-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hello,
Jumping fast to certain coefficient (for frequency around 40/50 Hz and Q around 0.7) cut sound on iOS (iPad) when i use libpd (it seems to be OK with MacOSX). I need to restart the iOS after. Do you know what I have to do to solve this issue ? Thanx. ++
Jack
On 2015-07-14 16:11, Jack wrote:
Hello,
Jumping fast to certain coefficient (for frequency around 40/50 Hz and Q around 0.7) cut sound on iOS (iPad) when i use libpd (it seems to be OK with MacOSX). I need to restart the iOS after. Do you know what I have to do to solve this issue ?
provide a test-patch.
two ideas:
- the quick jump causes some very high sample values. the iOS system
limiter~ mutes the signal and takes ages to recover (too long to wait). a quick fix would be to use [clip~] just before the [dac~]s
- the quick jump might cause a NaN to be introduced in the signal
vector; due to the recursive nature of [biquad~] all successive samples become NaNs. no quick fix for this. try to get the sample values after muting. e.g. using [print~], or even just [env~]...
fmasdr IOhannes
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Thanx IOhannes,
First : Isolate the biquad~ "stuff" has solved the issue (without [clip~ -1 1])
Secondly : Adding a [clip~ -1 1] just before the [dac~] doesn't solve the problem.
Thirdly : But I don't know yet where the problem comes : if a [clip~ -1 1] is added after the biquad~ "stuff", it works, no crash.
I have to dig a little bit more... ++
Jack
Le 14/07/2015 16:53, IOhannes m zmoelnig a écrit :
On 2015-07-14 16:11, Jack wrote:
Hello,
Jumping fast to certain coefficient (for frequency around 40/50 Hz and Q around 0.7) cut sound on iOS (iPad) when i use libpd (it seems to be OK with MacOSX). I need to restart the iOS after. Do you know what I have to do to solve this issue ?
provide a test-patch.
two ideas: - the quick jump causes some very high sample values. the iOS system limiter~ mutes the signal and takes ages to recover (too long to wait). a quick fix would be to use [clip~] just before the [dac~]s
- the quick jump might cause a NaN to be introduced in the signal
vector; due to the recursive nature of [biquad~] all successive samples become NaNs. no quick fix for this. try to get the sample values after muting. e.g. using [print~], or even just [env~]...
fmasdr IOhannes
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-
IOhannes m zmoelnig -
Jack