DSP563xx Emulator (Access Virus, Nord Lead, Waldorf MW...)
- KVRAF
- 23101 posts since 7 Jan, 2009 from Croatia
It's doing well.
-
- KVRist
- Topic Starter
- 119 posts since 17 Jan, 2021
Emulation update: Finally, after weeks of debugging and bug hunting, the JIT / dynamic recompiler is taking its first baby steps and generates valid audio
https://cdn.discordapp.com/attachments/ ... Arp_BC.wav
https://cdn.discordapp.com/attachments/ ... Arp_BC.wav
-
- KVRist
- Topic Starter
- 119 posts since 17 Jan, 2021
Yep, the JIT engine works and is being optimized for performance right now. The current execution speed peaks a 102 MIPS (we need 136 MIPS for a Virus C in realtime). Numbers are promising, given that it only took a few hours to get there and there is lots of potential left.
-
- KVRian
- 821 posts since 14 May, 2014
numerouno wrote: ↑Tue Jun 22, 2021 10:14 amYep, the JIT engine works and is being optimized for performance right now. The current execution speed peaks a 102 MIPS (we need 136 MIPS for a Virus C in realtime). Numbers are promising, given that it only took a few hours to get there and there is lots of potential left.
That's fantastic!! hopefully it helps open the doorway for other synths with more chips down the line!
- KVRian
- 1154 posts since 17 Feb, 2010
Hi numerouno,
from its early stages, it was clear this was a cycle-exact emulator.
Will the "performance optimizations" have effect on the accuracy side (introducing high-level approximations of the DSP instructions) or does DSP563XX - and the additional, device-specific emulated chips - mantain its original DSP instruction/cycle-accurate approach ?
from its early stages, it was clear this was a cycle-exact emulator.
Will the "performance optimizations" have effect on the accuracy side (introducing high-level approximations of the DSP instructions) or does DSP563XX - and the additional, device-specific emulated chips - mantain its original DSP instruction/cycle-accurate approach ?
-
- KVRist
- Topic Starter
- 119 posts since 17 Jan, 2021
Not all all, there are no accuracy changes or anything like that, it will always be a 1:1 emulationxhunaudio wrote: ↑Thu Jun 24, 2021 7:31 pm Hi numerouno,
from its early stages, it was clear this was a cycle-exact emulator.
Will the "performance optimizations" have effect on the accuracy side (introducing high-level approximations of the DSP instructions) or does DSP563XX - and the additional, device-specific emulated chips - mantain its original DSP instruction/cycle-accurate approach ?
-
- KVRist
- Topic Starter
- 119 posts since 17 Jan, 2021
Greetings, it’s been a very busy and eventful month for us.
We're glad to announce that we finally have a functional VST (internal at the moment for testing purposes) that can run the synths by themselves in real time.
Realtime VST demos (Virus C) running on an i7 4790k:
https://dbree.org/v/cd8dfd
https://dbree.org/v/29660d
https://dbree.org/v/cfacb4
https://dbree.org/v/e4d9f9
Underruns are caused by the CPU, although the emulator performance is being improved daily
We're glad to announce that we finally have a functional VST (internal at the moment for testing purposes) that can run the synths by themselves in real time.
Realtime VST demos (Virus C) running on an i7 4790k:
https://dbree.org/v/cd8dfd
https://dbree.org/v/29660d
https://dbree.org/v/cfacb4
https://dbree.org/v/e4d9f9
Underruns are caused by the CPU, although the emulator performance is being improved daily
-
- KVRAF
- 5664 posts since 7 Feb, 2013
Amazing!
That Carte Blanche melody
That Carte Blanche melody
You may think you can fly ... but you better not try
-
- KVRian
- 892 posts since 27 Oct, 2004 from Inside the kick drum
My Eset is warning me:numerouno wrote: ↑Thu Jul 01, 2021 11:16 am Greetings, it’s been a very busy and eventful month for us.
We're glad to announce that we finally have a functional VST (internal at the moment for testing purposes) that can run the synths by themselves in real time.
Realtime VST demos (Virus C) running on an i7 4790k:
https://dbree.org/v/cd8dfd
https://dbree.org/v/29660d
https://dbree.org/v/cfacb4
https://dbree.org/v/e4d9f9
Underruns are caused by the CPU, although the emulator performance is being improved daily
Threat found
This web page contains potentially dangerous content.
Threat: HTML/ScrInject.B trojan
Access to it has been blocked. Your computer is safe.
Is there videos on those links? Can you upload it somehwere else?
-
- KVRist
- Topic Starter
- 119 posts since 17 Jan, 2021
Just in case you want to hear more, we uploaded a bunch of audio examples now on SoundCloud
https://soundcloud.com/dsp56300
https://soundcloud.com/dsp56300
-
- KVRian
- 821 posts since 14 May, 2014
Thank you very much! this is the most exciting synth development to happen for an extremely long time!!
Do you think it's feasible now to do synths with multiple chips now?
Do you think it's feasible now to do synths with multiple chips now?
-
- KVRist
- Topic Starter
- 119 posts since 17 Jan, 2021
Not just yet, we have plenty of single DSP synths to get through first
-
- KVRian
- 821 posts since 14 May, 2014
right!! I meant, does it seem technically feasible? I remember it being a concern that multi-chip synths wouldn't be possible because of how CPU unfriendly they could be. Since the new emulator is much farther ahead and now works in real-time, I'm wondering if that's changed