drylbear / jointascandsed

jointASCandSED

jointascandsed's Issues

Problems of work reproduction

Dear author,

I am trying to reproduce your paper (Towards joint sound scene and polyphonic sound event recognition), but I have encountered some problems and hope to get your reply.

Q1: The paper says: "the final sound scene and event dataset consists of 3000", but the dataset I downloaded contains two folders: train and test, each of which contains 3000 "wav" audio files. There are 6,000 in total, which seems to be inconsistent with the description in the paper. Should I only use the files in the train folder?

Q2: The paper says "using stratified five-fold cross validation", could you please tell me how you divide it?

Q3: I see here (https://github.com/drylbear/jointASCandSED/blob/master/feat_extraction.ipynb), there are related files for dividing the dataset, but it lacks specific functions, such as: readEventsForScenes, countBackgrounds, etc., could you please share the complete code with me?

Hope to get your help.