Recent Posts

Just letting everyone know that this works just fine on the Raspberry Pi under Raspbian (Effectively Debian). It's a lengthy and tricky compile process (lots of dependencies and compiling libraries from source), but the program runs really well.

/Dextrus

Dear Vamp Plug-ins Forum,

I am getting errors when I try to run Queen Mary's structural segmentation vamp plug-in in sonic annotator.  I am using 32 bit Ubuntu 12.04, and I am using the latest version of sonic annotator and the Queen Mary vamp plug-ins.  When I try running it, I get the following error in my terminal...

root@david-VirtualBox:~/Downloads/sonic-annotator-0.7-linux-i386# ./sonic-annotator -d vamp:qm-vamp-plugins:qm-segmenter:segmentation "/home/david/The Foundation/testfile.wav" -w csv --csv-basedir "/home/david/The Foundation/"
basedir -> /home/david/The Foundation/
Have audio source: "/home/david/The Foundation/testfile.wav"
Determining default rate and channel count from first input file... Done
File or URL "/home/david/The Foundation/testfile.wav" opened successfully
Taking default channel count of 2 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
NOTE: Using default sample rate of 44100 for default transform
NOTE: Loaded and initialised plugin for transform "vamp:qm-vamp-plugins:qm-segmenter:segmentation" with plugin step size 8820 and block size 26460 (adapter step and block size 16384)
Extracting features for: "/home/david/The Foundation/testfile.wav"
Audio file "/home/david/The Foundation/testfile.wav": 2ch at 44100Hz
Extracting and writing features... 100%./sonic-annotator: symbol lookup error: /usr/lib/vamp/qm-vamp-plugins.so: undefined symbol: clapack_dgetrf

Any help would be greatly appreciated.

Sincerely,

David Garber
SmartSound Software, Inc.
8550 Balboa Blvd.
Suite 180
Northridge, CA 91325

jVamp is a JNI (Java Native Interface) library that permits Java applications to load and use Vamp audio analysis plugins.

jVamp is intended for application developers who are interested in adding capabilities to Java applications or a Java environment; it's not currently something that an end-user would normally wish to install. But if you are working in Java, you might well find it interesting.

Note that jVamp is intended to be compatible with the Android NDK as well as desktop Java, though it hasn't been tested there (and existing binary distributions of plugins do not usually include ARM builds).

See http://code.soundsoftware.ac.uk/projects/jvamp for more information.

Chris

Hi justin,

Just for information, I also developped a plug-in related to yours:
https://github.com/wslihgt/IMMF0salience

It only does the representation part, but does it with another algorithm (Source/Filter NMF, as you may know). Any comment welcome!

Best regards,
Jean-Louis

P.S.: since the new version of SV for macosx, in native 64bits, I m not quite sure the compiled version works anymore. yet another TODO...

The MELODIA plug-in automatically estimates the pitch of a song's main melody. More specifically, it implements an algorithm that automatically estimates the fundamental frequency corresponding to the pitch of the predominant melodic line of a piece of polyphonic (or homophonic or monophonic) music.

You can find out more about the plug-in and download it here: http://mtg.upf.edu/technologies/melodia

Cheers,

Justin

I'm trying to understand in detail exactly how the results of the adaptive spectrogram matches up with musical time. On sonic annotator, I ran the vamp:qm-vamp-plugins:qm-adaptivespectrogram:output transform on some music (127.663 seconds), using the -d option. "with plugin step size 1024 and block size 2048 (adapter step and block size 16384)"  The output is a 22106 X 1024 matrix.

My question: how do I interpret the data in terms of time? For each of the 22106 vectors, I would like to associate it with the proper span of musical time, and I'm not sure how to do that.

The documentation of the adaptive spectrogram does help me understand what the algorithm is doing, but it doesn't really answer my question.

Extra question 1: from the programmer's guide I take it the first block is not centred on time zero but rather starts at the first sample of the audio right? (double checking, as this could cause alignment issues when checking against ground-truths centred on time 0).

Depends on the host, but you can tell from the timestamp provided.

The docs (http://code.soundsoftware.ac.uk/embedded/vamp-plugin-sdk/classVamp_1_1Plugin.html#ae4aed3bebfe80a2e2fccd3d37af26996) say that "[t]he timestamp will be the real time in seconds of the centre of the FFT input window". Therefore, if the first timestamp is zero, that should mean you are being passed a window centred on the start of the audio rather than starting with the first sample.

Just in case anyone is following this thread - I just realised this is not the case if you use the time domain input - in that case the timestamp will be of the beginning of the block! Sorry Chris, I realise that's stated quite clearly in the link you posted, I should've read it more carefully.

Anyway, for those of us working with time-domain input, if you want "correct" timestamps you should add windowduration/2 to your timestamps (where windowduration is expressed in seconds, i.e. m_blockSize / inputSampleRate).

Cheers

Hi- I just downloaded the new version.

I'm having sample-rate troubles. Sound files at 44100Hz are being read as 48000Hz.

How do I fix it?

Hi Chris!

I'm glad to say I already have a very early alpha version of my plugin working. I'll probably have some more questions as I get closer to a proper release, but for now just a couple of small questions:

1) The plug-in has several outputs. Of these, one is a matrix type (with 600 bins) and the other is a single value per frame (in cents). Ideally, I want both to be on the same scale, so the matrix bins represent values 0-5990 cents (i.e. bin*10), and then if the single value output has a value of say 4000 cents, it should appear at the same height as bin 400 (which represents value 400*10 = 4000 cents). In trying to do this I have 2 problems (a) apart from giving the matrix bins names, can I set them to represent numerical values? (b) since the single value has a greater range than the matrix (matrix goes from bin 0 to bin 599, but single value can also have negative values), they are both squeezed to fit on the screen meaning the values are not aligned. Any possible solution?

2) When loading QM plugins, there is a little button at the top right corner of the plugin parameters window with an 'i', which when clicked opens a url (in this case http://vamp-plugins.org/plugin-doc/qm-vamp-plugins.html). Would it be possible to have this 'i' button on my plug-in and direct it to a url of my choice?

Thanks!

justin

Version 2.4 of the Vamp plugin SDK is now available.

   http://www.vamp-plugins.org/

Vamp is a plugin API for audio analysis and feature extraction plugins written in C or C++.  Its SDK features an easy-to-use set of C++ classes for plugin and host developers, a reference host implementation, example plugins, and documentation.  It is supported across Linux, OS/X, and Windows.

A documentation guide to writing plugins using the Vamp SDK can be found at http://www.vamp-plugins.org/guide.pdf.

Version 2.4 is a maintenance and bugfix release. The main addition is a simple FFT implementation for plugins to use. For more details, see the changelog at

http://code.soundsoftware.ac.uk/projects/vamp-plugin-sdk/repository/entry/CHANGELOG


Chris