Thanks for this hint Oliver!

Can't test right now but am curious if audio drops out (eg. playing a sine test tone while copying some 10secs of table data) and if there is any (dis)advantages over the get/set method below.

cheersz, P

• oliver@klingt.org oliver@klingt.org [2024-04-23 10:22]:

Also just found out by chance that writing and reading to/from clipboard also works !

[write clipboard( | [array define a1]

[read clipboard( | [array define a2]

Can you confirm this ?

Best

Oliver

Am 23. April 2024 10:15:00 MESZ schrieb "Peter P." peterparker@fastmail.com:

...

Replying to myself here, after this solution has been suggested to me off-list:

[bang(
|
[array get one]
|
[array set two]

seems to work for larger tables without audio dropouts.

Much appreciated help, thanks! P

• Peter P. peterparker@fastmail.com [2024-04-23 08:54]:

...

Hi list,

I am sure this has been asked many times, but I can't find a solution in the mailing list archive. How can I copy data between arrays, in my case arrays of 480000 samples without using externals and without blocking the DSP computation (causing dropouts)? I know that I can write to disk and read from it again, but is there a way that does not require disk access?

Thanks for all ideas! Peter

---

Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list

--

Gesendet von meinem Endgerät

---

Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list

If you need 'instantaneous' (zero logical time) copy, then [array get] and [array set] are probably a good option, but even though this will not stop the DSP, it may well cause an audio dropout, typically in the case of large copies on a system with small blocksize and small buffer size and non-trivial CPU usage. These factors all interact in determining how much headroom (in terms of CPU time) there is for a copy to happen within a single audio callback without causing a buffer underrun. If you need a real-time friendly[1] copy, consider using a pair of tabread~ / tabwrite~. In this case, if you need to copy faster than real time[2] (i.e.: copy one second worth of data in less than one second), you should be able to do that in an oversampled subpatch(via [block~]).

[1]: real-time friendly in the sense of the very loose definition of an operation that completes in a finite amount of time and that has an (almost) constant CPU time for each audio callback. [2]: real time as in real-world time in seconds given by arraySize/samplingRate

Benjamin Wesch wrote on 23/04/2024 10:38:

hi,

to avoid confusion here: the "clipboard" in this case is actually a file. it certainly works to write and read arrays to/from disk like that (as documented in "other-messages" subpatch of the [array define] help) - but this is not a RAM copy/paste as it might seem. :)

cheers, ben

Am Di., 23. Apr. 2024 um 10:29 Uhr schrieb Peter P. peterparker@fastmail.com:

...

Thanks for this hint Oliver!

Can't test right now but am curious if audio drops out (eg. playing a sine test tone while copying some 10secs of table data) and if there is any (dis)advantages over the get/set method below.

cheersz, P

• oliver@klingt.org oliver@klingt.org [2024-04-23 10:22]:

...

Also just found out by chance that writing and reading to/from clipboard also works !

[write clipboard( | [array define a1]

[read clipboard( | [array define a2]

Can you confirm this ?

Best

Oliver

Am 23. April 2024 10:15:00 MESZ schrieb "Peter P." peterparker@fastmail.com:

...

Replying to myself here, after this solution has been suggested to me off-list:

[bang( | [array get one] | [array set two]

seems to work for larger tables without audio dropouts.

Much appreciated help, thanks! P

• Peter P. peterparker@fastmail.com [2024-04-23 08:54]:

...

Hi list,

I am sure this has been asked many times, but I can't find a solution in the mailing list archive. How can I copy data between arrays, in my case arrays of 480000 samples without using externals and without blocking the DSP computation (causing dropouts)? I know that I can write to disk and read from it again, but is there a way that does not require disk access?

Thanks for all ideas! Peter

---

Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list

--

Gesendet von meinem Endgerät

...

---

Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list

---

Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list

---

Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list