Zero-Latency Plug-Ins... how?

[Picture Start]

I'm curious how many native plug-ins can have 0 samples of latency even without ADC enabled. Anybody know?

[crazyhorserocks]

hopefully not too stupid a question, but what exactly to you mean by "even without ADC enabled"?

tons of my waves plugins induce zero samples of latency, at least according to the waves web site. I asked the same questions about the new avid channel strip today and one response says zero samples (ie. no additional latency).

[Picture Start]

Well, I cut my teeth in TDM land and nothing—NOTHING—was zero latency. You at least had 3-4 samples. In native, somehow sending a signal out to the host processor via a plug-in and back creates no delay of any kind. I just find that surprising.

[nst7]

That still seems to be the case. If you look on the current Waves latency chart, it still seems to be that the same plugins in TDM add an additional 3 mil. latency. Don't know why that is.

http://www.wavesupport.net/content.aspx?id=2213

[CME]

I believe it has to do with the playback buffer. Some plugs are able to do their thing within the buffer setting and not cause any additional latency. In a native system you will always have at least 32-samples of latency. I believe TDM could approach zero latency, but any digital system will have some latency.


Sent from my iPhone using Tapatalk

[Picture Start]

Originally Posted by CME

I believe it has to do with the playback buffer. Some plugs are able to do their thing within the buffer setting and not cause any additional latency. In a native system you will always have at least 32-samples of latency. I believe TDM could approach zero latency, but any digital system will have some latency.


Sent from my iPhone using Tapatalk

But this only applies to recording through the plug-in. I'm talking simply, the plug-in is instantiated and reports no latency. On playback.

[JFreak]

Native plug should not have latency, if properly designed. So yes, reporting zero is allright. Native systems however have latency by design, and it's the size of playback buffer. Everything goes through the buffer and there are problems only if

a) plugin cannot process fast enough and produces pops and clicks, and you have to double your buffer to get over it. You only get as much processing power your host computer can handle within the buffer. So it's a tradeoff between more processing power (bigger buffer) and less latency (smaller buffer).
b) plugin is designed to operate with latency, let's say "a thousand samples", which means that the first thousand samples output silence and the signal starts playing a thousand samples late. And continues to play a thousand samples after the signal input is silenced. THIS is the type of latency that needs to be reported to the mixer.

TDM is different. It is designed to be "zero-latency" with empty mixer. Add a TDM plugin and it needs X samples of processing time, add another which needs Y samples of processing time and the two combined X+Y=Z gives you total latency what is required to process the given signal. This is then reported to the mixer and it is compensated when the signal gets summed with other signals. It is possible to create a very low latency mixer, but heavy plugin use cumulates fast and in real world TDM sessions likely have higher latency than the super-low 32 sample native buffer.

And, of course, whenever you insert your first Native plugin into a TDM mixer, there's going to be both latencies, the TDM processing and the Native buffer.

[Picture Start]

That's helpful, thank you.

[lexaudio]

Well wouldn't native plugins communicate faster than that of dsp?

It wou ld seem to make ksense that native would be faster in terms of latency.

But, the dsp take the strain off the system. 3 samples is nothing

[Jack Ruston]

The TDM plug ins always have a 3 or 4 sample addition to whatever their processing requires. Native plug ins don't. Both systems then have a latency incurred by the converters, and native adds the buffer that native users are familiar with.

TDM has always had a fast, but not latency free mixer. The plug in latency will be basically the same. And the converter latency set by the converter. This is all assuming the same sample rate. With the new Avid interfaces we get a considerably faster conversion. That means an extremely low latency for HDX and an HD Native latency similar to TDM with a 192 IO. Plug ins remain the same. It doesn't matter what system you use, if you add a high latency plug in, that issue can not be skirted. So while people claim that an advantage of TDM is the ability to track onto an already loaded mix, the reality is that it depends what plug ins you're using. I always use to print the mix in those situations on Hd3.

J