mpeg7audioenc-devel Mailing List for MPEG-7 Audio Encoder
Status: Beta
Brought to you by:
holger_c
Menu
▾
▴
mpeg7audioenc-devel — List for developers, testers and anyone that can help in development
You can subscribe to this list here.
| 2003 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(10) |
Nov
(8) |
Dec
(5) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2004 |
Jan
|
Feb
(1) |
Mar
|
Apr
|
May
(3) |
Jun
(2) |
Jul
(2) |
Aug
|
Sep
|
Oct
(1) |
Nov
|
Dec
|
| 2005 |
Jan
(1) |
Feb
(1) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2006 |
Jan
|
Feb
(2) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Holger C. <cry...@ie...> - 2006-02-15 11:13:01
|
Hello developers, I am currently working on the next release. For the next version some things needs be finished and I am working on that: - Replace GPL text in every file by LGPL text - Include AudioFundamentalFrequencyType and AudioTempoType output in MP7DocumentBuilder (Currently only supported by MP7Writer) Does someone know other tasks that need to be finished? Holger Holger Crysandt wrote: > Hi, > > I am currently working on a new release. If you have time feel free to > test it. When you have problems or suggestions contact me immediately. > > Holger |
|
From: Holger C. <cry...@ie...> - 2006-02-07 12:20:34
|
Hi, I am currently working on a new release. If you have time feel free to test it. When you have problems or suggestions contact me immediately. Holger |
|
From: Holger C. <cry...@ie...> - 2005-02-25 10:16:04
|
Hello mpeg7audioenc developers,
since the mp3 support has been a problem for months I tried something new:
I added a file called "build.properties". Within this file the path to the
JavaLayer MP3 library can be set (cp.javalayer=...)
During the compilation ("ant jar") ant checks if this is a valid classpath.
Take a look at the messages for more details.
Let me know, if you need additional information.
Have a nice day & weekend
Holger
--
Dipl.-Ing. Holger Crysandt
Institute of Communications Engineering, Aachen University
Melatener Str. 23, D-52074 Aachen
EMail mailto:cry...@ie...
Phone +49 241 80-27676
Fax +49 241 80-22196
Public Key: http://www.ient.rwth-aachen.de/team/crysandt/crysandt.asc
|
|
From: Will N. <wi...@mi...> - 2005-01-12 18:46:48
|
Are there any plans to support the Melody DS parts of MPEG 7 in this tool? Would this feature be feasible to add? Thanks, |
|
From: Holger C. <cry...@ie...> - 2004-10-28 10:23:29
|
Hello Guido, thank you for the file. It has an ID3-tag at the beginning. I added this possibility to my mp3-test. Now the Encoder also works with mp3-files which have the id3 information at the beginning and not at the end. Holger Guido Raparo wrote: > I've created it recorging it with SoundForge and saving it as mp3 |
|
From: Holger C. <cry...@ie...> - 2004-07-06 09:01:42
|
Thank you Guido for supporting the encoder project. I think, we should add multichannel support. But only where needed. IMHO it does not make much sense to enable multichannel support to all descriptors: If someone is interested in the power of the left AND the right channel, he should create 2 encoders (one for each channel) and then merge the two descriptions using XPath for example (javax.xpath.*, since Java 1.5) But for some descriptors like CrossChannelCorrelation it is necessary to have both channels at the same time to compare them. One remark from my side: I made the 10ms Resizer because the first descriptors which I implemented worked all with a segmentation of 10ms. When you have descriptors which don't use the 10ms segmentation forget the Resizer and create your own path through the Encoder and meet at the MP7DocumentBuilder again. Regards Holger Guido Raparo wrote: > Holger and I discussed mutichannel support for the encoder, because I > need it for AudioSignalQuality descriptors(document ISO/IEC > 15938-4:2002/FDAM 1:2003(E)). > > The way to do that can be including new function, new Resizer, new > AudioSignature or bypass the MonoResizer and add a multichannel path: > > : > > 1-new function put(float[], int channels) in the class Encoder,including > informations on channels > > 2-make Resizer and AudioSpectrum support multiple channels or 2) call > put( convertToMono(signal,N)) within the put(float[],N) function and > forget the Resizer and AudioSpectrum and make everything on your own. -- Dipl.-Ing. Holger Crysandt Institute of Communications Engineering, Aachen University Melatener Str. 23, D-52074 Aachen EMail mailto:cry...@ie... Phone +49 241 80-27676 Fax +49 241 80-22196 Public Key: http://www.ient.rwth-aachen.de/team/crysandt/crysandt.asc |
|
From: Guido R. <gui...@li...> - 2004-07-06 08:30:48
|
Holger and I discussed mutichannel support for the encoder, because I = need it for AudioSignalQuality descriptors(document ISO/IEC = 15938-4:2002/FDAM 1:2003(E)). The way to do that can be including new function, new Resizer, new = AudioSignature or bypass the MonoResizer and add a multichannel path: : 1-new function put(float[], int channels) in the class Encoder,including = informations on channels 2-make Resizer and AudioSpectrum support multiple channels or 2) call = put( convertToMono(signal,N)) within the put(float[],N) function and = forget the Resizer and AudioSpectrum and make everything on your own. |
|
From: Holger C. <cry...@ie...> - 2004-06-02 10:02:44
|
Hello Giovanni! You are right. I am working on that memory problem. But (currently) I don't know, who consumes that much memory. What I found out is that converting millions of 32-bit float values to Strings costs some time (especially AudioSpectrumEnvelope and AudioSpectrumFlatness). And most of the memory is consumed during creating the xml-tree from the messages. There is no big difference in memory using de.crysandt.xml.XML or org.w3c.dom.Document. IMHO its the converting function from float32 to String. Collecting the all messages and then creating the xml-tree is NOT the problem. I will let you know when I have new results about the memory problem. Regards Holger Giovanni Tummarello wrote: > Hi holger, great idea! > What about that memory debugging thing, did it yeald to any resoult? Solving the > issue of getting the library to convert larger things (or with more complex > descriptions) should probably be the first priority. > Ciao ciao :-) |
|
From: Holger C. <cry...@ie...> - 2004-06-02 08:24:45
|
Hello developers, thank you for supporting this project! The project is growing and growing. As there were no big changes in the last two weeks, I think we reached a stable point. So I will release version 0.3 at the end of this week. If someone disagrees or needs more time, please let me know! Regards Holger -- Dipl.-Ing. Holger Crysandt Institute of Communications Engineering, Aachen University Melatener Str. 23, D-52074 Aachen EMail cry...@ie... Phone +49 241 80-27676 Fax +49 241 80-22196 |
|
From: <ben...@id...> - 2004-05-25 09:18:49
|
Dear Open Source developer I am doing a research project on "Fun and Software Development" in which I kindly invite you to participate. You will find the online survey under http://fasd.ethz.ch/qsf/. The questionnaire consists of 53 questions and you will need about 15 minutes to complete it. With the FASD project (Fun and Software Development) we want to define the motivational significance of fun when software developers decide to engage in Open Source projects. What is special about our research project is that a similar survey is planned with software developers in commercial firms. This procedure allows the immediate comparison between the involved individuals and the conditions of production of these two development models. Thus we hope to obtain substantial new insights to the phenomenon of Open Source Development. With many thanks for your participation, Benno Luthiger PS: The results of the survey will be published under http://www.isu.unizh.ch/fuehrung/blprojects/FASD/. We have set up the mailing list fa...@we... for this study. Please see http://fasd.ethz.ch/qsf/mailinglist_en.html for registration to this mailing list. _______________________________________________________________________ Benno Luthiger Swiss Federal Institute of Technology Zurich 8092 Zurich Mail: benno.luthiger(at)id.ethz.ch _______________________________________________________________________ |
|
From: Holger C. <cry...@ie...> - 2004-05-12 17:39:14
|
I implemented my idea in the main(...) function of de.crysandt.audio.mpeg7audio.MP7DocumentBuilder. It works fine with Java2, version 1.4.2 AND 1.5 without any additional libraries. Feel free to take a look at it (especially the last lines of the function) Holger Holger Crysandt wrote: > Hello MPEG7AudioEnc-Developers! > > I am currently thinking of creating the XML-tree of the MPEG-7 > description as a "org.w3c.dom.Document" and not as a > "de.crysandt.xml.XMLNode". > > Advantages of "org.w3c.dom.Document": > The Document has the advantage that it is part of the java > implementation. It needs a XML-Parser. But they are included by default > (java2 1.4: crimson; java2 1.5: xerces). But it is also possible to use > other XML-Parser (set property "org.xml.sax.driver"). I have heard > Oracle has its own. With this implementation it is no problem to use > this one if the user has any benefit using it. > > Disadvantages of "org.w3c.dom.Document": > When I first started the project, I was not familiar with XML (Now I am > an expert ;-) ). So I wrote my own XML-Tree which is faster than the > Document but supports much less features. > > IMHO we should change from "de.crysandt.xml.*" to "org.w3c.dom.*". I can > convert the MP7Writer class to MP7DocumentBuilder (I already started). > > Any comments, suggestions, ...? Please let me (us) know. > > Regards > > Holger |
|
From: Holger C. <cry...@ie...> - 2004-05-12 09:01:51
|
Hello MPEG7AudioEnc-Developers! I am currently thinking of creating the XML-tree of the MPEG-7 description as a "org.w3c.dom.Document" and not as a "de.crysandt.xml.XMLNode". Advantages of "org.w3c.dom.Document": The Document has the advantage that it is part of the java implementation. It needs a XML-Parser. But they are included by default (java2 1.4: crimson; java2 1.5: xerces). But it is also possible to use other XML-Parser (set property "org.xml.sax.driver"). I have heard Oracle has its own. With this implementation it is no problem to use this one if the user has any benefit using it. Disadvantages of "org.w3c.dom.Document": When I first started the project, I was not familiar with XML (Now I am an expert ;-) ). So I wrote my own XML-Tree which is faster than the Document but supports much less features. IMHO we should change from "de.crysandt.xml.*" to "org.w3c.dom.*". I can convert the MP7Writer class to MP7DocumentBuilder (I already started). Any comments, suggestions, ...? Please let me (us) know. Regards Holger |
|
From: Holger C. <cry...@ie...> - 2004-02-19 10:30:25
|
Hello MPEG7 Audio Enc Developers!
Since java 1.5 is out I want to make some major changes in the code. Lets talk
about it:
- xerces is now included in the jre by default (com.sun.org.apache.xerces). So
why not export the mpeg7 description as a DOM tree instead of a self made tree?!
- include coding in xml-description "<?xml version="1.0" encoding="UTF-8"?>"
- generic collections (ArrayList<String>). We can start using them when
eclipse supports it. Right now it doesn't make much sense.
- I want to make another release (0.2)
- Change the xml tree to:
<Mpeg7 ...>
<Description xsi:type="ContentEntityType" >
<MultimediaContent xsi:type="AudioType" >
<Audio xsi:type="AudioSegmentType" >
<!-- Lots of descriptors and DS -->
</Audio>
</MultimediaContent>
</Description>
</Mpeg7>
It has one node less than the current structure. The only modules which does
not fit into this structure are Silence and SoundModelDS. SoundModelDS already
has its own structure which is parallel to the old and new structure. (I hope
you understand what I want to say)
Let me know if you (dis-)agree or when you have other ideas
Holger
--
Dipl.-Ing. Holger Crysandt
Institute of Communications Engineering, Aachen University
Melatener Str. 23, D-52074 Aachen
EMail cry...@ie...
Phone +49 241 80-27676
Fax +49 241 80-22196
|
|
From: Holger C. <cry...@ie...> - 2003-12-23 10:50:58
|
Hello developers Take a look at this link: http://130.149.51.3/ It is a web-frontend for the mpeg-7 encoder from the "technical university of berlin" which is written in C++. We should use it to check if we/they create similar results. Holger |
|
From: Giovanni T. <gio...@wu...> - 2003-12-20 11:38:16
|
Hi there, we announce that the whole library now uses "single" as base time. this shoul have no quality problem whatsoever and at the same time it already speeds up operation by over 30% .. and of course memory usage as well. The new optimized fft is used as well although the improovements are not as drammatics as they were expected.. we'll see if we can do better soon. A "database" module is appering that will allow us to have more compact rappresentation of audio mpeg7 data (in a DB like) which should(as first benefit) allow us to have automatic tests for the library Enjoy :-) |
|
From: Holger C. <cry...@ie...> - 2003-12-19 12:53:52
|
Hello developers! I added a new feature: the encoder can be configured with an xml-file. I added a sample file called "mpeg7audioenc.xml". To enable (disable) a descriptor then the enable-value to true(false). By default the value is set to true. If you want to use the configuration file you have to add an xml-parser to your classpath: java -cp xercesImpl.jar:xmlParserAPIs.jar:MPEG7Audioenc.jar de.crysandt.audio.mpeg7audio.Encoder audio.wav mpeg7audioenc.xml Other parsers than xerces are possible but not tested (yet). The other new file (mpeg7audioenc.xsd) is used by the parser to verify whether the config-file is valid or not. Marry X-mas and a happy new year Holger |
|
From: Holger C. <cry...@ie...> - 2003-12-11 09:43:51
|
Hello developers! Version 0.1 of the MPEG-7 Audio Encoder is released. Now it is time to think about the future. - "End of Signal" (eos) Problem: As we needed the eos-information for the AudioSpectrumBasis/-Projection I created a new interface called "Flushable" which contains only one function called flush(). Every MsgSpeaker is an instance of Flushable. When flush() is called at a speaker, the speaker tests for every Listener if the listener is Flushable or not. If the Listener is Flushable it calls the flush() method. So the flush is sent from the resizer to every class in the encoder. If a class needs the eos-information overwrite the flush() method, do what you want to do. DON'T forget to call super.flush() at the end of your new flush() method so every listener gets the information, too. - SoundModelDS, SoundClassificationModelDS, ... Currently I am working on the sound models starting with writing software in Java. It's a challenge to get familiar with the topic (hmm), but the idea behind it seams brilliant to me. I hope I can give you additional information soon. Holger |
|
From: Holger C. <cry...@ie...> - 2003-12-08 11:56:01
|
MPEG-7 developers! Today I released the sources of the "Java MPEG-7 Audio Encoders" at: http://sourceforge.net/projects/mpeg7audioenc/ It is a command line tools is written in Java which and currently extracts the following descriptors and description schemes: - AudioPowerType - AudioWaveformType - AudioSpectrumCentroidType - AudioSpectrumSpreadType - AudioSpectrumEnvelopeType - AudioSpectrumFlatnessType - AudioSpectrumBasisType - AudioSpectrumProjectionType - AudioFundamentalFrequencyType - AudioHarmonicityType - AudioSignatureType - Silence There is also a (simple) gui available for the encoder. If "Java Web Start" (javaws[.exe]) is installed on your computer (usually it is installed with Java2, version 1.4 by default), you can start it with the following link: http://www.ient.rwth-aachen.de/team/crysandt/software/mpeg7audioenc/MPEG7AudioEncApp.jnlp Note: If "Java Web Start" can't download the files from the internet you might need to set the proxy settings manually. Feel free to test it. Regards Holger Crysandt P.S.: This application is not for windows users only. I also tested it on an Apple notebook and on our Sun Enterprise Server. It works fine on all platforms. -- Dipl.-Ing. Holger Crysandt Institute of Communications Engineering, Aachen University Melatener Str. 23, D-52074 Aachen EMail cry...@ie... Phone +49 241 80-27676 Fax +49 241 80-22196 |
|
From: Holger C. <cry...@ie...> - 2003-11-28 15:15:53
|
Hello developers! Thanks for helping me implementing several descriptors. It was fun working with of you. I think we did a very good job! Next week I think of publishing the first version (0.1) of the Java MPEG-7 Audio Encoder. Please make only minor changes (bug fixes), don't add new features, please. If you need more time, mail me. Have a nice weekend Holger |
|
From: <gio...@wu...> - 2003-11-25 10:06:24
|
Yes, that's what we thought also after messing with it for considerable time.. :-) > IMHO (I am an not an xml expert) this is a bug in the schema. I will > contact the some MPEG-7 developers today. > > Holger > > > > > ------------------------------------------------------- > This SF.net email is sponsored by: SF.net Giveback Program. > Does SourceForge.net help you be more productive? Does it > help you create better code? SHARE THE LOVE, and help us help > YOU! Click Here: http://sourceforge.net/donate/ > _______________________________________________ > Mpeg7audioenc-devel mailing list > Mpe...@li... > https://lists.sourceforge.net/lists/listinfo/mpeg7audioenc-devel |
|
From: Holger C. <cry...@ie...> - 2003-11-25 09:02:11
|
jhm...@fr... schrieb: > I added a new descriptor: AudioHarmonicity. It seems to work but i still > have some problem with XML output validation. I hope to fix it soon. I don't manage to create an valid MPEG-7 description containing the AudioHarmonicity either. I always get the error message "The type (Harmonicity) definition must not be abstract". And this is true. IMHO (I am an not an xml expert) this is a bug in the schema. I will contact the some MPEG-7 developers today. Holger |
|
From: <gio...@wu...> - 2003-11-25 00:50:50
|
I have merged today the branch giovanni.. basically we added the EndOfSignal message propagating from the flush() and changed all the calsses implementing MsgListeners to be "msg safe" .. that is to check the type of the message before processing it (casting it). The implementation of the silence has also changed due to this and is not ready for prime time yet.. (it should be tomorrow) I believe we need some tests and .. a release :-) we need to have a way to check the end mpeg-7 .. some metric between some reference file maybe? I really want to build some mpeg-7 metric which will be fundamental in my thesys so this might be a good "excuse" ? :-) ciao |
|
From: <jhm...@fr...> - 2003-11-24 14:13:03
|
I added a new descriptor: AudioHarmonicity. It seems to work but i still have some problem with XML output validation. I hope to fix it soon. ByeBye! --------------------------------------------------------- Stanco dello spam nella tua email? Prova GRATIS il nuovo servizio ANTISPAM di superEva: http://webmail.supereva.it/spam.html --------------------------------------------------------- |
|
From: <jhm...@fr...> - 2003-11-14 10:18:36
|
Hello Holger! You were right about the bug concerning the "infinite center clipping". I made some correction an now everything seems working.I put the new verison in the repository. Check it out and let me know! ByeBye! --------------------------------------------------------- Stanco dello spam nella tua email? Prova GRATIS il nuovo servizio ANTISPAM di superEva: http://webmail.supereva.it/spam.html --------------------------------------------------------- |
|
From: Holger C. <cry...@ie...> - 2003-11-13 11:41:20
|
Hello developers!
I added the new Descriptor to the graphical user interface. But currently
it needs some bug fixes:
MPEG-7 description (MP7Writer):
- replace "AudioFundamentalFrequency" with "AudioFundamentalFrequencyType"
- replace "lolimit" with "loLimit"
- replace "hilimit" with "hiLimit"
I already made the modifications and checked the new version in.
But sometimes I get a NaN as confidence. This is probably caused by the
"infinite centered clipping" where also the negative samples are set to zero.
This creates a signal only containing zeros which leads to a 0/0=NaN in line
69 ("autocorrelation[k - minlag] = a/Math.sqrt((b * c));")
Holger
jhm...@fr... schrieb:
> Hello everyone! I added a new descriptor: AudioFundamentalFrequency which,
> as you can figure out, extracts the fundamental frequency of the signal.
>
> ByeBye!
|