Hello,
Has anyone got any experience with [array quantile]?
I'm getting some strange results and I've done everything I can think of.
I've attached a patch that should clarify the problem. Basically, sometimes, not always, [array quantile] returns some weird numbers that I can't explain.
And a related issue: array random doesn't seem to be doing what it should be doing. It returns very different values, compared with quantile fed with random values between 0 and 1. Again, there is an example of what I mean in the attached patch.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
I think the patch is occasionally (once every 512 times on average) sending zero to "array quantile" which then outputs the index of the first nonzero number in the table -- in this case a point with probability about 1e-45.
Maybe try random 1e8 (or so) and divide by 1e8 to get a more continuous, less grainy random sample out of the array.
cheers Miller
On Fri, Nov 22, 2013 at 01:51:19AM +0000, peiman khosravi wrote:
Hello,
Has anyone got any experience with [array quantile]?
I'm getting some strange results and I've done everything I can think of.
I've attached a patch that should clarify the problem. Basically, sometimes, not always, [array quantile] returns some weird numbers that I can't explain.
And a related issue: array random doesn't seem to be doing what it should be doing. It returns very different values, compared with quantile fed with random values between 0 and 1. Again, there is an example of what I mean in the attached patch.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Thank you very much. That does the trick.
Though, [array random] doesn't seem to be working as I expected. Am I right to think that it should produce the same result as [array quantile] fed with uniformly distributed random values? If so I'm getting very different results here.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 22 November 2013 04:16, Miller Puckette msp@ucsd.edu wrote:
I think the patch is occasionally (once every 512 times on average) sending zero to "array quantile" which then outputs the index of the first nonzero number in the table -- in this case a point with probability about 1e-45.
Maybe try random 1e8 (or so) and divide by 1e8 to get a more continuous, less grainy random sample out of the array.
cheers Miller
On Fri, Nov 22, 2013 at 01:51:19AM +0000, peiman khosravi wrote:
Hello,
Has anyone got any experience with [array quantile]?
I'm getting some strange results and I've done everything I can think of.
I've attached a patch that should clarify the problem. Basically, sometimes, not always, [array quantile] returns some weird numbers that I can't explain.
And a related issue: array random doesn't seem to be doing what it should be doing. It returns very different values, compared with quantile fed
with
random values between 0 and 1. Again, there is an example of what I mean
in
the attached patch.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
There could be something wrong. But array_random_bang() (in x_array.c) cooks up a pseudorandom number from 0 to 1 (I believe) and then calls array_quantile_float() with it. That's exactly what connecting random() to array_quantile in a patch should be doing.
cheers Miller
On Fri, Nov 22, 2013 at 09:38:59AM +0000, peiman khosravi wrote:
Thank you very much. That does the trick.
Though, [array random] doesn't seem to be working as I expected. Am I right to think that it should produce the same result as [array quantile] fed with uniformly distributed random values? If so I'm getting very different results here.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 22 November 2013 04:16, Miller Puckette msp@ucsd.edu wrote:
I think the patch is occasionally (once every 512 times on average) sending zero to "array quantile" which then outputs the index of the first nonzero number in the table -- in this case a point with probability about 1e-45.
Maybe try random 1e8 (or so) and divide by 1e8 to get a more continuous, less grainy random sample out of the array.
cheers Miller
On Fri, Nov 22, 2013 at 01:51:19AM +0000, peiman khosravi wrote:
Hello,
Has anyone got any experience with [array quantile]?
I'm getting some strange results and I've done everything I can think of.
I've attached a patch that should clarify the problem. Basically, sometimes, not always, [array quantile] returns some weird numbers that I can't explain.
And a related issue: array random doesn't seem to be doing what it should be doing. It returns very different values, compared with quantile fed
with
random values between 0 and 1. Again, there is an example of what I mean
in
the attached patch.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Thanks. In this case I think something isn't right with [array random]. Using the same array, I get very different patterns with the two methods (see attached patch). The difference is obvious with a gaussian distribution, which looks skewed when generated with [array random].
Best, Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 22 November 2013 15:48, Miller Puckette msp@ucsd.edu wrote:
There could be something wrong. But array_random_bang() (in x_array.c) cooks up a pseudorandom number from 0 to 1 (I believe) and then calls array_quantile_float() with it. That's exactly what connecting random() to array_quantile in a patch should be doing.
cheers Miller
On Fri, Nov 22, 2013 at 09:38:59AM +0000, peiman khosravi wrote:
Thank you very much. That does the trick.
Though, [array random] doesn't seem to be working as I expected. Am I
right
to think that it should produce the same result as [array quantile] fed with uniformly distributed random values? If so I'm getting very
different
results here.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 22 November 2013 04:16, Miller Puckette msp@ucsd.edu wrote:
I think the patch is occasionally (once every 512 times on average)
sending
zero to "array quantile" which then outputs the index of the first
nonzero
number in the table -- in this case a point with probability about
1e-45.
Maybe try random 1e8 (or so) and divide by 1e8 to get a more
continuous,
less grainy random sample out of the array.
cheers Miller
On Fri, Nov 22, 2013 at 01:51:19AM +0000, peiman khosravi wrote:
Hello,
Has anyone got any experience with [array quantile]?
I'm getting some strange results and I've done everything I can
think of.
I've attached a patch that should clarify the problem. Basically, sometimes, not always, [array quantile] returns some weird numbers
that I
can't explain.
And a related issue: array random doesn't seem to be doing what it
should
be doing. It returns very different values, compared with quantile
fed
with
random values between 0 and 1. Again, there is an example of what I
mean
in
the attached patch.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS
Feed
http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
Aha and bngo! In effect - array random is only looking at the lower 1/2 of the distribution. I made a stupid C data type fumble in the code.
I recently tripped over a bug, too, in "text set" - will attempt to fix them both and issue an updated pd-0.45 in the next day or 2.
Thanks for flagging this!
Miller
On Fri, Nov 22, 2013 at 03:56:54PM +0000, peiman khosravi wrote:
Thanks. In this case I think something isn't right with [array random]. Using the same array, I get very different patterns with the two methods (see attached patch). The difference is obvious with a gaussian distribution, which looks skewed when generated with [array random].
Best, Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 22 November 2013 15:48, Miller Puckette msp@ucsd.edu wrote:
There could be something wrong. But array_random_bang() (in x_array.c) cooks up a pseudorandom number from 0 to 1 (I believe) and then calls array_quantile_float() with it. That's exactly what connecting random() to array_quantile in a patch should be doing.
cheers Miller
On Fri, Nov 22, 2013 at 09:38:59AM +0000, peiman khosravi wrote:
Thank you very much. That does the trick.
Though, [array random] doesn't seem to be working as I expected. Am I
right
to think that it should produce the same result as [array quantile] fed with uniformly distributed random values? If so I'm getting very
different
results here.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 22 November 2013 04:16, Miller Puckette msp@ucsd.edu wrote:
I think the patch is occasionally (once every 512 times on average)
sending
zero to "array quantile" which then outputs the index of the first
nonzero
number in the table -- in this case a point with probability about
1e-45.
Maybe try random 1e8 (or so) and divide by 1e8 to get a more
continuous,
less grainy random sample out of the array.
cheers Miller
On Fri, Nov 22, 2013 at 01:51:19AM +0000, peiman khosravi wrote:
Hello,
Has anyone got any experience with [array quantile]?
I'm getting some strange results and I've done everything I can
think of.
I've attached a patch that should clarify the problem. Basically, sometimes, not always, [array quantile] returns some weird numbers
that I
can't explain.
And a related issue: array random doesn't seem to be doing what it
should
be doing. It returns very different values, compared with quantile
fed
with
random values between 0 and 1. Again, there is an example of what I
mean
in
the attached patch.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS
Feed
http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
Thanks for looking into it so quickly.
all the best, Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 22 November 2013 16:11, Miller Puckette msp@ucsd.edu wrote:
Aha and bngo! In effect - array random is only looking at the lower 1/2 of the distribution. I made a stupid C data type fumble in the code.
I recently tripped over a bug, too, in "text set" - will attempt to fix them both and issue an updated pd-0.45 in the next day or 2.
Thanks for flagging this!
Miller
On Fri, Nov 22, 2013 at 03:56:54PM +0000, peiman khosravi wrote:
Thanks. In this case I think something isn't right with [array random]. Using the same array, I get very different patterns with the two methods (see attached patch). The difference is obvious with a gaussian distribution, which looks skewed when generated with [array random].
Best, Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 22 November 2013 15:48, Miller Puckette msp@ucsd.edu wrote:
There could be something wrong. But array_random_bang() (in x_array.c) cooks up a pseudorandom number from 0 to 1 (I believe) and then calls array_quantile_float() with it. That's exactly what connecting
random()
to array_quantile in a patch should be doing.
cheers Miller
On Fri, Nov 22, 2013 at 09:38:59AM +0000, peiman khosravi wrote:
Thank you very much. That does the trick.
Though, [array random] doesn't seem to be working as I expected. Am I
right
to think that it should produce the same result as [array quantile]
fed
with uniformly distributed random values? If so I'm getting very
different
results here.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS
Feed
http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 22 November 2013 04:16, Miller Puckette msp@ucsd.edu wrote:
I think the patch is occasionally (once every 512 times on average)
sending
zero to "array quantile" which then outputs the index of the first
nonzero
number in the table -- in this case a point with probability about
1e-45.
Maybe try random 1e8 (or so) and divide by 1e8 to get a more
continuous,
less grainy random sample out of the array.
cheers Miller
On Fri, Nov 22, 2013 at 01:51:19AM +0000, peiman khosravi wrote:
Hello,
Has anyone got any experience with [array quantile]?
I'm getting some strange results and I've done everything I can
think of.
I've attached a patch that should clarify the problem. Basically, sometimes, not always, [array quantile] returns some weird
numbers
that I
can't explain.
And a related issue: array random doesn't seem to be doing what
it
should
be doing. It returns very different values, compared with
quantile
fed
with
random values between 0 and 1. Again, there is an example of
what I
mean
in
the attached patch.
Thanks Peiman
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk ||
RSS
Feed
http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
My pleasure ..... I want the stuff to work after all.
I don't know if anyone else is using the new stuff so didn't throw it on pd-announce, but I've quietly poseted a bug-fix release on my page (0.45-4, msp.ucsd.edu).
cheers Miller
On Sat, Nov 23, 2013 at 09:36:30AM +0000, peiman khosravi wrote:
Thanks for looking into it so quickly.
all the best, Peiman
Great, thanks very much. I can't find it here though: http://msp.ucsd.edu/software.html
I've become addicted to this new stuff!
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 23 November 2013 18:34, Miller Puckette msp@ucsd.edu wrote:
My pleasure ..... I want the stuff to work after all.
I don't know if anyone else is using the new stuff so didn't throw it on pd-announce, but I've quietly poseted a bug-fix release on my page (0.45-4, msp.ucsd.edu).
cheers Miller
On Sat, Nov 23, 2013 at 09:36:30AM +0000, peiman khosravi wrote:
Thanks for looking into it so quickly.
all the best, Peiman
Oops, my bad. Should be up now.
M
On Sun, Nov 24, 2013 at 03:01:58AM +0000, peiman khosravi wrote:
Great, thanks very much. I can't find it here though: http://msp.ucsd.edu/software.html
I've become addicted to this new stuff!
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 23 November 2013 18:34, Miller Puckette msp@ucsd.edu wrote:
My pleasure ..... I want the stuff to work after all.
I don't know if anyone else is using the new stuff so didn't throw it on pd-announce, but I've quietly poseted a bug-fix release on my page (0.45-4, msp.ucsd.edu).
cheers Miller
On Sat, Nov 23, 2013 at 09:36:30AM +0000, peiman khosravi wrote:
Thanks for looking into it so quickly.
all the best, Peiman
Got it, thanks.
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 24 November 2013 04:07, Miller Puckette msp@ucsd.edu wrote:
Oops, my bad. Should be up now.
M
On Sun, Nov 24, 2013 at 03:01:58AM +0000, peiman khosravi wrote:
Great, thanks very much. I can't find it here though: http://msp.ucsd.edu/software.html
I've become addicted to this new stuff!
*www.peimankhosravi.co.uk http://www.peimankhosravi.co.uk || RSS Feed http://peimankhosravi.co.uk/miscposts.rss || Concert News http://spectralkimia.wordpress.com/*
On 23 November 2013 18:34, Miller Puckette msp@ucsd.edu wrote:
My pleasure ..... I want the stuff to work after all.
I don't know if anyone else is using the new stuff so didn't throw it
on
pd-announce, but I've quietly poseted a bug-fix release on my page (0.45-4, msp.ucsd.edu).
cheers Miller
On Sat, Nov 23, 2013 at 09:36:30AM +0000, peiman khosravi wrote:
Thanks for looking into it so quickly.
all the best, Peiman
-
Miller Puckette -
peiman khosravi