Official Serum thread!
- KVRAF
- 2355 posts since 3 Mar, 2010
Very cool! One question - the other page states "Serum has a variety of methods and options for importing + analyzing audio (WAV or AIFF) for breaking it apart into individual waveforms." How long can the imported sample be - for example, can I load a short (5 second or less) vocal phrase) and have a wavetable created from that?
-
- KVRAF
- 4218 posts since 15 Sep, 2010
Very interesting! Will look into it for sure.
- KVRAF
- 14985 posts since 26 Jun, 2006 from San Francisco Bay Area
Yeah, count me as interested. Looked on the site and it's still not listed as a product though.
Zerocrossing Media
4th Law of Robotics: When turning evil, display a red indicator light. ~[ ●_● ]~
4th Law of Robotics: When turning evil, display a red indicator light. ~[ ●_● ]~
-
- KVRAF
- 2111 posts since 25 Jun, 2008 from Montreal, Canada
Very nice about the intro price, thanks!
-
- KVRAF
- 1637 posts since 21 Dec, 2012
Good man for making communication, give it to Deadmau5 :p
https://www.youtube.com/watch?v=8EMhjocxlpQ
https://www.youtube.com/watch?v=8EMhjocxlpQ
- KVRian
- Topic Starter
- 1110 posts since 31 Aug, 2001 from Los Angeles, CA
Yes, you can import a vocal phrase of that length and create a wavetable from it. However for that length, because there is a 256-cycle limit, you'd have to use FFT import options (which are spectrally overlapping larger chunks of audio) or have a very low voice (50 hz??) if you really wanted 5 seconds worth. My natural speaking voice allows for a little over 2 seconds of raw audio (every wave cycle of my voice as a table frame).Very cool! One question - the other page states "Serum has a variety of methods and options for importing + analyzing audio (WAV or AIFF) for breaking it apart into individual waveforms." How long can the imported sample be - for example, can I load a short (5 second or less) vocal phrase) and have a wavetable created from that?
There's a lot of SSE2 optimized code, just to pull off what it does (which is an essentially flat frequency response to 20 khz with an aliasing point well below any other WT synth I'm aware of). That sort of quality comes with a price (CPU wise) but nothing in the absurd realm (every tester asked has found it reasonable or a non-issue). In the future I should be able to provide a 'draft' mode, but I was more concerned with attaining a superior quality while more or less matching the per-voice CPU expense of other WT synths.how good is the CPU usage?
-
penguinfromdeep penguinfromdeep https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=193898
- KVRAF
- 1993 posts since 18 Nov, 2008
That's the attitude I like! Consider me interested!bitcrusher wrote:That sort of quality comes with a price (CPU wise). In the future I should be able to provide a 'draft' mode, but I was more concerned with attaining a superior quality while more or less matching the per-voice CPU expense of other WT synths.
circuit modeling and 0-dfb filters are cool
-
- KVRAF
- 1637 posts since 21 Dec, 2012
-Are the filters fixed or is there an option to paint your own filters?
- What is the format where you saved your wavetables? or are there different one possible?
- Possible to get a "public beta" version to test it or to publish it here like other companies (U-he...) ?
Thanks in advance
- What is the format where you saved your wavetables? or are there different one possible?
- Possible to get a "public beta" version to test it or to publish it here like other companies (U-he...) ?
Thanks in advance
- KVRAF
- 2912 posts since 13 Apr, 2008 from Charleston, SC
- KVRAF
- 2925 posts since 29 May, 2009 from New Zealand
Fixed I believe (based on what I saw on Stephen's (SeamlessR) stream last night.MillerSam wrote:-Are the filters fixed or is there an option to paint your own filters?
Based on a previous stream a while back, wav iirc....?
- What is the format where you saved your wavetables? or are there different one possible?
- KVRian
- Topic Starter
- 1110 posts since 31 Aug, 2001 from Los Angeles, CA
GeorgeZ is correct, they are algorithmic. Painting your own requires IFFT/FFT (huge latency), or an additive synth engine to begin with (or perhaps some form of black magic I'm not disciplined in).-Are the filters fixed or is there an option to paint your own filters?
The format is .wav with additional info for Serum in a header (e.g. interpolation modes, frame size, factory table or user-made). I added a .256 export for Wiard and related modular hardware (untested with hardware, but the file format is simple enough). There are a whole ton of "undocumented" WT import file formats (not tested enough to officially support / reverse engineering file formats involves some guesswork). Other file export formats may be possible to implement in the future upon request.What is the format where you saved your wavetables? or are there different one possible?
Beta is finished up for the most part, I keep that to a closed set of people who communicate with me, I have some professional testers (literally) and they do a great job. There will be a demo version released concurrently. I'm going to pre-release the demo as "beta" for Xfer customers on he product forums (maybe tomorrow, we'll see).Possible to get a "public beta" version to test it or to publish it here like other companies (U-he...) ?
Last edited by bitcrusher on Tue Sep 16, 2014 9:35 pm, edited 1 time in total.
- KVRAF
- 2355 posts since 3 Mar, 2010
Beautiful - thank you!bitcrusher wrote:Yes, you can import a vocal phrase of that length and create a wavetable from it. However for that length, because there is a 256-cycle limit, you'd have to use FFT import options (which are spectrally overlapping larger chunks of audio) or have a very low voice (50 hz??) if you really wanted 5 seconds worth. My natural speaking voice allows for a little over 2 seconds of raw audio (every wave cycle of my voice as a table frame).Very cool! One question - the other page states "Serum has a variety of methods and options for importing + analyzing audio (WAV or AIFF) for breaking it apart into individual waveforms." How long can the imported sample be - for example, can I load a short (5 second or less) vocal phrase) and have a wavetable created from that?
There's a lot of SSE2 optimized code, just to pull off what it does (which is an essentially flat frequency response to 20 khz with an aliasing point well below any other WT synth I'm aware of). That sort of quality comes with a price (CPU wise) but nothing in the absurd realm (every tester asked has found it reasonable or a non-issue). In the future I should be able to provide a 'draft' mode, but I was more concerned with attaining a superior quality while more or less matching the per-voice CPU expense of other WT synths.how good is the CPU usage?
-
- KVRAF
- 2111 posts since 25 Jun, 2008 from Montreal, Canada
It's my birthday today (37, damn, 40 is approaching), can I have a free copy!
- KVRian
- Topic Starter
- 1110 posts since 31 Aug, 2001 from Los Angeles, CA