Live 9.2 in beta. Latency compensation fixes, and legacy OSes dropped

[2020]

The silly buffer latency for any external plugin is indeed gone now Theo. And what joy! I couldn't be happier!

Best,
2020

so basically, if one is not using effects with noticeable latency, they can get away with everything now incl automation and tempo effects.

Exactly. In that case you don't even need to have Latency Compensation activated, since there's nothing to compensate for anymore.

I'm pleasently surprised. I honestly never thought they would change this.

[2ZrgE]

Not sure how I'm supposed to read the latency meter they added yesterday. The info it shows means nothing to me. Doesn't matter anyway all sounds in time to me.

I guess: The more latency, the higher the probablity of screwing up the overall timing, espcially with a couple of high latencies in a row. So you can decide whether to keep a high latency plug or get rid of it. But I'm not really that deep into this tecnical stuff.

Frankly only care if it is dau proof or not

[chroma]

you're supposed to use it to catch cases where the plugin doesn't accurately communicate its latency - if you know a plugin has latency but live says it's 0 samples, then you know why it's failing and can complain to the plugin maker instead of Ableton, which will be more effective (though I guess ableton could manually override the latencies for errant plugins occasionally).

it's also potentially useful to check latencies for plugins that change latency - e.g. if you add lookahead to a compressor, the latency will change. live appears to detect these changes sometimes, but not always...

[memyselfandus]

can someone explain in more detail when we will see latency and when we won't now? a video would be superb.

[tooneba]

The silly buffer latency for any external plugin is indeed gone now Theo. And what joy! I couldn't be happier!

Best,
2020

Then what the heck was the silly buffer latency for??? It existed for nothing

[2020]

Then what the heck was the silly buffer latency for??? It existed for nothing

You'd have to ask Ableton about that. If I had to venture a guess I think it had to do with the gapless audio engine in Live and the state of processor power at the time. But then again, there was no such buffer latency for the built in devices so, yeah, it never made any sense.

[2ZrgE]

you're supposed to use it to catch cases where the plugin doesn't accurately communicate its latency - if you know a plugin has latency but live says it's 0 samples, then you know why it's failing and can complain to the plugin maker instead of Ableton, which will be more effective (though I guess ableton could manually override the latencies for errant plugins occasionally).

it's also potentially useful to check latencies for plugins that change latency - e.g. if you add lookahead to a compressor, the latency will change. live appears to detect these changes sometimes, but not always...

That makes sense, thanks for the info.

Man this endless Live PDC story has grown to a big "No one really knows what is exactly happening" story, or am I mistaken? Information policy from Abelton on this matter is pathetic, especially as a non-native speaker I still don't get the whole problem.

I can understand that people grab Bitwig...

[dolomick]

I asked this earlier but didn't get a clear response, so I want to ask again - if one puts an auto filter or beat repeat first in the chain, followed by some latency inducing plugins, (and that channel has automation), will that stay in time with the new PDC fixes?

[Mister Natural]

over at the LIVE forum : thiophilus says, "basically, in the past, VSTs would get a buffer equivalent to the current audio buffer size, even if it was a zero-latency effect. This was often so small to not be noticeable, but when you started stacking VSTs it would add up.

anyways, i have a test set that has a bunch of VSTs in, all zero latency VSTs (actually I just stacked a bunch of volume shapers). In live 9.1.7, the master shows tons of latency, audio is out of sync with the midi clock, etc.

in live 9.2b1, the master shows _zero_ latency, and the audio is in sync with the midi clock. (i used a buffer value of 2048 here in both cases, to make the issue worse when it appears).

this is also pretty big I think... in live 9.1.7 and previous, if you threw 5 effects on a track and then something like FilterShaper, LFOTool, or Camelspace, it would already be misaligned to the bar, and the amount would be based on your audio buffer, _even if all the effects were zero-latency effects_. in live 9.2b1, this case will pass; only long-latency plugins will still cause the issue."

[2020]

I asked this earlier but didn't get a clear response, so I want to ask again - if one puts an auto filter or beat repeat first in the chain, followed by some latency inducing plugins, (and that channel has automation), will that stay in time with the new PDC fixes?

Hi dolomick,

Since this is a beta I think people are reluctant to give you an definite answer that might change before the final release. I just tested a couple of things that Ableton said were fixed and reported them here.

I don't have time to test this at the moment and I'm not sure what you mean by "will that stay in time", but I think what you describe should already properly compensate in previous Live versions.

Do you have access to the Live beta? Maybe you could ask in the beta forum directly?

[chroma]

I asked this earlier but didn't get a clear response, so I want to ask again - if one puts an auto filter or beat repeat first in the chain, followed by some latency inducing plugins, (and that channel has automation), will that stay in time with the new PDC fixes?

not sure the question... beat-synced effects in the chain were always synced, even in 9.1.7. but automation on those effects wasn't compensated, and now it is (with some caveats I am looking at...)

edit: of course i meant 'beat-synced effects _first_ in the chain were always synced'... elsewhere in the chain, they may not be.

[dolomick]

I don't have beta access, but it seems like Chroma answered my question. As he said, automation was never compensated on those effects, and now it is.