Re: [chuck] generative program, audio problems when sporking
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 yeah, I could put the samples up, but i'd rather not because they are very large. any files at all should work with minor tweaking, just swap out the string being passed as the name. the attributes are five-digit integers, ordered with the least significant (number 0) on the right - sorry I forget the technical name for this. for the purposes of testing, they really make no difference though. matt Michael Heuer wrote: | On Wed, 27 Apr 2005, Michael Heuer wrote: | | |>Ah, oops, I suppose I would need the samples in order to give it a proper |>try. Could you make these available somewhere? Otherwise a bit of doc on |>what the attribute values mean and I could substitute my own set of |>samples... | | | Wait a minute, I see now, the attribute values are in reverse order. | | I'll stop bothering you with emails now... :) | | michael | | - -- Matthew Shanley little secrets records and stuff http://littlesecretsrecords.com/ E-mail: matthewshanley@littlesecretsrecords.com GnuPG Public Key ID: 0x4DA4915C -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFCcCClIrtD7k2kkVwRAsZ5AJ4q/vvQ+Ee4YjYYpWchK0vy4PalbgCfQxOW BN/P/60UBwYSinLyd5v8KIs= =k9A4 -----END PGP SIGNATURE-----
Hi Matthew, Looking through the code, I think ChucK's 'sndbuf' is the source of the audio problems. For now, the sndbuf tries to load the entire file into memory - bad!!! - we need to fix this soon. The file I/O is most likely causing havoc, especially if the files are large. It has been on our schedule to allow many readers to the same sndbuf, but this requires the new object framework so it isn't done yet. Also, we will add more objects as we are deeper into 1.2, that will allow preloading and other file I/O features. For now, If possible, sndbuf's should load files at the beginning of chuck time. But this looks difficult since the program might like to allocate sndbuf's at each recursive call/spork. Perhaps it is possible to work out some scheme to share a group of global sndbufs? Meanwhile we will continue to find a better solution to this problem. Best, Ge! On Apr 27, 2005, at 7:31 PM, Matthew Shanley wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
yeah, I could put the samples up, but i'd rather not because they are very large. any files at all should work with minor tweaking, just swap out the string being passed as the name. the attributes are five-digit integers, ordered with the least significant (number 0) on the right - sorry I forget the technical name for this. for the purposes of testing, they really make no difference though.
matt
Michael Heuer wrote: | On Wed, 27 Apr 2005, Michael Heuer wrote: | | |>Ah, oops, I suppose I would need the samples in order to give it a proper |>try. Could you make these available somewhere? Otherwise a bit of doc on |>what the attribute values mean and I could substitute my own set of |>samples... | | | Wait a minute, I see now, the attribute values are in reverse order. | | I'll stop bothering you with emails now... :) | | michael | |
- -- Matthew Shanley little secrets records and stuff http://littlesecretsrecords.com/ E-mail: matthewshanley@littlesecretsrecords.com
GnuPG Public Key ID: 0x4DA4915C -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFCcCClIrtD7k2kkVwRAsZ5AJ4q/vvQ+Ee4YjYYpWchK0vy4PalbgCfQxOW BN/P/60UBwYSinLyd5v8KIs= =k9A4 -----END PGP SIGNATURE----- _______________________________________________ chuck mailing list chuck@lists.cs.princeton.edu https://lists.cs.princeton.edu/mailman/listinfo/chuck
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Thanks, Ge - I altered my code around that suggestion and it made things sound great. If anyone is interested, I updated the code on the wiki: http://wiki.cs.princeton.edu/index.php/IvyTree matt Ge Wang wrote: | Hi Matthew, | | Looking through the code, I think ChucK's 'sndbuf' is the source of | the audio problems. For now, the sndbuf tries to load the entire file | into memory - bad!!! - we need to fix this soon. The file I/O is most | likely causing havoc, especially if the files are large. It has been on | our schedule to allow many readers to the same sndbuf, but this | requires the new object framework so it isn't done yet. Also, we will | add more objects as we are deeper into 1.2, that will allow preloading | and other file I/O features. | | For now, If possible, sndbuf's should load files at the beginning of chuck | time. But this looks difficult since the program might like to allocate | sndbuf's at each recursive call/spork. Perhaps it is possible to work | out some scheme to share a group of global sndbufs? Meanwhile we | will continue to find a better solution to this problem. | | Best, | Ge! - -- Matthew Shanley little secrets records and stuff http://littlesecretsrecords.com/ E-mail: matthewshanley@littlesecretsrecords.com GnuPG Public Key ID: 0x4DA4915C -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFCcXXuIrtD7k2kkVwRAuYLAJ4kBdnS1nO8JIdYPF1bRjVcWHtuuQCeI6hu VZQfYmwIDKsiR1YObGv+cTc= =aRSe -----END PGP SIGNATURE-----
participants (2)
-
Ge Wang -
Matthew Shanley