Anukari 3D Physics Synthesizer: Beta Released

[jbraner]

a 512 sample buffer size, can have for instance a 12ms output latency,

Plus you need about the same amount coming back in.
That's 24ms latency, which would be VERY good with a buffer size of 512.
I reckon it's more likely to be double that for most people - and that is pretty difficult to play along with.
It's fine if you hold down one key (or chord) but not if you need to play short notes in time.

[WasteLand]

a 512 sample buffer size, can have for instance a 12ms output latency,

Plus you need about the same amount coming back in.
That's 24ms latency, which would be VERY good with a buffer size of 512.
I reckon it's more likely to be double that for most people - and that is pretty difficult to play along with.
It's fine if you hold down one key (or chord) but not if you need to play short notes in time.

24 ms is still good..
but RTL is with ITB not a problem... what is coming back? there is no round trip latency, even if you record the audio. of course the audio will ITB then, but does not come from the interface.

there is only output latency.... with soft synth....

[jbraner]

Well, after the output latency, you'll want to play a MIDI note so it has to come back in - so RTL, isn't it?
Anyway - I don't want to argue with you - but I just think it's awkward doing anything at buffer = 512.

I checked on my system in Reaper and it's 12/23 ms @ buffers =512.
I don't know which is in and which is out - but that's 35ms RTL.
I have to admit, I thought it would be even higher. I was just guessing as I always run at buffers = 64 or 128 when it can't handle that.

[Funky40]

1. For folks with powerful macOS machines who tried Anukari and it didn't perform well, I'm working with Apple to fix this and the first release (0.9.5) with improvements is available at https://anukari.com/download. I'd love to hear feedback on whether this helps. More improvements will be on the way after this release. (Note: you still need a buffer size of 512+.)

I’ve played around briefly with the latest beta last weekend, and performance was indeed noticeable better on my M4 Pro. But it still felt a bit clunky, and after reading your blog post about your conversation with Apple I decided to wait for the next update.

I really hope you/Apple manage to bring down latency somehow. Needing a 512 samples buffer size is a bitter pill to swallow, especially with MPE.

a 512 sample buffer size, can have for instance a 12ms output latency, for the audio interface.
which is a theoretical value anyway.

which in my book, with MPE, is no problem.

normally i do 256 sample buffer size, which is around 6.3 ms output latency.

i can't say it is a bitter pill, i work with it, don't notice it...

you can adapt (or not?).... latency sensitivity is personal, and what you are used to.

GP shows me 10.7mS Latency with 512samples buffer at 48khz.
Which is ofcoarse just a theoretical value

applesilicon macs do perform btw. the best with buffer sizes at around 64-128ms.
But things are also dependend on some "problematic" plugins. Sometimes it´s better for me to set the buffer size higher. But i´d never go above 256buffer size with AS macs.. If i have to, then something´s wrong somewhere.
( talking solely vs. single instrument realtime play usecases)

There seems to be a whole page in this thread that has disappeared btw. . I made a long post.
I tested Anukari vs. 512, 256, 128 samples buffer size. Anukari worked out for me with 128 samples buffer size ( M4Mac-Studio / 40core GPU). I have still slight crackling. also dependend on how complex a preset in Anuakri is, so i run again at a 512 buffer. That´s absolutely playable for me.

I do run realtime play patches with up to 200ms latency. there is sometimes no way around that. I quite often run patches with 40-80ms Latency. 200ms is ofcoarse nearly nolonger "really" playable. You have to hit the keys permanently to keep somehow a groove right.

and it´s true, as has been aformentioned, if you have to start playing in sync after a pause, the latency is a big thing. But up to around 12ms is imho absolutely manageable.
Most of my patches have a higher latency. (but i don´t play in bandcontext. And never have -> "as a keyboarder" )

I´ve seen real stage pros laughing at all the lateccy pickyness some forums people show.

[v1md]

There seems to be a whole page in this thread that has disappeared btw

I think you are mixing up the two Anukari threads.

[Funky40]

I have a question:
some Factory presets don´t play over the whole keyboard range.
What and where do i have to check ?
If it´s in the manual, i go there, any hints where to look ? ( i hate manuals)

i really enjoy to jam along with that thing !

I think you are mixing up the two Anukari threads.

ahh, i see, haha. Thanks

[nanostream]

you can adapt (or not?).... latency sensitivity is personal, and what you are used to.

Absolutely, up to a certain extent it’s subjective, and it depends on the workflow and other factors.

In my case, after switching to a M1 Mac, I found that overall performance was far better with smaller buffer sizes. That was very liberating, because I tend to use lots of plugins which come with additional latency, and it adds up. But for the first time in my life I didn’t have to think about it.

I can add half a dozen of Arturia saturation plugins on each track, some fft based craziness on another, and it just works, i.e. I don’t have to worry too much about CPU overloads or unplayable latency. It’s even better on M4, and I love it. I’ve been messing around with electronic music since the 90s, and sure, it’s possible to work with higher latencies. But I don’t miss those times, and I don’t want to go back.

That said, I’m willing to break my „rules“ if it’s worth it. Anukari seems to be in that ballpark. I can even imagine using it only to generate new samples, and switching back to my usual low latency settings afterwards. That would be the worst case, though, and I’m not sure if it’s worth it yet. I’m certainly not going to make a 512 samples buffer my new default, that’s for sure, let alone for a single plugin I’d probably use only sporadically anyway

In any case, it’s my understanding that the beta and this very thread is meant to give feedback and to iron out bugs. It’s up to the developer to prioritize certain requests or to evaluate if something is even technically feasible. My hope is that, maybe with a bit of help from Apple, it’s possible to bring down latency. At least I’m going to continue to push for it, because it’s important to me

[anukari-music]

I have a question:
some Factory presets don´t play over the whole keyboard range.
What and where do i have to check ?
If it´s in the manual, i go there, any hints where to look ? ( i hate manuals)

i really enjoy to jam along with that thing !

It's under the Preset Properties panel as "Note Range," here's a screenshot:

https://imgur.com/a/s9IRFDO

Generally speaking, we've tried to only limit the note range for instruments that don't really work well outside that range, usually because the physics system goes crazy. Super stiff springs with light masses can sometimes get unstable with really high notes, etc.

[anukari-music]

Regarding the latency discussion... I wholeheartedly want to reduce latency. Everyone's going to have a different tolerance for how much latency is acceptable given their particular workflow, but probably we all agree that having lower latency available is always nice, even if it's just providing headroom/safety.

I do expect some further progress in this area. But it's still an open-ended question as to how low it can go. I'll certainly post updates as I have them.

And Funky40 -- I did see your detailed comments about testing/performance. Really appreciate it! All the feedback from folks on this thread has been very valuable.

[Funky40]

It's under the Preset Properties panel as "Note Range," here's a screenshot:

https://static.anukari.com/support/shar ... enshot.png

Ahh, i see. I totally overlooked that. Thanks alots.

the screenshot was btw. not accessible. Set to "private" or so......

[anukari-music]

the screenshot was btw. not accessible. Set to "private" or so......

Thanks, moved it to imgur: https://imgur.com/a/s9IRFDO

[v1md]

So, just to be clear, are the modular nets issued on a per-voice basis? Issuing a global network could help keep the CPU demand under control for larger configurations. I’m still getting several audio dropouts and there is no headroom for extra processing effects on my M2, unfortunately.

[WasteLand]

Well, after the output latency, you'll want to play a MIDI note so it has to come back in - so RTL, isn't it?

if you play a midi note, it does not come back in...

it generates, depending of the source, a sound. that is the output.

midi is the input.

so in your case 23ms latency.... which is quite well...

and to make it more difficult: a midi input device has a latency... etc... so there is an input latency but not from the sound interface...

and latency is made a bigger problem than it is.
of course some people are very sensitive for it.

but in real live, yes the M1 to M4 chips work better at lower sample buffers. why i don't know. in general i think... Serum 2 ?

well bottom line 23 ms is good. 40ms 50ms you can really hear. but you can get used to it... an organ in a church has a big latency. and it can be played.

[jbraner]

If you play a MIDI note, there's latency coming in - from the MIDI device, into your DAW and into the VSTi.
There is latency from the ms you touch a key until it gets processed in the DAW and into your VSTi.
Is this not correct?
Then there is output latency, which you acknowledge.

So from the ms you touch a key, there is the input latency, then it gets processed, then there is output latency - and then you hear it.
Is this not correct?

I don't really care. If you are OK with 24ms latency - that's fine
Maybe we're just misunderstanding something - and I'll be the first to admit that it could be me

I'm a guitar player - and if I play my guitar through an amp sim with 24ms latency - I *do not* like it. I'm not hearing the note at the same time I feel it on the guitar. Maybe it's less of a problem for keyboard players - especially with strings and pad sounds.

[jbraner]

To put this another way - latency only matters when your playing into your DAW (MIDI or audio) and listening at the same time.
If you're not inputting MIDI (or audio) - who cares what the latency is? Everything will play in time - just slightly later than when you hit "play". So, if you're just mixing (so just listening, but not inputting anything) it doesn't really matter if your latency is 1000 ms. That would just mean it takes a second to start, when you hit "play".