conclusion.tex

   1 \section{Conclusion \& Future Research}
   2 This research shows that existing techniques for singing-voice detection
   3 designed for regular singing voices also work respectably on extreme singing
   4 styles like grunting. With a standard \gls{ANN} classifier using \gls{MFCC}
   5 features a performance of $85\%$ can be achieved which is similar to the same
   6 techniques on regular singing. This means that it might be suitable as a
   7 pre-processing step for lyrics forced alignment.
   8
   9 Future interesting research includes doing the actual forced alignment. This
  10 probably requires entirely different models. The models used for real speech
  11 are probably not suitable because the acoustic properties of a regular singing
  12 voice is very different from a growling voice, let alone speech.
  13
  14 Secondly, it would be interesting if a model could be trained that could
  15 discriminate a singing voice for all styles of singing including growling.
  16 Moreover, it is possible to investigate the performance of detecting growling
  17 on regular singing-voice trained models and the other way around.
  18
  19 %Discussion section
  20 \section{Discussion}
  21 The dataset used is not very big. Only three albums are annotated and used
  22 as training data. The albums chosen do represent the ends of the spectrum and
  23 therefore the resulting model can be very general. However, it could also mean
  24 that the model is able to recognize three islands in the entire space of
  25 grunting. This does not seem the case since the results show that totally alien
  26 data also has a good performance.
  27
  28 The model clearly has trouble with pauses between singing.
  29
  30 \emph{Singing}-voice detection and \emph{singer}-voice Singing-voice detection
  31 can be seen as a crude way of genre-discrimination. Therefore it be
  32 generalizable to extensive genre recognition
  33 might.