2021-06-07

Wren demos

Some audio demos, and patches they were made with. These are all from self playing patches (noodles).

These were all run on a samplerate of 48k with a control rate decimation of 32 to make it possible to have large patches.

2021-06-04-01-11 🕯️


download patch

02-2021-05-27-01-10 Gewoon Toevallig


download patch

03-2021-05-15-10 Het Sleutelkruid


download patch

  1. wakyct

    Tuesday, June 8, 2021 - 02:29:59

    Nice jan! I finally stress tested my tower PC with one of your big patches :). I’m curious what your audio setup is — mine currently is using Windows DirectSound with a buffer of 2048. The patch (Gewoon) maxes my CPU :).

  2. jan

    Tuesday, June 8, 2021 - 03:12:10

    I’m running it on my laptop with an RME audio interface using ASIO .. laptop is not the newest thing … the patch you mentioned runs at around 50% to 60% here for the audio thread (what Wren reports).

    Let me see what I actually have …

    An ASUS GL503VD with an i7 7700 HQ processor 2.80 GHz, windows says, and it actually runs at about 3.25 GHz with a total load over all cores of around 20% - 25%, so there is some other stuff running as well. The fan is audible but not crazy with a room temperature of 21C currently.

    The computer is from around 2018-01-01, well that was when windows was installed..

    My audio interface is an RME Fireface UFX, connected trough USB.

    I have a fairly big ASIO buffer size as well, 1024 samples, with an input and output latency of about 22ms.

  3. wakyct

    Tuesday, June 8, 2021 - 03:57:21

    Interesting, sounds like I should play with the sound setup, buffer and latency, since this benchmark seems to say my cpu should work OK, https://cpu.userbenchmark.com/Compare/Intel-Core-i7-7700HQ-vs-AMD-Ryzen-5-2600X/m211019vs3956

    I’m guessing the RME interface doesn’t have much to do with Wren just running the patch?

  4. jan

    Tuesday, June 8, 2021 - 15:02:13

    Yes, looks like your tower should outperform my laptop,

    But, I run the patch at 48k with a control rate decimation of 32, that might explain it?

    What I’ve noticed is that when a patch gets more complex it gets more difficult to hear the difference between 48k and 96k.

    Just ran a quick test, and I see no significant change in CPU usage when I use the laptop’s internal audio over DirectSound instead of using the RME thing with ASIO - there is more latency using DirectSound, but that’s not very relevant for a self playing patch.

    I did not really do any weird things to this laptop for speeding it up, or at least can’t recall. I probably set the CPU to high performance and processor scheduling to Background Services .. that’s about it.

    Oh and probably killed some processes, but seeing you have 6 cores / 12 threads there should be plenty of CPU for such stuff.

    But anyway, I guess it will be a sample rate thing.

  5. wakyct

    Thursday, June 10, 2021 - 04:02:30

    Yep, simple as halving the sample rate and it plays around 60% CPU. Sounds great! Like Miles Davis in space :).

Add comment

Fill out the form below to add your own comments