directgumby

I'm working on a project where I am using a digital SLR camera coupled with audio recordings. What I would like to do is sync the audio file with the photos via looking for the click from the shutter (it is noticeable in the audio at this time).

The goal is to take the audio file, coupled with a known directory of images, then run analysis on the audio file to know where to line up the images. Really, what I need is a way to have the audio file analyzed, when it finds the "click" return a timestamp, and then I can use that timestamp to link a file.

1) Would anyone have any thoughts on the best way to analyzed the audio?

2) I was considering using FFTs to do this work but I am not sure if this is the best approach in Max/MSP

3) From there I was going to run a tool like iZotope RX doctor and remove the click from the camera shutter.

I'm working on a project where I am using a digital SLR camera coupled with audio recordings.  What I would like to do is sync the audio file with the photos via looking for the click from the shutter (it is noticeable in the audio at this time). 

The goal is to take the audio file, coupled with a known directory of images, then run analysis on the audio file to know where to line up the images.  Really, what I need is a way to have the audio file analyzed, when it finds the "click" return a timestamp, and then I can use that timestamp to link a file. 

1) Would anyone have any thoughts on the best way to analyzed the audio?
2) I was considering using FFTs to do this work but I am not sure if this is the best approach in Max/MSP
3) From there I was going to run a tool like iZotope RX doctor and remove the click from the camera shutter.  

wanting-to-find-a-known-click-in-an-audio-file-for-sync-purposes

i think it can be done with some clever combination of comparison operators.

given that the click looks like 0. 1. 0., you could do something like:

then add these three results and do [==~ 3.]

any occuring 1. (true) result means that it was- almost certain - your sync click

of 0. 1. 0. which had been played there (1 sample ago).

you could also search for 0. 0. 1. or only 0. 1. when the 1 sample delay is a problem

for you (multitrack application?) - but 0. 1. might happen elsewhere more often than

cant try it in max now, but you can do that yourself.

[==~ 0.] - [delay~ 2]
[==~ 1.] - [delay~ 1]
[==~ 0.] - 

then add these three results and do  [==~ 3.]

any occuring 1. (true) result means that it was- almost certain - your sync click
of 0. 1. 0. which had been played there (1 sample ago).

you could also search for 0. 0. 1. or only 0. 1. when the 1 sample delay is a problem
for you (multitrack application?) - but 0. 1. might happen elsewhere more often than
0. 1. 0.

Or use delta~ (which in a sense is also what Roman describes)

Wanting to find a known "click" in an audio file for sync purposes