Spectrogram visualisation of the training process of a conditional waveGAN architecture learning to synthesise knocking sound effects with different emotions. Cool to see how the label makes them diverge over time! #DeepLearning#gameaudio#sounddesign
For those of you who didn't make it to the @audiodevcon this year, here is a short video of the procedural audio models with Chunity demo. It features several simple models (impact, a fan, rain & wind) to get started. Get the code here: https://t.co/eQB9JnVBCb #gameaudio#phdchat
*** SURVEY ON VIDEO GAME PLAYER TRAITS ***
Hello! If you play video games (and 18+), please fill in my questionnaire on player traits (rts help too <3)! It takes ~15 minutes, and you could win one of 20 Amazon or Steam vouchers worth £20! Take part here: https://t.co/LvzWEtKfbw
Conditional generation using waveGAN. Interesting to see how the label (concrete, gravel, carpet and wood) affects the synthesised sound over time, considering they are all generated by the same noise vector. #GameAudio#proceduralaudio#DeepLearning#phdchat
First 12000 epochs of a GAN (using the WaveGAN architecture) learning to synthesise footsteps sound effects. Quite promising results so far! Next step: class conditioning. #phdchat#deeplearning#gameaudio#proceduralaudio
Thrilled my demo 'Procedural Audio Models for Video Games with Chunity' got accepted to the Audio Developer Conference! See you in London in November #adc19#phdchat#GameAudio
Thrilled to have the opportunity to present my paper "Perceptual Evaluation of Modal Synthesis for Impact-Based Sounds" at the Sound and Music Computing Conference, which this year happens to be in Málaga, my hometown! Have a read here: https://t.co/DaaoHZgSPg