Telephone Bell    |   Mammals    |   Birds    |   Rain    |   Bouble pour    |   Fire    |   Evaluation Results    |   Combinations

Mammals

 

Tested with Chrome 58.0.3029.110 (64-bit)

Interface comments

  • The taxonomy used for the sliders is a mix between perceptual and technical labels (e.g. Filter Q and Harshness) It would be good to group and saparate technical and perceptual taxonomies so that the interface is conceptually more clear.

  • For this model, better to avoid "obsucre" technical parameters (e.g. pulse open closed and return) and privilage more perceptual controls.

  • Generally, for technical labels, add number boxes to display sliders values

  • While discussing with Will, it emerged that, in order to simplify the interface, it would be good to group the parameters "pulse open, pulse close, pulse return, noisiness and harshness" into a singol slider . The parameter could refer to a perceptual feature such as "harshness or growlness". Pulse sliders change the sound as less harsh (with low values) and more harsh (with high values).

  • The slider Filter Q refer to "F0", "F1", "F2", "F3" (which are the central frequencies of band pass filters). Change "filter Q" in something that refer to the formant sounds (e.g. formants brightness).

  • "Duration" - add a number box to specify slider value (in seconds?)

  • Better define the paramiters "F0", "F1", "F2", "F3": more explicitly (e.g. Fundamental Frequency, First formant, Second formant, Third formant)


Model comments

  • While playing sounds artifacts are generated (clicks)

  • Each time the algorithm is triggered the resulting sounds is slitly different. This seems to be related to (i) the variation of the amplitude evenvelop (genralli divided in two parts) and (ii) the variation of the fundamental frequency (glissando). Will suggested that this bug is probably related to the "duration" parameter. This does not happen when I play the pd patch. Ideally, if the parameters are not changed the sound should be always the same.

  • Increase the frequency range of the "low pass filter" in order to be able to decrease the energy in the lower frequency range

  • Note that the Growl samples often have the highest component disappering in the second half of the sound, this variation could be introduced for Growl sounds applying a highpass filter

  • RTSFX sounds seems to lack brightness and noisyness. Add more frequency components to get noisy energy in the hihger frequency range



Sound designers comments

  • Not good. I could not find a single usable sound.



RTSFX syntheses - Recorded samples comparison.


RTSFX Fox Roar (preset)

RTSFX Fox Spectrogram

Sample Fox Roar

Sample Fox Spectrogram

RTSFX Leopard Growl (preset)

RTSFX Leopard Spectrogram

Sample Leopard Growl

Sample Leopard Spectrogram

RTSFX Puma Growl (preset)

RTSFX Puma Spectrogram

Sample Puma Growl

Sample Puma Spectrogram

RTSFX Lion Roar (preset)

RTSFX Lion Spectrogram

Sample Lion Roar

Sample Lion Spectrogram

RTSFX Tiger Roar (preset)

RTSFX Tiger Spectrogram

Sample Tiger Roar

Sample Tiger Spectrogram

RTSFX Wolf Growl (preset)

RTSFX Wolf Spectrogram

Sample Wolf Growl

Sample Wolf Spectrogram