Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: How to decipher this Adaptive algorithm used for Voice Activity Detect (Read 5109 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

How to decipher this Adaptive algorithm used for Voice Activity Detect

Recently in the task of implementing a adaptive energy algorithm for VOICE ACTIVITY DETECTION ,i came across an algorithm which is working properly and well but the problem is i am not able to understand the algorithm in the phase of updating the threshold value,can any one point out why the threshold is increased with a logarithm step, how the threshold is affected if the Emin decreases

Algorithm - source of VAD algo



 

How to decipher this Adaptive algorithm used for Voice Activity Detect

Reply #1 –
There should be no logarithm, just k * E_min.
You'd better ask the author.

How to decipher this Adaptive algorithm used for Voice Activity Detect

Reply #2 –
Quote from: Alexey Lukin on
There should be no logarithm, just k * E_min.
You'd better ask the author.


the author has never replied  it must be having some relation or may be it was giving him more good results by using a log than actual value but there must be a reason for it any one there to explain the mathematical suspense in it

How to decipher this Adaptive algorithm used for Voice Activity Detect

Reply #3 –
The log is likely an approximation to the speech probability/energy distribution. It's actually even closer to gamma distribution, which some other algorithms use, such as various MMSE noise suppressors.
For example, ones written by prof. Israel Cohen: http://webee.technion.ac.il/Sites/People/IsraelCohen/
ruxvilti'a