Hello,
What's the best way to convert a signal into a message? There are possibilities like snapshot~ or vsnapshot~, but they are not accurate. The error they produce when you convert into a message and reconvert to a signal (with sig~, line~ or vline~) is clearly audible. ciao
Wolfgang
On Sun, 2009-05-10 at 17:29 +0200, Wolfgang Jäger wrote:
Hello,
What's the best way to convert a signal into a message? There are possibilities like snapshot~ or vsnapshot~, but they are not accurate. The error they produce when you convert into a message and reconvert to a signal (with sig~, line~ or vline~) is clearly audible-
yeah, because you most likely capture not _all_ samples and convert them to messages. in order to do so, you would have to use [vsnapshot~] and bang it 44.1 times per milliseconds. [snapshot~] wouldn't work, since it always captures the first (or last?) sample of the block.
an easier way is to use [pack~] and [unpack~] from zexy. those convert a signal vector (one block of audio samples) to a list of floats and vice versa. if you want to have a stream of floats instead of lists, use [drip].
either way you do it, it's going to be expensive.
roman
Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de
well, the question is for what reason are you doing it? I think it is better to get some features from say sigmund~ and use those to control or recreate, but i don't know what you want....
J
On Sun, May 10, 2009 at 8:54 AM, Roman Haefeli reduzierer@yahoo.de wrote:
On Sun, 2009-05-10 at 17:29 +0200, Wolfgang Jäger wrote:
Hello,
What's the best way to convert a signal into a message? There are possibilities like snapshot~ or vsnapshot~, but they are not accurate. The error they produce when you convert into a message and reconvert to a signal (with sig~, line~ or vline~) is clearly audible-
yeah, because you most likely capture not _all_ samples and convert them to messages. in order to do so, you would have to use [vsnapshot~] and bang it 44.1 times per milliseconds. [snapshot~] wouldn't work, since it always captures the first (or last?) sample of the block.
an easier way is to use [pack~] and [unpack~] from zexy. those convert a signal vector (one block of audio samples) to a list of floats and vice versa. if you want to have a stream of floats instead of lists, use [drip].
either way you do it, it's going to be expensive.
roman
Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Jaime Oliver wrote:
well, the question is for what reason are you doing it?
I want to send audio data packed in OSC messages/bundles over Ethernet.
I think it is better to get some features from say sigmund~ and use those to control or recreate, but i don't know what you want....
I guess in my case it's better to work with exact audio data, an estimation would limit applications on the receiver side.
Wolfgang
J
On Sun, May 10, 2009 at 8:54 AM, Roman Haefeli reduzierer@yahoo.de wrote:
On Sun, 2009-05-10 at 17:29 +0200, Wolfgang Jäger wrote:
Hello,
What's the best way to convert a signal into a message?
On May 10, 2009, at 5:40 PM, Wolfgang Jäger wrote:
Jaime Oliver wrote:
well, the question is for what reason are you doing it?
I want to send audio data packed in OSC messages/bundles over
Ethernet.
Whoa, unless you really want OSC, I think there are better ways to do
this. Check out streamin~/streamout~ for example.
.hc
I think it is better to get some features from say sigmund~ and use those to control or recreate, but i don't know what you want....
I guess in my case it's better to work with exact audio data, an estimation would limit applications on the receiver side.
Wolfgang
J
On Sun, May 10, 2009 at 8:54 AM, Roman Haefeli
reduzierer@yahoo.de wrote:On Sun, 2009-05-10 at 17:29 +0200, Wolfgang Jäger wrote:
Hello,
What's the best way to convert a signal into a message?
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
The arc of history bends towards justice. - Dr. Martin Luther
King, Jr.
--- On Sun, 5/10/09, Roman Haefeli reduzierer@yahoo.de wrote:
From: Roman Haefeli reduzierer@yahoo.de Subject: Re: [PD] signal / message To: "Wolfgang Jäger" wolfgang.jaeger@gmx.at Cc: pd-list@iem.at Date: Sunday, May 10, 2009, 5:54 PM On Sun, 2009-05-10 at 17:29 +0200, Wolfgang Jäger wrote:
Hello,
What's the best way to convert a signal into a
message?
There are possibilities like snapshot~ or vsnapshot~,
but they are not
accurate. The error they produce when you convert into
a message and
reconvert to a signal (with sig~, line~ or vline~) is
clearly audible-
yeah, because you most likely capture not _all_ samples and convert them to messages. in order to do so, you would have to use [vsnapshot~] and bang it 44.1 times per milliseconds. [snapshot~] wouldn't work, since it always captures the first (or last?) sample of the block.
Isn't it the last sample of the block? (see attached)
-Jonathan
Isn't it the last sample of the _last_ block? I heard that somewhere.
So if you're processing block N on the current DSP tick what you see is the [snapshot~] of N-1 ???
On Sun, 10 May 2009 13:28:01 -0700 (PDT) Jonathan Wilkes jancsika@yahoo.com wrote:
Isn't it the last sample of the block? (see attached)
-Jonathan
--- On Sun, 5/10/09, Andy Farnell padawan12@obiwannabe.co.uk wrote:
From: Andy Farnell padawan12@obiwannabe.co.uk Subject: Re: [PD] signal / message To: "Jonathan Wilkes" jancsika@yahoo.com Cc: wolfgang.jaeger@gmx.at, reduzierer@yahoo.de, pd-list@iem.at Date: Sunday, May 10, 2009, 9:57 PM Isn't it the last sample of the _last_ block? I heard that somewhere.
So if you're processing block N on the current DSP tick what you see is the [snapshot~] of N-1 ???
Thanks for the clarification. You must be right; otherwise, snapshot~ would have to delay its output until the current block finishes.
-Jonathan
On Sun, 10 May 2009 13:28:01 -0700 (PDT) Jonathan Wilkes jancsika@yahoo.com wrote:
Isn't it the last sample of the block? (see
attached)
-Jonathan
-- Use the source
On Sun, 2009-05-10 at 13:28 -0700, Jonathan Wilkes wrote:
--- On Sun, 5/10/09, Roman Haefeli reduzierer@yahoo.de wrote:
From: Roman Haefeli reduzierer@yahoo.de Subject: Re: [PD] signal / message To: "Wolfgang JÀger" wolfgang.jaeger@gmx.at Cc: pd-list@iem.at Date: Sunday, May 10, 2009, 5:54 PM On Sun, 2009-05-10 at 17:29 +0200, Wolfgang JÀger wrote:
Hello,
What's the best way to convert a signal into a
message?
There are possibilities like snapshot~ or vsnapshot~,
but they are not
accurate. The error they produce when you convert into
a message and
reconvert to a signal (with sig~, line~ or vline~) is
clearly audible-
yeah, because you most likely capture not _all_ samples and convert them to messages. in order to do so, you would have to use [vsnapshot~] and bang it 44.1 times per milliseconds. [snapshot~] wouldn't work, since it always captures the first (or last?) sample of the block.
Isn't it the last sample of the block? (see attached)
-Jonathan
i guess, you're right. thanks for the patch.
roman
___________________________________________________________ Der frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: http://mail.yahoo.de
Hallo, Roman Haefeli hat gesagt: // Roman Haefeli wrote:
an easier way is to use [pack~] and [unpack~] from zexy. those convert a signal vector (one block of audio samples) to a list of floats and vice versa. if you want to have a stream of floats instead of lists, use [drip].
Or use [tabsend~] and read out the table between dsp cycles.
Ciao
Frank
Frank Barknecht wrote:
Hallo, Roman Haefeli hat gesagt: // Roman Haefeli wrote:
an easier way is to use [pack~] and [unpack~] from zexy. those convert a signal vector (one block of audio samples) to a list of floats and vice versa. if you want to have a stream of floats instead of lists, use [drip].
Or use [tabsend~] and read out the table between dsp cycles.
Ciao
Thanks a lot, I'm definitly going to try this, even if my computer seems to be a little bit too slow.
Wolfgang
-
Andy Farnell -
Frank Barknecht -
Hans-Christoph Steiner -
Jaime Oliver -
Jonathan Wilkes -
Roman Haefeli -
Wolfgang Jäger