[Tts-project] festvox voices
elbrus at debian.org
Sat Dec 23 19:32:08 UTC 2017
As you may have noticed, I have worked on several festival/festvox
packages lately. However, I was wondering about some issues, so I
solicit for advice. Mainly because I never use any of these packages myself.
1) We have three¹ 8kHz voices in our archive, with 16kHz siblings. Are
they still interesting to keep? I thing they served their purpose in
another era. Upstream hasn't listed them since 2007 (last one was in
festival 1.96). I propose we request for removal and have the 16kHz
package take over (via provides).
2) Festival/Festvox ship a lot of new voices since 2014². How valuable
do we consider shipping multiple voices? I would expect that these
modern voices have better quality, is it worth it to have so many? I
guess that hearing a voice is a matter of taste and therefor worth to
have a choice. Looking at popcon³ though, it seems most people settle
for one: festvox-kallpc16k, but it may be that that is the default
somehow, as it is the first dependency of festival. As I expect the most
important audience (pun intended) is rather small and a substantial
amount is probably reading this I value your opinion over popcon.
¹ festvox-kallpc8k, festvox-kdlpc8k, festvox-rablpc8k
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 488 bytes
Desc: OpenPGP digital signature
More information about the Tts-project