Is TDM on the way out???? (sorry, lame opener)

[Soundtrack2life]

FWIW even with my modest HD2 project studio I can not even think about not having TDM! I hope they continue development and or support. It truly has change the way I work with a DAW. . . . and in a good way
Joe

[OliJ]

One thing that some people seem to be forgetting about purely Native systems, is that plugin latency is affected by The Hardware Buffer. Double the H/W buffer and you double the latency of RTAS plugins. I'm sure most users will grow to max out their plugin count, and keep increasing the buffer size. Unfortunately, if you're using a lot of bussing, pretty soon your going to find that delay compensation can't pull your RTAS convolution verbs back in to time, as the amount of delay exceeds the compensation available.
TDM will continue remain desirable for me for a little while yet. That being said, who knows what new approaches to DSP/Native may surface in a couple of years.

[Dism]

Originally Posted by OliJ

One thing that some people seem to be forgetting about purely Native systems, is that plugin latency is affected by The Hardware Buffer. Double the H/W buffer and you double the latency of RTAS plugins. I'm sure most users will grow to max out their plugin count, and keep increasing the buffer size. Unfortunately, if you're using a lot of bussing, pretty soon your going to find that delay compensation can't pull your RTAS convolution verbs back in to time, as the amount of delay exceeds the compensation available.
TDM will continue remain desirable for me for a little while yet. That being said, who knows what new approaches to DSP/Native may surface in a couple of years.

Since when?

Buffer size has never affected the latency of plugins for me, unless you're referring to tracking, which is a different beast altogether.

[OliJ]

Try bringing up an Aux track, assign a processor hungry RTAS Plugin, preferably a convolution reverb. Now have a look at how many samples of delay compensation is being applied to that track. OK, now double your hardware buffer, and check again, how many samples of delay compensation is being applied to that track.

[necjamc]

Originally Posted by OliJ

Try bringing up an Aux track, assign a processor hungry RTAS Plugin, preferably a convolution reverb. Now have a look at how many samples of delay compensation is being applied to that track. OK, now double your hardware buffer, and check again, how many samples of delay compensation is being applied to that track.

It doesn't change at all on mine. I tried several plugs and they all stayed the same.

Edit: I tried this on ver 8 on mac, and ver 9 windows. They both stayed the same.

[OliJ]

Originally Posted by necjamc

It doesn't change at all on mine. I tried several plugs and they all stayed the same.

Are you sure that you have either short or long "delay compensation" option selected in the 'Playback Engine" setup ?

[necjamc]

Originally Posted by OliJ

Are you sure that you have either short or long "delay compensation" option selected in the 'Playback Engine" setup ?

I had long enabled (on 9). I made a new session and specifically set that first. On 8 I was using m-powered and only checking the reported latency.

[OliJ]

So in the compensation window at the bottom of the channel strip, the reported delay should increase, whenever you increase the hardware buffer. I just realized that my earlier post was a bit misleading, as there won't be any delay compensation applied to the track with the plugin of course, but instead will be applied other tracks that have no plugins, or ones that are causing less delay.

[necjamc]

The reported delay does not increase at all. if you are on 9, load up just 9 (no HD) and try it yourself. I tried several different ways. No increase.

[OliJ]

Mmmm, strange. I'm still on PT8 on an HD system. So either things have changed, but I thought that this effect of the Hardware buffer on native plugin delay was unavoidable , or there is some problem in the repoting of plugin delay in whichever PT9 version your using. Perhaps some one from Digi ..oops Avid may be ale to confirm this ? I'd be interested to know