LIM

Demo
Audio Dynamics Automatic Equalization Inspired by Visual Perception

These examples refer to the following paper:

Ludovico, L. A., Presti, G., & Rizzi, A. Audio dynamics automatic equalization inspired by visual perception. Multimedia Tools and Applications, 1-13.
▸view paper

For further information: ▸giorgio.presti@unimi.it

 

Example 1
Band-Limited Music Signal

A music sample with almost no high frequencies. To restore missing spectral content, music producers usually emply exciters. In this case, ADAE shows a more pronounced mid-range frequency boost, ending up in a less natural but more present sound, if compared with the traditional processing.

Original file

ADAE-processed file (slope = 8, look-behind = 0.005 s, look-ahead = 0.005 s)

Traditional processing (commercial exciter)

Example 2
Band-Limited Vocal Signal with Excessive Dynamics

Band limited vocal tracks with dynamic excursion issues are typical of cheap microphones. Fixing this kind of artefact is generally complex, and can be done with a combination of tools. ADAE provides a dynamics restoration comparable with that of the traditional processing, but with much more saturation, thus resulting in a less natural vocal sound.

Original file

ADAE-processed file (slope = 6, look-behind = 0.6 s, look-ahead = 0.01 s)

Traditional processing (compressor and exciter)

Example 3
Bass Saturation

Bass distortion is frequently used in hard rock productions. Musicians and mix engineers are always in the quest for new distortions to play with. Here ADAE provides a pleasant distortion, as well as a traditional overdrive effect, but at the same time it avoids the dynamic range reduction typical of almost all distortion effects.

Original file

ADAE-processed file (slope = 100, look-behind = 0.004 s, look-ahead = 0.001 s)

Traditional processing (a basic overdrive effect)

Example 4
Music with Reduced Dynamics

An excessive compression is a non-linear process that cannot be reversed; however, using a combination of equalizers and dynamic processors, some transients can be restored. In this case ADAE dynamics restoration sounds more natural than that of a traditional expander, which shows too much volume modulation.

Original file

ADAE-processed file (slope = 1, look-behind = 0.2 s, look-ahead = 0.2 s)

Traditional processing (filter and expander)

 

Laboratorio di Informatica Musicale
Dipartimento di Informatica
Università degli Studi di Milano
Via G. Celoria 18, 20133 Milano, ITALY
 Google Maps
Goffredo Haus, director
Phone: +39 02 50316 222
E-mail: goffredo.haus@unimi.it
  Luca A. Ludovico, technical contact
Phone: +39 02 50316 382
E-mail: luca.ludovico@unimi.it
  Laboratory
Phone: +39 02 50316 382
E-mail: lim@di.unimi.it