Audio DSP guy. No interesting illnesses or genetic disposition. Opinions are my own, of course.
Joined August 2009
- Tweets348
- Following45
- Followers374
- Likes67
aivoicedetector.com what a hoax. Suppress a bit of background noise and your recording is "100% AI", add a tiny bit of pink noise and your 100% synthetic voice is suddenly "natural". Wow, what an achievement.
2
154
Natürlich berichten die öffentlich-rechtlichen Medien stets unabhängig und keinesfalls regierungsnah (follow the money!) :)
1
I'm wondering. Has there been any progress in 3D audio over the past 2 decades at all? The tech that I see being used is pretty much the same stuff that we had in the 90s, just with more hardware built around it...
Made some recordings using the Sennheiser Ambeo headset. This is both awesome and fun! I wanted to have something like this ever since I recorded my own HRTFs in the 90s using a frog clicker and in-ear headphones.
I really don't understand why no developer supports 64bit integer data format in WAV and AIFF. I need this to do measurements and the specs don't preclude it. 32bit int/float just isn't enough for < -200dB DFS.
1
3
This is great. Now make this a C++ toolbox so people can build actual products with it. :)
Audio source separation on (a)steroids. This greatly speeds up having reproducible research on speech separation (especially time-domain models). We are working on adding music separation (open-unmix) and MUSDB18 data-loaders. I'm happy to be part of this initiative.
2
1
WORMHOLE's algorithm goes way back to Prosoniq PiWarp developed in 1997 and was also used in the Hartmann Neuron synth... it's great to be (indirectly) part of many current productions like this: Behind The Witcher’s Wild Sound / j.mp/2Z3iqGf via @asoundeffect
Stephan Bernsee retweeted
Spectacular academic paper!
Th brain recognizes familiar music in less than half a second.
This is faster then most visual stimuli.
Interesting that it correlates to studies in song lyric recall in about the same time capsule.
We are wired for sound. nature.com/articles/s41598-0…
Recommended reading: "'GANs' vs 'ODEs': the end of mathematical modeling?" by Alexandr Honchar: towardsdatascience.com/gans-…