Autocorrelation and pitch detection

In summary, autocorrelation is a technique used in pitch detection and echo removal. It involves comparing an input waveform at different time intervals in the time domain. It can be implemented in practice using tools like Matlab and there are resources available, such as the music-dsp mailing list and the book "Digital Processing of Speech Signals" by Rabiner and Schafer. However, implementing a summation in practice may require further knowledge or resources.
  • #1
sathya_rajan
2
0
I have heard that autocorrelation is used in pitch detection, but no proper or convincing explanation of how it is used. I had previously worked on echo removal using autocorrelation. But what puzzles me is that autocorrelation is done in time domain but how can this be used to detect fundamental frequency of some unknown source?
 
Engineering news on Phys.org
  • #2
Go to www.google.com

Type in "autocorrelation pitch detection", hit the "I am feeling lucky" button.
 
  • #3
Floid: I had done that before. The input waveform is compared (autocorrelated) at different time intervals. I don't know whether this can be implemented in practice. I am interested to perform this in matlab. If you have time, then you can explain. Please don't give trivial replies. Thank you.
 
  • #4
also try joining the music-dsp mailing list or posting to the comp.dsp newsgroup.

also there is a book by Rabiner and Schafer Digital Processing of Speech Signals.

now, do you know how to implement a summation in practice?
 

FAQ: Autocorrelation and pitch detection

1. What is autocorrelation and how does it relate to pitch detection?

Autocorrelation is a mathematical technique used to measure the similarity between a signal and a time-shifted version of itself. In the context of pitch detection, autocorrelation is used to find the fundamental frequency of a sound by identifying the repeating patterns in the signal.

2. What types of signals can be analyzed using autocorrelation for pitch detection?

Autocorrelation can be used on any type of signal that contains harmonic content, such as musical instruments, human speech, or synthetic sounds. It is particularly effective for periodic signals, which have a repeating pattern.

3. How does autocorrelation differ from other methods of pitch detection?

Unlike other methods of pitch detection, autocorrelation does not require any prior knowledge about the signal or the fundamental frequency. It is a simple and efficient method that can be used on a wide range of signals.

4. Can autocorrelation be used for polyphonic pitch detection?

No, autocorrelation is primarily used for monophonic pitch detection, meaning it can only detect one pitch at a time. Polyphonic signals, which contain multiple pitches, require more complex methods of pitch detection.

5. Are there any limitations or challenges associated with using autocorrelation for pitch detection?

One limitation of autocorrelation is that it is sensitive to noise and other disturbances in the signal, which can affect the accuracy of the results. Additionally, it may not work well for signals with a low signal-to-noise ratio or for signals with a complex harmonic structure.

Similar threads

Back
Top