Wednesday, June 7, 2017

NSI Instruments

Magenta is a Google Brain project that has created something called NSynth which seems to be extracting the audio DNA from samples and then mixing it all together in a user controllable synth. While I can't give you the synth (it is highly confusing) which will hopefully someday come out as a VST, I can apparently use the NSynth Dataset to create more traditional though possibly funky sounding instruments. So far I've done a grand total of 15 instruments but more are probably on the way as the dataset contains about 1000...

Also I'm trying out Kontakt GUI Maker. The interface is really simple, but it does have a convolution reverb which sounds much better than regular reverb. Note these are all Kontakt instruments and need Kontakt 5.6.6.

PS If you want Kontakt for 125 bucks instead of $400 this month (June 2017) go get free DrumMic'a and register it, then use it to crossgrade to Kontakt during the current sale.

NST Resources (mandatory for instruments to work, only download once)

Because the resource files are bigger than the instruments, I put them in a separate download. This makes every instrument download 90 megs smaller.
Unzip NST Resources in the same folder as the NST instruments you download (in 7-Zip use Extract Here). This will make a NST Resources folder as you see in the picture below. This folder will have 3 files, a .nkc, a .nkr.& a .png file. All NST instruments will use this one resources folder.


Acoustic Piano 00
Acoustic Piano 01
Acoustic Piano 02
Acoustic Piano 03
Acoustic Piano 04   Soundcloud Demo 
Acoustic Piano 05
Acoustic Piano 06 
Acoustic Piano 07

Electronic Keyboard 00

Acoustic Guitar 00

Electric Bass 00

Acoustic String 00
Acoustic String 01
Acoustic String 02
Acoustic String 03
Acoustic String 04

Acoustic Brass 00
Acoustic Brass 01

Acoustic Reed 00
Acoustic Reed 01
Acoustic Reed 02

Acoustic Mallet 00

Acoustic Vocal 00
Acoustic Vocal 01
Acoustic Vocal 02

License: The dataset is made available by Google Inc. under a Creative Commons Attribution 4.0 International (CC BY 4.0) license.

Paper: Jesse Engel, Cinjon Resnick, Adam Roberts, Sander Dieleman, Douglas Eck, Karen Simonyan, and Mohammad Norouzi. "Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders." 2017.

No comments:

Post a Comment