view testdata/evaluation/outputs/trios-70773820e719-flatten2b-0.04.txt @ 249:43617fe1954e norm

Latest piano test output

Yeti 0.9.9+
1.0
make -C constant-q-cpp -f Makefile.linux libcq.a
make[1]: Entering directory '/home/cannam/code/silvet/constant-q-cpp'
make[1]: 'libcq.a' is up to date.
make[1]: Leaving directory '/home/cannam/code/silvet/constant-q-cpp'
make -C flattendynamics -f Makefile flattendynamics.so
make[1]: Entering directory '/home/cannam/code/silvet/flattendynamics'
make[1]: 'flattendynamics.so' is up to date.
make[1]: Leaving directory '/home/cannam/code/silvet/flattendynamics'

Input files are:
/home/cannam/Music/TRIOS_dataset/mozart/piano.wav
/home/cannam/Music/TRIOS_dataset/mozart/viola.wav
/home/cannam/Music/TRIOS_dataset/mozart/mix.wav
/home/cannam/Music/TRIOS_dataset/mozart/clarinet.wav
/home/cannam/Music/TRIOS_dataset/lussier/piano.wav
/home/cannam/Music/TRIOS_dataset/lussier/mix.wav
/home/cannam/Music/TRIOS_dataset/lussier/bassoon.wav
/home/cannam/Music/TRIOS_dataset/lussier/trumpet.wav
/home/cannam/Music/TRIOS_dataset/take_five/piano.wav
/home/cannam/Music/TRIOS_dataset/take_five/mix.wav
/home/cannam/Music/TRIOS_dataset/take_five/saxophone.wav

Evaluating for file /home/cannam/Music/TRIOS_dataset/mozart/piano.wav...

For piece mozart, arrangement piano, using instrument 1...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="1"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 1%
Extracting and writing features... 2%
Extracting and writing features... 3%
Extracting and writing features... 4%
Extracting and writing features... 5%
Extracting and writing features... 6%
Extracting and writing features... 7%
Extracting and writing features... 8%
Extracting and writing features... 10%
Extracting and writing features... 11%
Extracting and writing features... 12%
Extracting and writing features... 13%
Extracting and writing features... 14%
Extracting and writing features... 15%
Extracting and writing features... 16%
Extracting and writing features... 17%
Extracting and writing features... 18%
Extracting and writing features... 20%
Extracting and writing features... 21%
Extracting and writing features... 22%
Extracting and writing features... 23%
Extracting and writing features... 24%
Extracting and writing features... 25%
Extracting and writing features... 26%
Extracting and writing features... 27%
Extracting and writing features... 28%
Extracting and writing features... 30%
Extracting and writing features... 31%
Extracting and writing features... 32%
Extracting and writing features... 33%
Extracting and writing features... 34%
Extracting and writing features... 35%
Extracting and writing features... 36%
Extracting and writing features... 37%
Extracting and writing features... 38%
Extracting and writing features... 40%
Extracting and writing features... 41%
Extracting and writing features... 42%
Extracting and writing features... 43%
Extracting and writing features... 44%
Extracting and writing features... 45%
Extracting and writing features... 46%
Extracting and writing features... 47%
Extracting and writing features... 48%
Extracting and writing features... 50%
Extracting and writing features... 51%
Extracting and writing features... 52%
Extracting and writing features... 53%
Extracting and writing features... 54%
Extracting and writing features... 55%
Extracting and writing features... 56%
Extracting and writing features... 57%
Extracting and writing features... 58%
Extracting and writing features... 59%
Extracting and writing features... 61%
Extracting and writing features... 62%
Extracting and writing features... 63%
Extracting and writing features... 64%
Extracting and writing features... 65%
Extracting and writing features... 66%
Extracting and writing features... 67%
Extracting and writing features... 68%
Extracting and writing features... 69%
Extracting and writing features... 71%
Extracting and writing features... 72%
Extracting and writing features... 73%
Extracting and writing features... 74%
Extracting and writing features... 75%
Extracting and writing features... 76%
Extracting and writing features... 77%
Extracting and writing features... 78%
Extracting and writing features... 79%
Extracting and writing features... 81%
Extracting and writing features... 82%
Extracting and writing features... 83%
Extracting and writing features... 84%
Extracting and writing features... 85%
Extracting and writing features... 86%
Extracting and writing features... 87%
Extracting and writing features... 88%
Extracting and writing features... 89%
Extracting and writing features... 91%
Extracting and writing features... 92%
Extracting and writing features... 93%
Extracting and writing features... 94%
Extracting and writing features... 95%
Extracting and writing features... 96%
Extracting and writing features... 97%
Extracting and writing features... 98%
Extracting and writing features... 99%
Extracting and writing features... Done


Validating against ground truth at 50 ms:
precision 47.8, recall 67.8, accuracy 38.9, F 56  <-- main mozart/piano

Validating against MIREX submission at 50 ms:
precision 48.3, recall 59.8, accuracy 36.4, F 53.4

Validating MIREX against ground truth at 50 ms:
precision 56.2, recall 64.3, accuracy 42.9, F 60

Validating against ground truth at 100 ms:
precision 57, recall 80.8, accuracy 50.2, F 66.8

Validating against MIREX submission at 100 ms:
precision 64.7, recall 80.2, accuracy 55.8, F 71.6

Validating MIREX against ground truth at 100 ms:
precision 69.4, recall 79.4, accuracy 58.8, F 74.1


Evaluating for file /home/cannam/Music/TRIOS_dataset/mozart/viola.wav...

For piece mozart, arrangement viola, using instrument 4...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="4"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 1%
Extracting and writing features... 2%
Extracting and writing features... 3%
Extracting and writing features... 4%
Extracting and writing features... 5%
Extracting and writing features... 6%
Extracting and writing features... 7%
Extracting and writing features... 8%
Extracting and writing features... 10%
Extracting and writing features... 11%
Extracting and writing features... 12%
Extracting and writing features... 13%
Extracting and writing features... 14%
Extracting and writing features... 15%
Extracting and writing features... 16%
Extracting and writing features... 17%
Extracting and writing features... 18%
Extracting and writing features... 20%
Extracting and writing features... 21%
Extracting and writing features... 22%
Extracting and writing features... 23%
Extracting and writing features... 24%
Extracting and writing features... 25%
Extracting and writing features... 26%
Extracting and writing features... 27%
Extracting and writing features... 28%
Extracting and writing features... 30%
Extracting and writing features... 31%
Extracting and writing features... 32%
Extracting and writing features... 33%
Extracting and writing features... 34%
Extracting and writing features... 35%
Extracting and writing features... 36%
Extracting and writing features... 37%
Extracting and writing features... 38%
Extracting and writing features... 40%
Extracting and writing features... 41%
Extracting and writing features... 42%
Extracting and writing features... 43%
Extracting and writing features... 44%
Extracting and writing features... 45%
Extracting and writing features... 46%
Extracting and writing features... 47%
Extracting and writing features... 48%
Extracting and writing features... 50%
Extracting and writing features... 51%
Extracting and writing features... 52%
Extracting and writing features... 53%
Extracting and writing features... 54%
Extracting and writing features... 55%
Extracting and writing features... 56%
Extracting and writing features... 57%
Extracting and writing features... 58%
Extracting and writing features... 59%
Extracting and writing features... 61%
Extracting and writing features... 62%
Extracting and writing features... 63%
Extracting and writing features... 64%
Extracting and writing features... 65%
Extracting and writing features... 66%
Extracting and writing features... 67%
Extracting and writing features... 68%
Extracting and writing features... 69%
Extracting and writing features... 71%
Extracting and writing features... 72%
Extracting and writing features... 73%
Extracting and writing features... 74%
Extracting and writing features... 75%
Extracting and writing features... 76%
Extracting and writing features... 77%
Extracting and writing features... 78%
Extracting and writing features... 79%
Extracting and writing features... 81%
Extracting and writing features... 82%
Extracting and writing features... 83%
Extracting and writing features... 84%
Extracting and writing features... 85%
Extracting and writing features... 86%
Extracting and writing features... 87%
Extracting and writing features... 88%
Extracting and writing features... 89%
Extracting and writing features... 91%
Extracting and writing features... 92%
Extracting and writing features... 93%
Extracting and writing features... 94%
Extracting and writing features... 95%
Extracting and writing features... 96%
Extracting and writing features... 97%
Extracting and writing features... 98%
Extracting and writing features... 99%
Extracting and writing features... Done


Validating against ground truth at 50 ms:
precision 23.5, recall 65.6, accuracy 21, F 34.7  <-- main mozart/viola

Validating against MIREX submission at 50 ms:
precision 32.5, recall 23.3, accuracy 15.7, F 27.2

Validating MIREX against ground truth at 50 ms:
precision 20.9, recall 81.2, accuracy 20, F 33.3

Validating against ground truth at 100 ms:
precision 26.9, recall 75, accuracy 24.7, F 39.6

Validating against MIREX submission at 100 ms:
precision 40.4, recall 29, accuracy 20.3, F 33.8

Validating MIREX against ground truth at 100 ms:
precision 22.5, recall 87.5, accuracy 21.8, F 35.8


Evaluating for file /home/cannam/Music/TRIOS_dataset/mozart/mix.wav...

For piece mozart, arrangement mix, using instrument 0...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="0"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 1%
Extracting and writing features... 2%
Extracting and writing features... 3%
Extracting and writing features... 4%
Extracting and writing features... 5%
Extracting and writing features... 6%
Extracting and writing features... 7%
Extracting and writing features... 8%
Extracting and writing features... 10%
Extracting and writing features... 11%
Extracting and writing features... 12%
Extracting and writing features... 13%
Extracting and writing features... 14%
Extracting and writing features... 15%
Extracting and writing features... 16%
Extracting and writing features... 17%
Extracting and writing features... 18%
Extracting and writing features... 20%
Extracting and writing features... 21%
Extracting and writing features... 22%
Extracting and writing features... 23%
Extracting and writing features... 24%
Extracting and writing features... 25%
Extracting and writing features... 26%
Extracting and writing features... 27%
Extracting and writing features... 28%
Extracting and writing features... 30%
Extracting and writing features... 31%
Extracting and writing features... 32%
Extracting and writing features... 33%
Extracting and writing features... 34%
Extracting and writing features... 35%
Extracting and writing features... 36%
Extracting and writing features... 37%
Extracting and writing features... 38%
Extracting and writing features... 40%
Extracting and writing features... 41%
Extracting and writing features... 42%
Extracting and writing features... 43%
Extracting and writing features... 44%
Extracting and writing features... 45%
Extracting and writing features... 46%
Extracting and writing features... 47%
Extracting and writing features... 48%
Extracting and writing features... 50%
Extracting and writing features... 51%
Extracting and writing features... 52%
Extracting and writing features... 53%
Extracting and writing features... 54%
Extracting and writing features... 55%
Extracting and writing features... 56%
Extracting and writing features... 57%
Extracting and writing features... 58%
Extracting and writing features... 59%
Extracting and writing features... 61%
Extracting and writing features... 62%
Extracting and writing features... 63%
Extracting and writing features... 64%
Extracting and writing features... 65%
Extracting and writing features... 66%
Extracting and writing features... 67%
Extracting and writing features... 68%
Extracting and writing features... 69%
Extracting and writing features... 71%
Extracting and writing features... 72%
Extracting and writing features... 73%
Extracting and writing features... 74%
Extracting and writing features... 75%
Extracting and writing features... 76%
Extracting and writing features... 77%
Extracting and writing features... 78%
Extracting and writing features... 79%
Extracting and writing features... 81%
Extracting and writing features... 82%
Extracting and writing features... 83%
Extracting and writing features... 84%
Extracting and writing features... 85%
Extracting and writing features... 86%
Extracting and writing features... 87%
Extracting and writing features... 88%
Extracting and writing features... 89%
Extracting and writing features... 91%
Extracting and writing features... 92%
Extracting and writing features... 93%
Extracting and writing features... 94%
Extracting and writing features... 95%
Extracting and writing features... 96%
Extracting and writing features... 97%
Extracting and writing features... 98%
Extracting and writing features... 99%
Extracting and writing features... Done


Validating against ground truth at 50 ms:
precision 58.5, recall 52.6, accuracy 38.3, F 55.4  <-- main mozart/mix

Validating against MIREX submission at 50 ms:
precision 59.4, recall 52.8, accuracy 38.8, F 55.9

Validating MIREX against ground truth at 50 ms:
precision 50.4, recall 51, accuracy 33.9, F 50.7

Validating against ground truth at 100 ms:
precision 72.9, recall 65.5, accuracy 52.7, F 69

Validating against MIREX submission at 100 ms:
precision 77.4, recall 68.8, accuracy 57.3, F 72.8

Validating MIREX against ground truth at 100 ms:
precision 67.2, recall 68, accuracy 51, F 67.6


Evaluating for file /home/cannam/Music/TRIOS_dataset/mozart/clarinet.wav...

For piece mozart, arrangement clarinet, using instrument 9...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="9"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 1%
Extracting and writing features... 2%
Extracting and writing features... 3%
Extracting and writing features... 4%
Extracting and writing features... 5%
Extracting and writing features... 6%
Extracting and writing features... 7%
Extracting and writing features... 8%
Extracting and writing features... 10%
Extracting and writing features... 11%
Extracting and writing features... 12%
Extracting and writing features... 13%
Extracting and writing features... 14%
Extracting and writing features... 15%
Extracting and writing features... 16%
Extracting and writing features... 17%
Extracting and writing features... 18%
Extracting and writing features... 20%
Extracting and writing features... 21%
Extracting and writing features... 22%
Extracting and writing features... 23%
Extracting and writing features... 24%
Extracting and writing features... 25%
Extracting and writing features... 26%
Extracting and writing features... 27%
Extracting and writing features... 28%
Extracting and writing features... 30%
Extracting and writing features... 31%
Extracting and writing features... 32%
Extracting and writing features... 33%
Extracting and writing features... 34%
Extracting and writing features... 35%
Extracting and writing features... 36%
Extracting and writing features... 37%
Extracting and writing features... 38%
Extracting and writing features... 40%
Extracting and writing features... 41%
Extracting and writing features... 42%
Extracting and writing features... 43%
Extracting and writing features... 44%
Extracting and writing features... 45%
Extracting and writing features... 46%
Extracting and writing features... 47%
Extracting and writing features... 48%
Extracting and writing features... 50%
Extracting and writing features... 51%
Extracting and writing features... 52%
Extracting and writing features... 53%
Extracting and writing features... 54%
Extracting and writing features... 55%
Extracting and writing features... 56%
Extracting and writing features... 57%
Extracting and writing features... 58%
Extracting and writing features... 59%
Extracting and writing features... 61%
Extracting and writing features... 62%
Extracting and writing features... 63%
Extracting and writing features... 64%
Extracting and writing features... 65%
Extracting and writing features... 66%
Extracting and writing features... 67%
Extracting and writing features... 68%
Extracting and writing features... 69%
Extracting and writing features... 71%
Extracting and writing features... 72%
Extracting and writing features... 73%
Extracting and writing features... 74%
Extracting and writing features... 75%
Extracting and writing features... 76%
Extracting and writing features... 77%
Extracting and writing features... 78%
Extracting and writing features... 79%
Extracting and writing features... 81%
Extracting and writing features... 82%
Extracting and writing features... 83%
Extracting and writing features... 84%
Extracting and writing features... 85%
Extracting and writing features... 86%
Extracting and writing features... 87%
Extracting and writing features... 88%
Extracting and writing features... 89%
Extracting and writing features... 91%
Extracting and writing features... 92%
Extracting and writing features... 93%
Extracting and writing features... 94%
Extracting and writing features... 95%
Extracting and writing features... 96%
Extracting and writing features... 97%
Extracting and writing features... 98%
Extracting and writing features... 99%
Extracting and writing features... Done


Validating against ground truth at 50 ms:
precision 93.2, recall 79.7, accuracy 75.3, F 85.9  <-- main mozart/clarinet

Validating against MIREX submission at 50 ms:
precision 83, recall 68, accuracy 59.7, F 74.8

Validating MIREX against ground truth at 50 ms:
precision 72.2, recall 75.3, accuracy 58.4, F 73.7

Validating against ground truth at 100 ms:
precision 100, recall 85.5, accuracy 85.5, F 92.1

Validating against MIREX submission at 100 ms:
precision 89.8, recall 73.6, accuracy 67.9, F 80.9

Validating MIREX against ground truth at 100 ms:
precision 75, recall 78.2, accuracy 62, F 76.5


Evaluating for file /home/cannam/Music/TRIOS_dataset/lussier/piano.wav...

For piece lussier, arrangement piano, using instrument 1...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="1"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 2%
Extracting and writing features... 4%
Extracting and writing features... 6%
Extracting and writing features... 8%
Extracting and writing features... 10%
Extracting and writing features... 12%
Extracting and writing features... 14%
Extracting and writing features... 16%
Extracting and writing features... 18%
Extracting and writing features... 20%
Extracting and writing features... 22%
Extracting and writing features... 24%
Extracting and writing features... 26%
Extracting and writing features... 28%
Extracting and writing features... 31%
Extracting and writing features... 33%
Extracting and writing features... 35%
Extracting and writing features... 37%
Extracting and writing features... 39%
Extracting and writing features... 41%
Extracting and writing features... 43%
Extracting and writing features... 45%
Extracting and writing features... 47%
Extracting and writing features... 49%
Extracting and writing features... 51%
Extracting and writing features... 53%
Extracting and writing features... 55%
Extracting and writing features... 57%
Extracting and writing features... 59%
Extracting and writing features... 62%
Extracting and writing features... 64%
Extracting and writing features... 66%
Extracting and writing features... 68%
Extracting and writing features... 70%
Extracting and writing features... 72%
Extracting and writing features... 74%
Extracting and writing features... 76%
Extracting and writing features... 78%
Extracting and writing features... 80%
Extracting and writing features... 82%
Extracting and writing features... 84%
Extracting and writing features... 86%
Extracting and writing features... 88%
Extracting and writing features... 90%
Extracting and writing features... 92%
Extracting and writing features... 95%
Extracting and writing features... 97%
Extracting and writing features... 99%
Extracting and writing features... Done


Validating against ground truth at 50 ms:
precision 69.6, recall 57.9, accuracy 46.3, F 63.3  <-- main lussier/piano

Validating against MIREX submission at 50 ms:
precision 55.5, recall 46.2, accuracy 33.7, F 50.4

Validating MIREX against ground truth at 50 ms:
precision 45.3, recall 45.3, accuracy 29.3, F 45.3

Validating against ground truth at 100 ms:
precision 77.7, recall 64.7, accuracy 54.6, F 70.6

Validating against MIREX submission at 100 ms:
precision 76.7, recall 63.8, accuracy 53.5, F 69.7

Validating MIREX against ground truth at 100 ms:
precision 62.1, recall 62.1, accuracy 45.1, F 62.1


Evaluating for file /home/cannam/Music/TRIOS_dataset/lussier/mix.wav...

For piece lussier, arrangement mix, using instrument 0...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="0"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 2%
Extracting and writing features... 4%
Extracting and writing features... 6%
Extracting and writing features... 8%
Extracting and writing features... 10%
Extracting and writing features... 12%
Extracting and writing features... 14%
Extracting and writing features... 16%
Extracting and writing features... 18%
Extracting and writing features... 20%
Extracting and writing features... 22%
Extracting and writing features... 24%
Extracting and writing features... 26%
Extracting and writing features... 28%
Extracting and writing features... 31%
Extracting and writing features... 33%
Extracting and writing features... 35%
Extracting and writing features... 37%
Extracting and writing features... 39%
Extracting and writing features... 41%
Extracting and writing features... 43%
Extracting and writing features... 45%
Extracting and writing features... 47%
Extracting and writing features... 49%
Extracting and writing features... 51%
Extracting and writing features... 53%
Extracting and writing features... 55%
Extracting and writing features... 57%
Extracting and writing features... 59%
Extracting and writing features... 62%
Extracting and writing features... 64%
Extracting and writing features... 66%
Extracting and writing features... 68%
Extracting and writing features... 70%
Extracting and writing features... 72%
Extracting and writing features... 74%
Extracting and writing features... 76%
Extracting and writing features... 78%
Extracting and writing features... 80%
Extracting and writing features... 82%
Extracting and writing features... 84%
Extracting and writing features... 86%
Extracting and writing features... 88%
Extracting and writing features... 90%
Extracting and writing features... 92%
Extracting and writing features... 95%
Extracting and writing features... 97%
Extracting and writing features... 99%
Extracting and writing features... Done


Validating against ground truth at 50 ms:
precision 51.9, recall 32.6, accuracy 25, F 40.1  <-- main lussier/mix

Validating against MIREX submission at 50 ms:
precision 57.2, recall 46, accuracy 34.2, F 51

Validating MIREX against ground truth at 50 ms:
precision 40.4, recall 31.7, accuracy 21.6, F 35.5

Validating against ground truth at 100 ms:
precision 63.3, recall 39.9, accuracy 32.4, F 48.9

Validating against MIREX submission at 100 ms:
precision 72.5, recall 58.2, accuracy 47.7, F 64.6

Validating MIREX against ground truth at 100 ms:
precision 57.6, recall 45.1, accuracy 33.9, F 50.6


Evaluating for file /home/cannam/Music/TRIOS_dataset/lussier/bassoon.wav...

For piece lussier, arrangement bassoon, using instrument 11...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="11"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 2%
Extracting and writing features... 4%
Extracting and writing features... 6%
Extracting and writing features... 8%
Extracting and writing features... 10%
Extracting and writing features... 12%
Extracting and writing features... 14%
Extracting and writing features... 16%
Extracting and writing features... 18%
Extracting and writing features... 20%
Extracting and writing features... 22%
Extracting and writing features... 24%
Extracting and writing features... 26%
Extracting and writing features... 28%
Extracting and writing features... 31%
Extracting and writing features... 33%
Extracting and writing features... 35%
Extracting and writing features... 37%
Extracting and writing features... 39%
Extracting and writing features... 41%
Extracting and writing features... 43%
Extracting and writing features... 45%
Extracting and writing features... 47%
Extracting and writing features... 49%
Extracting and writing features... 51%
Extracting and writing features... 53%
Extracting and writing features... 55%
Extracting and writing features... 57%
Extracting and writing features... 59%
Extracting and writing features... 62%
Extracting and writing features... 64%
Extracting and writing features... 66%
Extracting and writing features... 68%
Extracting and writing features... 70%
Extracting and writing features... 72%
Extracting and writing features... 74%
Extracting and writing features... 76%
Extracting and writing features... 78%
Extracting and writing features... 80%
Extracting and writing features... 82%
Extracting and writing features... 84%
Extracting and writing features... 86%
Extracting and writing features... 88%
Extracting and writing features... 90%
Extracting and writing features... 92%
Extracting and writing features... 95%
Extracting and writing features... 97%
Extracting and writing features... 99%
Extracting and writing features... Done


Validating against ground truth at 50 ms:
precision 73, recall 87.6, accuracy 66.2, F 79.7  <-- main lussier/bassoon

Validating against MIREX submission at 50 ms:
precision 53.8, recall 30, accuracy 23.8, F 38.5

Validating MIREX against ground truth at 50 ms:
precision 31.4, recall 67.6, accuracy 27.3, F 42.9

Validating against ground truth at 100 ms:
precision 75.6, recall 90.7, accuracy 70.2, F 82.5

Validating against MIREX submission at 100 ms:
precision 70.5, recall 39.2, accuracy 33.7, F 50.4

Validating MIREX against ground truth at 100 ms:
precision 37.1, recall 80, accuracy 33.9, F 50.7


Evaluating for file /home/cannam/Music/TRIOS_dataset/lussier/trumpet.wav...

For piece lussier, arrangement trumpet, using instrument 0...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="0"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 2%
Extracting and writing features... 4%
Extracting and writing features... 6%
Extracting and writing features... 8%
Extracting and writing features... 10%
Extracting and writing features... 12%
Extracting and writing features... 14%
Extracting and writing features... 16%
Extracting and writing features... 18%
Extracting and writing features... 20%
Extracting and writing features... 22%
Extracting and writing features... 24%
Extracting and writing features... 26%
Extracting and writing features... 28%
Extracting and writing features... 31%
Extracting and writing features... 33%
Extracting and writing features... 35%
Extracting and writing features... 37%
Extracting and writing features... 39%
Extracting and writing features... 41%
Extracting and writing features... 43%
Extracting and writing features... 45%
Extracting and writing features... 47%
Extracting and writing features... 49%
Extracting and writing features... 51%
Extracting and writing features... 53%
Extracting and writing features... 55%
Extracting and writing features... 57%
Extracting and writing features... 59%
Extracting and writing features... 62%
Extracting and writing features... 64%
Extracting and writing features... 66%
Extracting and writing features... 68%
Extracting and writing features... 70%
Extracting and writing features... 72%
Extracting and writing features... 74%
Extracting and writing features... 76%
Extracting and writing features... 78%
Extracting and writing features... 80%
Extracting and writing features... 82%
Extracting and writing features... 84%
Extracting and writing features... 86%
Extracting and writing features... 88%
Extracting and writing features... 90%
Extracting and writing features... 92%
Extracting and writing features... 95%
Extracting and writing features... 97%
Extracting and writing features... 99%
Extracting and writing features... Done


Validating against ground truth at 50 ms:
precision 34.9, recall 91.6, accuracy 33.8, F 50.5  <-- main lussier/trumpet

Validating against MIREX submission at 50 ms:
precision 44.4, recall 40.5, accuracy 26.9, F 42.4

Validating MIREX against ground truth at 50 ms:
precision 28.9, recall 83.3, accuracy 27.3, F 43

Validating against ground truth at 100 ms:
precision 34.9, recall 91.6, accuracy 33.8, F 50.5

Validating against MIREX submission at 100 ms:
precision 65, recall 59.4, accuracy 45, F 62.1

Validating MIREX against ground truth at 100 ms:
precision 31.8, recall 91.6, accuracy 30.9, F 47.3


Evaluating for file /home/cannam/Music/TRIOS_dataset/take_five/piano.wav...

For piece take_five, arrangement piano, using instrument 1...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="1"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 1%
Extracting and writing features... 2%
Extracting and writing features... 3%
Extracting and writing features... 4%
Extracting and writing features... 5%
Extracting and writing features... 6%
Extracting and writing features... 7%
Extracting and writing features... 8%
Extracting and writing features... 9%
Extracting and writing features... 10%
Extracting and writing features... 11%
Extracting and writing features... 12%
Extracting and writing features... 13%
Extracting and writing features... 14%
Extracting and writing features... 15%
Extracting and writing features... 16%
Extracting and writing features... 17%
Extracting and writing features... 18%
Extracting and writing features... 19%
Extracting and writing features... 20%
Extracting and writing features... 21%
Extracting and writing features... 22%
Extracting and writing features... 23%
Extracting and writing features... 24%
Extracting and writing features... 25%
Extracting and writing features... 26%
Extracting and writing features... 27%
Extracting and writing features... 28%
Extracting and writing features... 29%
Extracting and writing features... 30%
Extracting and writing features... 31%
Extracting and writing features... 32%
Extracting and writing features... 33%
Extracting and writing features... 34%
Extracting and writing features... 35%
Extracting and writing features... 36%
Extracting and writing features... 37%
Extracting and writing features... 38%
Extracting and writing features... 39%
Extracting and writing features... 40%
Extracting and writing features... 41%
Extracting and writing features... 42%
Extracting and writing features... 43%
Extracting and writing features... 44%
Extracting and writing features... 45%
Extracting and writing features... 46%
Extracting and writing features... 47%
Extracting and writing features... 48%
Extracting and writing features... 49%
Extracting and writing features... 50%
Extracting and writing features... 51%
Extracting and writing features... 52%
Extracting and writing features... 53%
Extracting and writing features... 54%
Extracting and writing features... 55%
Extracting and writing features... 56%
Extracting and writing features... 57%
Extracting and writing features... 58%
Extracting and writing features... 59%
Extracting and writing features... 60%
Extracting and writing features... 61%
Extracting and writing features... 62%
Extracting and writing features... 63%
Extracting and writing features... 64%
Extracting and writing features... 65%
Extracting and writing features... 66%
Extracting and writing features... 67%
Extracting and writing features... 68%
Extracting and writing features... 69%
Extracting and writing features... 70%
Extracting and writing features... 71%
Extracting and writing features... 72%
Extracting and writing features... 73%
Extracting and writing features... 74%
Extracting and writing features... 75%
Extracting and writing features... 76%
Extracting and writing features... 77%
Extracting and writing features... 78%
Extracting and writing features... 79%
Extracting and writing features... 80%
Extracting and writing features... 81%
Extracting and writing features... 82%
Extracting and writing features... 83%
Extracting and writing features... 84%
Extracting and writing features... 85%
Extracting and writing features... 86%
Extracting and writing features... 87%
Extracting and writing features... 88%
Extracting and writing features... 89%
Extracting and writing features... 90%
Extracting and writing features... 91%
Extracting and writing features... 92%
Extracting and writing features... 93%
Extracting and writing features... 94%
Extracting and writing features... 95%
Extracting and writing features... 96%
Extracting and writing features... 97%
Extracting and writing features... 98%
Extracting and writing features... 99%
Extracting and writing features... 100%
Extracting and writing features... Done

Validating against ground truth at 50 ms:
precision 60.3, recall 79.2, accuracy 52.1, F 68.5  <-- main take_five/piano

Validating against MIREX submission at 50 ms:
precision 77.8, recall 67, accuracy 56.3, F 72

Validating MIREX against ground truth at 50 ms:
precision 50.1, recall 76.4, accuracy 43.4, F 60.5

Validating against ground truth at 100 ms:
precision 63.8, recall 83.8, accuracy 56.8, F 72.5

Validating against MIREX submission at 100 ms:
precision 90.5, recall 77.9, accuracy 72, F 83.7

Validating MIREX against ground truth at 100 ms:
precision 51.9, recall 79.2, accuracy 45.7, F 62.7


Evaluating for file /home/cannam/Music/TRIOS_dataset/take_five/mix.wav...

For piece take_five, arrangement mix, using instrument 0...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="0"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 1%
Extracting and writing features... 2%
Extracting and writing features... 3%
Extracting and writing features... 4%
Extracting and writing features... 5%
Extracting and writing features... 6%
Extracting and writing features... 7%
Extracting and writing features... 8%
Extracting and writing features... 9%
Extracting and writing features... 10%
Extracting and writing features... 11%
Extracting and writing features... 12%
Extracting and writing features... 13%
Extracting and writing features... 14%
Extracting and writing features... 15%
Extracting and writing features... 16%
Extracting and writing features... 17%
Extracting and writing features... 18%
Extracting and writing features... 19%
Extracting and writing features... 20%
Extracting and writing features... 21%
Extracting and writing features... 22%
Extracting and writing features... 23%
Extracting and writing features... 24%
Extracting and writing features... 25%
Extracting and writing features... 26%
Extracting and writing features... 27%
Extracting and writing features... 28%
Extracting and writing features... 29%
Extracting and writing features... 30%
Extracting and writing features... 31%
Extracting and writing features... 32%
Extracting and writing features... 33%
Extracting and writing features... 34%
Extracting and writing features... 35%
Extracting and writing features... 36%
Extracting and writing features... 37%
Extracting and writing features... 38%
Extracting and writing features... 39%
Extracting and writing features... 40%
Extracting and writing features... 41%
Extracting and writing features... 42%
Extracting and writing features... 43%
Extracting and writing features... 44%
Extracting and writing features... 45%
Extracting and writing features... 46%
Extracting and writing features... 47%
Extracting and writing features... 48%
Extracting and writing features... 49%
Extracting and writing features... 50%
Extracting and writing features... 51%
Extracting and writing features... 52%
Extracting and writing features... 53%
Extracting and writing features... 54%
Extracting and writing features... 55%
Extracting and writing features... 56%
Extracting and writing features... 57%
Extracting and writing features... 58%
Extracting and writing features... 59%
Extracting and writing features... 60%
Extracting and writing features... 61%
Extracting and writing features... 62%
Extracting and writing features... 63%
Extracting and writing features... 64%
Extracting and writing features... 65%
Extracting and writing features... 66%
Extracting and writing features... 67%
Extracting and writing features... 68%
Extracting and writing features... 69%
Extracting and writing features... 70%
Extracting and writing features... 71%
Extracting and writing features... 72%
Extracting and writing features... 73%
Extracting and writing features... 74%
Extracting and writing features... 75%
Extracting and writing features... 76%
Extracting and writing features... 77%
Extracting and writing features... 78%
Extracting and writing features... 79%
Extracting and writing features... 80%
Extracting and writing features... 81%
Extracting and writing features... 82%
Extracting and writing features... 83%
Extracting and writing features... 84%
Extracting and writing features... 85%
Extracting and writing features... 86%
Extracting and writing features... 87%
Extracting and writing features... 88%
Extracting and writing features... 89%
Extracting and writing features... 90%
Extracting and writing features... 91%
Extracting and writing features... 92%
Extracting and writing features... 93%
Extracting and writing features... 94%
Extracting and writing features... 95%
Extracting and writing features... 96%
Extracting and writing features... 97%
Extracting and writing features... 98%
Extracting and writing features... 99%
Extracting and writing features... 100%
Extracting and writing features... Done

Validating against ground truth at 50 ms:
precision 68.8, recall 69.2, accuracy 52.7, F 69  <-- main take_five/mix

Validating against MIREX submission at 50 ms:
precision 73.1, recall 60.9, accuracy 49.7, F 66.4

Validating MIREX against ground truth at 50 ms:
precision 56.5, recall 68.3, accuracy 44.8, F 61.8

Validating against ground truth at 100 ms:
precision 78, recall 78.5, accuracy 64.3, F 78.3

Validating against MIREX submission at 100 ms:
precision 88.2, recall 73.5, accuracy 66.9, F 80.2

Validating MIREX against ground truth at 100 ms:
precision 65.5, recall 79.1, accuracy 55.9, F 71.7


Evaluating for file /home/cannam/Music/TRIOS_dataset/take_five/saxophone.wav...

For piece take_five, arrangement saxophone, using instrument 0...
force -> 
one-file -> /tmp/16432
Have audio source: "/tmp/16432norm.wav"

Determining default rate and channel count from first input file... Done
File or URL "/tmp/16432norm.wav" opened successfully
Taking default channel count of 1 from file
Taking default sample rate of 44100Hz from file
(Note: Default may be overridden by transforms)
RDFTransformFactory: NOTE: Transform is: 
<transform
    id="vamp:silvet:silvet:"
    pluginVersion=""
    program=""
    stepSize="1024"
    blockSize="1024"
    windowType="hanning"
    startTime="0.000000000"
    duration="0.000000000"
    sampleRate="0">
  <parameter name="finetune" value="0"/>
  <parameter name="instrument" value="0"/>
  <parameter name="mode" value="1"/>
</transform>

NOTE: Transform does not specify a sample rate, using default rate of 44100
NOTE: Loaded and initialised plugin for transform "vamp:silvet:silvet:" with plugin step size 1024 and block size 1024 (adapter step and block size 16384)
Extracting features for: "/tmp/16432norm.wav"
Audio file "/tmp/16432norm.wav": 1ch at 44100Hz

Extracting and writing features... 1%
Extracting and writing features... 2%
Extracting and writing features... 3%
Extracting and writing features... 4%
Extracting and writing features... 5%
Extracting and writing features... 6%
Extracting and writing features... 7%
Extracting and writing features... 8%
Extracting and writing features... 9%
Extracting and writing features... 10%
Extracting and writing features... 11%
Extracting and writing features... 12%
Extracting and writing features... 13%
Extracting and writing features... 14%
Extracting and writing features... 15%
Extracting and writing features... 16%
Extracting and writing features... 17%
Extracting and writing features... 18%
Extracting and writing features... 19%
Extracting and writing features... 20%
Extracting and writing features... 21%
Extracting and writing features... 22%
Extracting and writing features... 23%
Extracting and writing features... 24%
Extracting and writing features... 25%
Extracting and writing features... 26%
Extracting and writing features... 27%
Extracting and writing features... 28%
Extracting and writing features... 29%
Extracting and writing features... 30%
Extracting and writing features... 31%
Extracting and writing features... 32%
Extracting and writing features... 33%
Extracting and writing features... 34%
Extracting and writing features... 35%
Extracting and writing features... 36%
Extracting and writing features... 37%
Extracting and writing features... 38%
Extracting and writing features... 39%
Extracting and writing features... 40%
Extracting and writing features... 41%
Extracting and writing features... 42%
Extracting and writing features... 43%
Extracting and writing features... 44%
Extracting and writing features... 45%
Extracting and writing features... 46%
Extracting and writing features... 47%
Extracting and writing features... 48%
Extracting and writing features... 49%
Extracting and writing features... 50%
Extracting and writing features... 51%
Extracting and writing features... 52%
Extracting and writing features... 53%
Extracting and writing features... 54%
Extracting and writing features... 55%
Extracting and writing features... 56%
Extracting and writing features... 57%
Extracting and writing features... 58%
Extracting and writing features... 59%
Extracting and writing features... 60%
Extracting and writing features... 61%
Extracting and writing features... 62%
Extracting and writing features... 63%
Extracting and writing features... 64%
Extracting and writing features... 65%
Extracting and writing features... 66%
Extracting and writing features... 67%
Extracting and writing features... 68%
Extracting and writing features... 69%
Extracting and writing features... 70%
Extracting and writing features... 71%
Extracting and writing features... 72%
Extracting and writing features... 73%
Extracting and writing features... 74%
Extracting and writing features... 75%
Extracting and writing features... 76%
Extracting and writing features... 77%
Extracting and writing features... 78%
Extracting and writing features... 79%
Extracting and writing features... 80%
Extracting and writing features... 81%
Extracting and writing features... 82%
Extracting and writing features... 83%
Extracting and writing features... 84%
Extracting and writing features... 85%
Extracting and writing features... 86%
Extracting and writing features... 87%
Extracting and writing features... 88%
Extracting and writing features... 89%
Extracting and writing features... 90%
Extracting and writing features... 91%
Extracting and writing features... 92%
Extracting and writing features... 93%
Extracting and writing features... 94%
Extracting and writing features... 95%
Extracting and writing features... 96%
Extracting and writing features... 97%
Extracting and writing features... 98%
Extracting and writing features... 99%
Extracting and writing features... 100%
Extracting and writing features... Done

Validating against ground truth at 50 ms:
precision 78.9, recall 89.5, accuracy 72.3, F 83.9  <-- main take_five/saxophone

Validating against MIREX submission at 50 ms:
precision 84, recall 71.4, accuracy 62.8, F 77.2

Validating MIREX against ground truth at 50 ms:
precision 67.8, recall 90.4, accuracy 63.3, F 77.5

Validating against ground truth at 100 ms:
precision 83.1, recall 94.2, accuracy 79.2, F 88.3

Validating against MIREX submission at 100 ms:
precision 92.4, recall 78.5, accuracy 73.8, F 84.9

Validating MIREX against ground truth at 100 ms:
precision 70, recall 93.3, accuracy 66.6, F 80


real	3m30.120s
user	8m44.897s
sys	0m2.810s