Please enable JavaScript.
Coggle requires JavaScript to display documents.
PERCEPTUAL VS. AUTOMATED JUDGEMENTS OF MUSIC COPYRIGHT INFRINGEMENT -…
PERCEPTUAL VS. AUTOMATED JUDGEMENTS OF MUSIC
COPYRIGHT INFRINGEMENT
ABSTRACT
Collected perceptual data from 20 participants for 17
adjudicated copyright cases from USA and Japan
predicated that listening to melody-only versions would result in perceptual judgements that more closely matched actual past legal decisions
both algorithms were able to match past decisions with identical accuracy of 71%(12/17 cases)
to classify suggests that melody, lyrics and other factors sometimes interact in complex ways difficult to capture using quantitative metrics
purpose directions for further investigation of the role of similarity in music copyright law using larger and more diverse samples of cases and enhanced methods and adapting our perceptual experiment method
to avoid relying for ground truth data only on court decisions
to important practical debates such as whether jury members should be allowed to listen to full audio recordings during copyright cases
DISCUSSION
significantly correlated with perceptual similarity for both melody-only and full-audio
a perceptually relevant measure of melodic similarity and is consistent with the idea that melodic similarity
both automated methods able to accurately predict 71% (12/17) of previous court decisions while perceptual
accuracy were 58% and 54% % under full-audio and
melody-only conditions respectively
the inclusion of such samples might conceivably have skewed judgements by including levels of dissimilarity rarely included in real court cases.
we failed to collect such
data, any familiarity effects when participants were aware of the cases would be predicted to increase, rather than decrease, accuracy
fact that our prediction was not only not
significant but was in the wrong direction suggests that
limited statistical power cannot explain this result.
The melody-only condition led to higher accuracy for
some cases but lower accuracy for others
(contra predictions).
These discrepancies show
how results from the PMI method can be affected by errors and uncertainties in the transcription process.
FUTURE DIRECTION
the primary limitation of our study at present is its
limited size and scope with a dataset of only 17 court decisions
some of cases include non-musical aspects that make it difficult for current automated methods focusing on musical similarity to identify those exceptions
plan to identify more non-US cases, particularly from Japan and China where music industry revenues are substantial
to expand from a focus purely on music copyright infringement to also include the related domain of cover-song detection