[CLAM] Inquiry about chordextractor output file

[Keiichiro Hoashi]

Thanks :-)

We're encountering another problem w/ chordextractor. When inputting 
relatively long MP3s, the program sometimes stops w/ this error message:

-----
At file /home/vokimon/CajitasDeArena/mingw/local/include/CLAM/Array.hxx
line 124
Array::Resize(): You cannot invoke this method on an array that does not
own any memory
Unable to adquire symbols names for the backtrace
-----

Is this a known problem? If so, how can we avoid this?

Regards,

KH


(2011/01/10 18:19), David García Garzón wrote:
> Fully right. :-)
>
> A Dijous 06 Gener 2011 13:31:49, Keiichiro Hoashi va escriure:
>> David,
>>
>> Thanks for your prompt reply.
>>
>> Let us check a little more detail about the XML file..
>>
>> Here's a snippet from an XML file that was output by the chordextractor:
>>
>> ---
>> <AttributePool name="Chords_Harte" size="1028">0.464399 0.835918
>> 0.835918 1.48608 1.48608 1.67184 1.67184 3.43655 3.43655 3.52943 ...
>> </AttributePool>
>> ...
>> <ScopePool name="ExtractedChord" size="514">
>>       <AttributePool name="Root">
>>         <Enumerated>G</Enumerated>
>>         <Enumerated>C</Enumerated>
>>         ...
>>       </AttributePool>
>>       <AttributePool name="Mode">
>>         <Enumerated>Major</Enumerated>
>>         <Enumerated>Major</Enumerated>
>>       </AttributePool>
>>       </AttributePool>
>>     </ScopePool>
>> ---
>>
>> Based on our analysis, we're assuming that the above "Chords_Harte"
>> values indicate the start and end times (in secs) of each "segment".
>> Furthermore, we're also assuming that the chord extraction results
>> written in<AttributePool name="Root">  and<AttributePool name="Mode">
>> express the chords of each extracted "segment". For example, the first
>> segment of the above input song is from 0.464399sec ~ 0.0835918sec, and
>> its estimated chord is "G" + "Major". Is this understanding correct?
>>
>> Sincerely,
>>
>> KH
>>
>> (2011/01/05 23:51), David García Garzón wrote:
>>> Not about the XML format itself but about the concepts used by the format
>>> (Segmentations, FrameDivisions, Attributes and Scopes)...
>>> http://clam-project.org/wiki/Manual_Annotator
>>>
>>> In short: Attributes are the data holders. They hold a value for each
>>> element of a given scope (each song, frame, or segment). There is a top
>>> level scope (Song) and then some special attributes define new scopes
>>> (FrameDivisions and Segmentations).
>>>
>>> You have two xml files:
>>> - description scheme: defines a collection of attributes and their
>>> properties: type, name, scope, documentation...
>>> - description data: holds the attribute values for a given song.
>>>
>>> At the end of the NetworkEditor tutorial there is an explanation of the
>>> data you can find in that concrete example:
>>> http://clam-project.org/wiki/Network_Editor_tutorial#Tonal_Analysis
>>>
>>> Also Chordata provides an interactive interface for all that:
>>> http://clam-project.org/wiki/Chordata_tutorial
>>>
>>> On Wednesday 05 January 2011 09:18:26 Keiichiro Hoashi wrote:
>>>> Hi,
>>>>
>>>> Are there any documentations which explain about the format of the XML
>>>> output file of chordextractor.exe?
>>>>
>>>> We have just started using CLAM a while ago, and are struggling to find
>>>> relevant info from the CLAM web page, wikis, etc...
>>>>
>>>> Thanks in advance for any advice,
>>>>
>>>> KH
>>>
>>> _______________________________________________
>>> clam-users mailing list
>>> clam-users at lists.clam-project.org
>>> http://clam-project.org
>


-- 
====---====---====---====---====---
Keiichiro Hoashi, Dr. Eng.
Intelligent Media Processing Lab,
KDDI R&D Laboratories, Inc.
e-mail: hoashi[at]kddilabs.jp
----===----===----===----===----===