Hi All, I've been playing about with the RECORD FILE agi function and am finding two distinct problems with the resulting wav file when using a non zero sample offset. Specifically, I call the function with a zero offset and a given filename (the "original" recording), and then later call it with the same filename and a non-zero offset (the "overdub"). When I do this, I experience the following: * The file has a strange signal superimposed on top of the normal recorded sound signal, from exactly the offset sample forward. In Audacity it looks like a perfect sawtooth wave, however one made up of distinct sharp samples. Needless to say, this produces an annoying buzzing sound on top of the normal sound from the offset point forward, that is to say that the "original" recording with offset 0 sounds great, and then the buzzing begins with the "overdub" recording starting at the non-zero offset from the second RECORD FILE call. * This strange signal, along with the intended recording, abruptly ends at the point at which the "orignal" recording ended, although the agi function does not return from the second RECORD FILE call until the proper escape digit is pressed. Also, the file length appears to be consistent with NOT cutting off the overdub. So it seems like the second RECORD FILE command actually runs to completion propertly, but does NOT actually add any samples to the file after the length of the original recording. This might be Audicity, however, just refusing to show data past the end of the sample size as read in the wav file header if, e.g. the header is not updated at the end of the second call. So, the my real question is whether or not these are known bugs, or if perhaps I've configured something wrong. Has anybody actually ever gotten this agi function to work properly while using a non-zero offset? I haven't seen any examples or discussion of this before... Any help, comments, etc. would be greatly appreciated... Kind regards, John
Hi All, I've been playing about with the RECORD FILE agi function and am finding two distinct problems with the resulting wav file when using a non zero sample offset. Specifically, I call the function with a zero offset and a given filename (the "original" recording), and then later call it with the same filename and a non-zero offset (the "overdub"). When I do this, I experience the following: * The file has a strange signal superimposed on top of the normal recorded sound signal, from exactly the offset sample forward. In Audacity it looks like a perfect sawtooth wave, however one made up of distinct sharp samples. Needless to say, this produces an annoying buzzing sound on top of the normal sound from the offset point forward, that is to say that the "original" recording with offset 0 sounds great, and then the buzzing begins with the "overdub" recording starting at the non-zero offset from the second RECORD FILE call. * This strange signal, along with the intended recording, abruptly ends at the point at which the "orignal" recording ended, although the agi function does not return from the second RECORD FILE call until the proper escape digit is pressed. Also, the file length appears to be consistent with NOT cutting off the overdub. So it seems like the second RECORD FILE command actually runs to completion propertly, but does NOT actually add any samples to the file after the length of the original recording. This might be Audicity, however, just refusing to show data past the end of the sample size as read in the wav file header if, e.g. the header is not updated at the end of the second call. So, the my real question is whether or not these are known bugs, or if perhaps I've configured something wrong. Has anybody actually ever gotten this agi function to work properly while using a non-zero offset? I haven't seen any examples or discussion of this before... Any help, comments, etc. would be greatly appreciated... Kind regards, John
On Tue, 2005-03-01 at 13:42 -0500, John Hammen wrote:> Hi All, > > I've been playing about with the RECORD FILE agi function and am > finding two distinct problems with the resulting wav file when using a > non zero sample offset. Specifically, I call the function with a zero > offset and a given filename (the "original" recording), and then later > call it with the same filename and a non-zero offset (the "overdub"). > When I do this, I experience the following:What file format are you using and what version of asterisk are you using? There had been a bug for all file formats with a header(wavs) that would mess up the recording. A bug was filed with a patch that was accepted during the 1.0.5 deployed time. If you are below 1.0.5 then you are sure to not have the patch. If your 1.0.5 is older than Feb-06, you don't have the patch. -- Steven Critchfield <critch@basesys.com>