Transcribe~ Release
Hello list,
As my first post to this list, I would like to announce the release
of [transcribe~], a real-time polyphonic audio to Midi transcriber,
for Pd and MaxMSP.
Suggestions and bug reports are very welcome!
Cheers,
Arshia Cont
http://cosmal.ucsd.edu/arshia/
Thanks Arshia, that sounds great.
BTW it's nice to see you're still alive :-)
Peiman
On 25/09/2007, Arshia Cont wrote:
>
> Hello list,
>
> As my first post to this list, I would like to announce the release
> of [transcribe~], a real-time polyphonic audio to Midi transcriber,
> for Pd and MaxMSP.
>
> http://cosmal.ucsd.edu/arshia/transcribe/
>
> Suggestions and bug reports are very welcome!
>
> Cheers,
>
>
> Arshia Cont
> http://cosmal.ucsd.edu/arshia/
>
>
>
>
>
This looks pretty good.
I'm having a little trouble with getting it to work correctly. The file
seems to play back at half the spped and the sounds degraded somewhat,
like as if the sample and bit rate have been changed without any
dithering (if that's the term).
It maybe just me...I'm not the most savvy when it comes to the absolute
nitty gritty of computer processing.
Is there a need for pluggo to work with this patch?
Kel
On 26/09/2007, at 4:30 PM, peiman khosravi wrote:
> Thanks Arshia, that sounds great.
>
> BTW it's nice to see you're still alive :-)
>
> Peiman
>
> On 25/09/2007, Arshia Cont wrote:
>>
>> As my first post to this list, I would like to announce the release
>> of [transcribe~], a real-time polyphonic audio to Midi transcriber,
>> for Pd and MaxMSP.
>>
>> http://cosmal.ucsd.edu/arshia/transcribe/
>>
>> Suggestions and bug reports are very welcome!
>>
>> Cheers,
>>
>>
>> Arshia Cont
>> http://cosmal.ucsd.edu/arshia/
>>
>>
>>
>>
Hi Kelly,
The source of "CPU distortion" is a result of CPU overload. It might
come from several sources that you need to check:
* Make sure you have a reasonable IO Vector size and Signal Vector
Size [in DSP Status]. Do not use VERY SMALL [i.e. 8 for ex.] signal
vector size or IO Vector Size (I suggest the normal 512/64 or 1024/64
combinations). The story briefly is that the processor should have
enough time before the next computation to do finish up the previous
calculation. Small vector sizes evade that.
* Look at the CPU status (included in the help patch) while running.
If it is too high, then try using a different hopsize (e.g. 1024 or
2048 instead of 512 -- second argument of the object).
* If the CPU percentage is still too high, then I suspect you are not
on a dual core machine or you don't have much RAM.
You don't need Plugo for this.
Let me know... .
Arshia Cont
http://cosmal.ucsd.edu/arshia/
Kelly wrote:
>I'm having a little trouble with getting it to work correctly. The
file seems to play back at half the spped and the sounds degraded
somewhat, like as >if the sample and bit rate have been changed
without any dithering (if that's the term).
>
> * If the CPU percentage is still too high, then I suspect you are not
> on a dual core machine or you don't have much RAM.
I'm with you on third point: I don't have a dual core processor and I'm
on 1Gb of RAM. It's looking like I'm going to need more....MORE POWER!!
Thanks, I'm looking forward to getting this working.
Kel
Hi again Kel,
As the project website suggests, the algorithm needs atleast a dual
core machine! On most dual core machines it does not take more than
30% of the CPU while running on a heavy orchestral input. In the next
coming versions I'll optimize it more and more.. .
On 26 sept. 07, at 16:58, Kelly Ottaway wrote:
>
>>
>> * If the CPU percentage is still too high, then I suspect you are
>> not on a dual core machine or you don't have much RAM.
>
> I'm with you on third point: I don't have a dual core processor and
> I'm on 1Gb of RAM. It's looking like I'm going to need more....MORE
> POWER!!
>
> Thanks, I'm looking forward to getting this working.
>
> Kel
>
>
Arshia Cont
http://cosmal.ucsd.edu/arshia/
Arshia Cont schrieb:
> As my first post to this list, I would like to announce the release of
> [transcribe~], a real-time polyphonic audio to Midi transcriber, for Pd
> and MaxMSP.
Thanks a lot for that work, indeed very interesting (though my Powerbook
is too slow... :-(
I was lurking on your web site and just wonder what the picture of the
factor oracle analysis of Charly Parker is actually showing. Is this a
visualisation of the sound or of the score?
When searching for factor oracle I found references to text searching
algorithms but not to musical applications...
Stefan
--
Stefan Tiedje------------x-------
--_____-----------|--------------
--(_|_ ----|-----|-----()-------
-- _|_)----|-----()--------------
----------()--------www.ccmix.com
Is this still around? the links seem to be broken :(
You also might try to contact Arshia Cont directly. (If he's not getting email alerts on the thread)
link still seems to be down. I'm also interested in trying out transcribe~ for a project.
Does anyone know, were there any guitar templates made?