[Hanzhang X]

Thanks!

[Hanzhang X]

Sorry I forgot to mention that. I used the default setting for warmup_epoch (20). What’s the difference between training_epochs and warmup_epoch?

I did not realize the problem of over fitting. Thanks for pointing that out and I will not change the number of epochs then.

[Hanzhang X]

Thanks!

[Hanzhang X]

I am working on a Dice machine and have my own copy of Ossian…

[Hanzhang X]

I did not figure the problem out, but when I ran the scripts again, it worked…

[Hanzhang X]

The first problem is solved. It turned out that I got the path of the output files wrong.

However, there is still an error message during synthesis:

Traceback (most recent call last):
File “./scripts/speak.py”, line 168, in <module>
main_work()
File “./scripts/speak.py”, line 118, in main_work
output_labfile=output_labfile)
File “/afs/inf.ed.ac.uk/user/s15/s1565865/Thesis/ossian_msc_2016_test/Ossian/scripts/main/Voice.py”, line 224, in synth_utterance
processor.apply_to_utt(utt, voice_mode=self.run_mode) ## utt is changed in place
File “/afs/inf.ed.ac.uk/user/s15/s1565865/Thesis/ossian_msc_2016_test/Ossian/scripts/processors/UtteranceProcessor.py”, line 89, in apply_to_utt
self.process_utterance(utterance)
File “/afs/inf.ed.ac.uk/user/s15/s1565865/Thesis/ossian_msc_2016_test/Ossian/scripts/processors/NN.py”, line 649, in process_utterance
self.world_resynth(streams, owave)
File “/afs/inf.ed.ac.uk/user/s15/s1565865/Thesis/ossian_msc_2016_test/Ossian/scripts/processors/NN.py”, line 671, in world_resynth
put_speech(data, ‘/tmp/tmp.%s’%(stream))
File “/afs/inf.ed.ac.uk/user/s15/s1565865/Thesis/ossian_msc_2016_test/Ossian/scripts/util/speech_manip.py”, line 43, in put_speech
write_floats(flat_data, outfile)
File “/afs/inf.ed.ac.uk/user/s15/s1565865/Thesis/ossian_msc_2016_test/Ossian/scripts/util/speech_manip.py”, line 50, in write_floats
f = open(outfile, “w”)
IOError: [Errno 13] Permission denied: u’/tmp/tmp.mgc’

I checked the “output” folder, there are files in folder “dnn_lab” and “lab_dur”, but nothing is in “utt” or “wav”.

[Hanzhang X]

A related question: It seems to me that the biggest difference between IFF and ASF is that IFF measures the distance between target and unit by weighted linguistic features, whereas ASF measures that by acoustic features. Is it right? Since these acoustic features are determined by linguistic features, there should be some direct mappings between linguistic features to acoustic features. So it seems to me that the two approaches are very similar. I do not understand why Taylor says they are two distinct approaches.

[Author]

Posts