Hi,
I m trying to write a multitrack recorder based on writesf~ and readsf~ . I m faced with a serious issue... There seems to be some latency either when writing to a file or reading from it (i can't tell). This means that i can't record a second track on top of another as an overdub and playback both. Assuming that Pd's diplayed latency in the audio settings is right (???), that is 100ms, i tried triggering the readsf~ after a 100 ms-long delay, to no avail. I thought it would be a piece of cake to program this, but well, apparently it's not...? Needless to say that i can't use table as i plan to use it for possibly long recordings (my goal is to use it on a phone with PdDroidParty).
So here's the question : is there any latency when writing or reading to a file, and if yes, is it measurable and foreseeable?
Thanks!
Pierre
I forgot to mention that i can't use sfplay~ since PdDroidParty doesn't support externals...
2011/9/11 Pierre Massat pimassat@gmail.com
Hi,
I m trying to write a multitrack recorder based on writesf~ and readsf~ . I m faced with a serious issue... There seems to be some latency either when writing to a file or reading from it (i can't tell). This means that i can't record a second track on top of another as an overdub and playback both. Assuming that Pd's diplayed latency in the audio settings is right (???), that is 100ms, i tried triggering the readsf~ after a 100 ms-long delay, to no avail. I thought it would be a piece of cake to program this, but well, apparently it's not...? Needless to say that i can't use table as i plan to use it for possibly long recordings (my goal is to use it on a phone with PdDroidParty).
So here's the question : is there any latency when writing or reading to a file, and if yes, is it measurable and foreseeable?
Thanks!
Pierre
writesf~ creates a subthread whose task is to write audio streams to disk. You need not provide any disk access time between "open" and "start", but between "stop" and the next "open" you must give the object time to flush all the output to disk.
(from writesf~ help)
On Mon, Sep 12, 2011 at 4:12 AM, Pierre Massat pimassat@gmail.com wrote:
I forgot to mention that i can't use sfplay~ since PdDroidParty doesn't support externals...
2011/9/11 Pierre Massat pimassat@gmail.com
Hi,
I m trying to write a multitrack recorder based on writesf~ and readsf~ . I m faced with a serious issue... There seems to be some latency either when writing to a file or reading from it (i can't tell). This means that i can't record a second track on top of another as an overdub and playback both. Assuming that Pd's diplayed latency in the audio settings is right (???), that is 100ms, i tried triggering the readsf~ after a 100 ms-long delay, to no avail. I thought it would be a piece of cake to program this, but well, apparently it's not...? Needless to say that i can't use table as i plan to use it for possibly long recordings (my goal is to use it on a phone with PdDroidParty).
So here's the question : is there any latency when writing or reading to a file, and if yes, is it measurable and foreseeable?
Thanks!
Pierre
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Ok, so there should be no latency linked to writesf~.
According to it's page in pdpedia, it seems like sfplay~ is the only playback object that makes it possible to do multitrack recording and overdubbing. I'm wondering why...
Thinking about it now, it looks like it should be a problem with the upstream latency before adc~. And i guess the only way i can fix this is by playing track (n-1) with a delay with respect to track n equal to this latency (since time travel doesn't exist yet).
Pierre
2011/9/12 hardoff goes bananas hard.off@gmail.com
writesf~ creates a subthread whose task is to write audio streams to disk. You need not provide any disk access time between "open" and "start", but between "stop" and the next "open" you must give the object time to flush all the output to disk.
(from writesf~ help)
On Mon, Sep 12, 2011 at 4:12 AM, Pierre Massat pimassat@gmail.com wrote:
I forgot to mention that i can't use sfplay~ since PdDroidParty doesn't support externals...
2011/9/11 Pierre Massat pimassat@gmail.com
Hi,
I m trying to write a multitrack recorder based on writesf~ and readsf~ . I m faced with a serious issue... There seems to be some latency either when writing to a file or reading from it (i can't tell). This means that i can't record a second track on top of another as an overdub and playback both. Assuming that Pd's diplayed latency in the audio settings is right (???), that is 100ms, i tried triggering the readsf~ after a 100 ms-long delay, to no avail. I thought it would be a piece of cake to program this, but well, apparently it's not...? Needless to say that i can't use table as i plan to use it for possibly long recordings (my goal is to use it on a phone with PdDroidParty).
So here's the question : is there any latency when writing or reading to a file, and if yes, is it measurable and foreseeable?
Thanks!
Pierre
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
I solved the problem by delaying the previous track during playback by a duration equal to my setup's latency. This doesn't make my multitrack recorder very portable at all, since i need to measure the latency beforehand... :(
Anyways...
Pierre
2011/9/12 Pierre Massat pimassat@gmail.com
Ok, so there should be no latency linked to writesf~.
According to it's page in pdpedia, it seems like sfplay~ is the only playback object that makes it possible to do multitrack recording and overdubbing. I'm wondering why...
Thinking about it now, it looks like it should be a problem with the upstream latency before adc~. And i guess the only way i can fix this is by playing track (n-1) with a delay with respect to track n equal to this latency (since time travel doesn't exist yet).
Pierre
2011/9/12 hardoff goes bananas hard.off@gmail.com
writesf~ creates a subthread whose task is to write audio streams to disk. You need not provide any disk access time between "open" and "start", but between "stop" and the next "open" you must give the object time to flush all the output to disk.
(from writesf~ help)
On Mon, Sep 12, 2011 at 4:12 AM, Pierre Massat pimassat@gmail.comwrote:
I forgot to mention that i can't use sfplay~ since PdDroidParty doesn't support externals...
2011/9/11 Pierre Massat pimassat@gmail.com
Hi,
I m trying to write a multitrack recorder based on writesf~ and readsf~ . I m faced with a serious issue... There seems to be some latency either when writing to a file or reading from it (i can't tell). This means that i can't record a second track on top of another as an overdub and playback both. Assuming that Pd's diplayed latency in the audio settings is right (???), that is 100ms, i tried triggering the readsf~ after a 100 ms-long delay, to no avail. I thought it would be a piece of cake to program this, but well, apparently it's not...? Needless to say that i can't use table as i plan to use it for possibly long recordings (my goal is to use it on a phone with PdDroidParty).
So here's the question : is there any latency when writing or reading to a file, and if yes, is it measurable and foreseeable?
Thanks!
Pierre
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-
hardoff goes bananas -
Pierre Massat