HDX+converter RTL at chosen sample rate.

2) However, If one were to add *native* aax plugins to the monitoring path of HDX, or on an aux or master fader, they DO add latency, even if they are zero latency plugins.. they would add, from memory, at least 2x the buffer chosen latency.. This is fine by me, as i will only use AAX DSP plugins on the master and on record channels. I have to investigate how badly UAD effects on busses would be affected though, i have to look into this as a separate situation.

Please let me know if anything I have said is wrong so far.. and the question i am having difficulty with and indeed the point of this thread is as follows:

3) For Vi's, one chooses a buffer setting in hardware, as we do in core audio PT vanilla, except it is using a proprietary avid HDX driver.
The buffer chosen in settings, is what affects VI live playing, i.e when a track is record armed.
However, my question is this.. does Pro tools HDX add an extra buffer, like it does with native EFFECTS on input paths, OR does it work just like PT vanilla/HDN when using native VI's.

That's basically it. However you just said
" The VI channel already has a delay to pass back to the DSP mixer (not the buffer setting)"

so does that mean there IS an extra delay being added to native Vi channels, vs the way i am using them now in PT vanilla?

The whole darn thing would be solved if avid would allow hd software more inputs with core audio. I'd buy 2 more apollos and it would be done. Why they have to make it so difficult is beyond me.
*THIS* is one of the very few situation where pro tools annoys me.

I also have one more question and i think it's better starting it here than a new topic. This one is about HD Native

4) Does HD native work better at low latencies, for VI's and external monitoring through PT mixer and native effects, than vanilla PT using core audio? At say 64 buffer?

[BScout]

Originally Posted by TNM

Ok thank you, now I am confused myself.. LOL.
1)HDX aax dsp plugins, do not add any delay to the monitoring path, other than whatever the plugin has itself. For example, many avid plugins add 10 samples. This will be reflected in the PT mixer. AAX DSP plugins are completely unaffected by chosen buffer in preferences, they are affected only by a plugin's own delay and the latency of HDX+converter RTL at chosen sample rate.

Yes, if delay comp is on.

2) However, If one were to add *native* aax plugins to the monitoring path of HDX, or on an aux or master fader, they DO add latency, even if they are zero latency plugins..

yes, but Pro Tools will compensate for the plugin latency

they would add, from memory, at least 2x the buffer chosen latency..

no. the latency of the buffer from dsp mixer to core is not the buffer latency selected.

I have to investigate how badly UAD effects on busses would be affected though, i have to look into this as a separate situation.

UAD is completely different as it isn't a plugin as much as a pass off to another subsystem and a return. The latency is greater than a "non-look-ahead, native min phase eq" for instance (that example as it would have least latency of a native plugin.) In some ways, the UAD is much like using Vienna Ensemble for VI as the plugin is actually just sending and returning sound from a separate subsystem that operates outside of the DAW host.

3) For Vi's, one chooses a buffer setting in hardware, as we do in core audio PT vanilla, except it is using a proprietary avid HDX driver.
The buffer chosen in settings, is what affects VI live playing, i.e when a track is record armed.
However, my question is this.. does Pro tools HDX add an extra buffer, like it does with native EFFECTS on input paths, OR does it work just like PT vanilla/HDN when using native VI's.

Like the efx. At the end of the day, the VI works on the host processor and the mixer works on a dsp. It's not much of a difference but, yes, the processing has to pass back to the dsp mixer from the VI. With an efx, you have double buffers (still not the buffer setting but something separate.) With a VI, you only have half (since that is the source generator and you aren't doing both sending and returning -- just sending into the DSP mixer.)

The whole darn thing would be solved if avid would allow hd software more inputs with core audio. I'd buy 2 more apollos and it would be done. Why they have to make it so difficult is beyond me.
*THIS* is one of the very few situation where pro tools annoys me.

They are in it to sell hardware. Only way for expensive (more channel count needed) studios to buy in. If there was not a limit, lots of MADI card systems would be sold and 128 channel out, etc would be cheap.

4) Does HD native work better at low latencies, for VI's and external monitoring through PT mixer and native effects, than vanilla PT using core audio? At say 64 buffer?

No DSP mixer on HD native cards. Everything is done on the host processor. Have to look at it as any other hardware audio interface (except you are looking at the card+interface; much like Motu and Lynx used to do.) So "vanilla" core audio is very much dependent on which audio interface. HD Native + Avid interface lists their specs but it is a "core audio" engine at the end of the day. It's an apples to apple comparison.

[TNM]

Thank you wow, so kind of you for such detail.

So the small HDX delay added to Vi's, do you have any idea what it might be at 44k?

Also, do you have any idea how big the buffer is that native effects add to a monitor path? Say a zero latency native effect?

--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Also, with HDN, i know it's all native processing for plugins, but I had some inkling that there was an FPGA on the card to help with low latency monitoring of audio inputs.

Since it's an apple to apple comparison, I will say that the core audio thunderbolt drivers for apollo are fantastic.

Are HDN drivers any good at low latency when monitoring external sources?

Right now, with core audio, i have 7 totally different interfaces here & i have tried them all, if i put it at 32 or 64 buffer, and monitor say 32 channels through pro tools (not apollo console), it goes berserk if i even add one reverb to any of the "live" channels. I was hoping HDN would be better,otherwise it's going to have to be HDX for me i guess.

Thanks again, you have been so helpful. I really appreciate it.

[YYR123]

Originally Posted by TNM

if i put it at 32 or 64 buffer, and monitor say 32 channels through pro tools (not apollo console), it goes berserk if i even add one reverb to any of the "live" channels.

I owned the Apollo for a few years....and you won’t be happy trying to run the session like that with an Apollo

You have to utilize the Console for monitoring to benefit from the Apollo (that and recording with plugs is the other benefit)

If you monitor from PT it’s not going to be a pleasure for you, I ran all my verbs while tracking from the AUX channel on the Console app.

You really need to utilize the Console to take advantage of the Zero Latency during recording.

Unless I’m missing something

But this is the main reason I went away from the Apollo is because I wanted to run all of my sends on my mixer from pro tools and not a 3rd party app.

[TNM]

Originally Posted by YYR123

I owned the Apollo for a few years....and you won’t be happy trying to run the session like that with an Apollo

You have to utilize the Console for monitoring to benefit from the Apollo (that and recording with plugs is the other benefit)

If you monitor from PT it’s not going to be a pleasure for you, I ran all my verbs while tracking from the AUX channel on the Console app.

You really need to utilize the Console to take advantage of the Zero Latency during recording.

Unless I’m missing something

But this is the main reason I went away from the Apollo is because I wanted to run all of my sends on my mixer from pro tools and not a 3rd party app.

I only use console..

This was for testing purposes.. i used 7 different interfaces and all of them showed ridiculous cpu usage with 32 inputs armed at 32 or 64 buffer when going through pro tools, and trying to add effects. One fab filter reverb on an input monitored track at 32 or 64 buffer sent it over the top. Logic could get 16 of the same plugin with 64 tracks being armed at 32 buffer. Crazy. Same core audio drivers.

This is why i am wondering if HDN is better in this regard.

Interfaces tried:

MOTU Ultralite MK3 FW (and yes if i use the dsp mixer and monitor through that, like apollo, it's fine, but i am talking about native monitoring here)

Mbox 3 USB

Firewire Solo FW

Apollo 8 x 2 TBolt

TBolt Quantum (on loan).

UR824 USB (again, the dsp mixer is great, but native through PT sucks)

RME 802 (again, through total mix is fine, but native sucks)



your reason is exactly why i want to move away from apollo to HDN or HDX..

so i can have PT as the one stop shop including all my monitoring.. 64 ins and will even be able to get rid of my 32 analog mixer in that scenario.

But i am simply saying that core audio through PT, at 32 and 64 buffer, monitoring a large number of tracks, sucks. (you can emulate high track count with a smaller interface by just assigning the same input to each track and input monitoring them all.. on the apollo i simulated 64 ins through native PT this way, etc).

And that's why i am wondering if HDN is better in that regard. I am scouring the web constantly for this question and can't seem to find answers, or a single video on how HDN reacts with 32+ tracks input monitor armed at lowest possible buffer. All i am trying to find out is if it's better than core audio..

If not, HDX it is.

Perhaps the HDN portion *should* have it's own topic, as others who may be able to answer might not be looking at this one.. hmmmm...

[YYR123]

Well at this point you really ought to go be friends with your dealer.

they will let you take that card home and try it out for a week, and then you can decide if it works out for you