Hmm. Bakuda is a bit more understandable than last episode but its still not that great. There are also some ups and downs in her pitch, making the whole thing not fitting with the in story description, which I wager was the point of the robot voice in the first place? Maybe use a bit less filter?
Yea, I really hate how I did Bakuda’s voice. One of the downsides to recording most of an Arc at once is that if you decide you don’t like a voice a few days later you need to change *all* the chapters in the Arc instead of just a few. My original idea for her was to try and do what I thought her actual voice would sound like, then find a way to force it to be monotone via something like autotune and add some static to it. Unfortunately – as you can hear – that plan didn’t work out the way I wanted it to. It also caused the editing time for her voice to take about 10x longer than it normally takes me. One of the problems with listening to my own recordings is that *I* can tell what she’s saying because I’m the one that said it, but I can’t tell how understandable it actually is to other people without their feedback.
Did you keep the original versions before editing? Because if you did, redoing the whole reading wouldn’t be needed. Just an appliance of the filters that works better.
Fortunately I do still have the project files for all the chapters I’ve recorded, but unfortunately there is no effect (that I’ve been able to find) that will produce the sound of Bakuda’s voice from the inflection-heavy original lines. If I just change the effects I used it still won’t sound like it’s supposed to. When I re-record them I’m going to do my best to have as little inflection as possible so the editing required should be pretty minimal.
hi! thank you so so much for putting all this together, it makes my work day much more fun! i don’t know if you read these anymore as it’s been years, but im a new fan and this audiobook has been so helpful. i am wondering, did you consider using a text to speech for Bakuda? if you did, im wondering why you made the choice to stick with a filtered voice? thank you again for all of your hard work, you have a very nice voice to listen to!
I do still read these and I’m glad you’re enjoying the audiobook!
I think I tried a TTS for her voice and really didn’t like how it sounded. TTS tech has come a LONG way since I made this so nowadays it would be a no-brainer, but back then it was not a lot better than Microsoft Sam lol.
Hmm. Bakuda is a bit more understandable than last episode but its still not that great. There are also some ups and downs in her pitch, making the whole thing not fitting with the in story description, which I wager was the point of the robot voice in the first place? Maybe use a bit less filter?
Yea, I really hate how I did Bakuda’s voice. One of the downsides to recording most of an Arc at once is that if you decide you don’t like a voice a few days later you need to change *all* the chapters in the Arc instead of just a few. My original idea for her was to try and do what I thought her actual voice would sound like, then find a way to force it to be monotone via something like autotune and add some static to it. Unfortunately – as you can hear – that plan didn’t work out the way I wanted it to. It also caused the editing time for her voice to take about 10x longer than it normally takes me. One of the problems with listening to my own recordings is that *I* can tell what she’s saying because I’m the one that said it, but I can’t tell how understandable it actually is to other people without their feedback.
Did you keep the original versions before editing? Because if you did, redoing the whole reading wouldn’t be needed. Just an appliance of the filters that works better.
Fortunately I do still have the project files for all the chapters I’ve recorded, but unfortunately there is no effect (that I’ve been able to find) that will produce the sound of Bakuda’s voice from the inflection-heavy original lines. If I just change the effects I used it still won’t sound like it’s supposed to. When I re-record them I’m going to do my best to have as little inflection as possible so the editing required should be pretty minimal.
It would still be better if more inflection comes through if the positive tradeoff is that people can actually understand everything.
hi! thank you so so much for putting all this together, it makes my work day much more fun! i don’t know if you read these anymore as it’s been years, but im a new fan and this audiobook has been so helpful. i am wondering, did you consider using a text to speech for Bakuda? if you did, im wondering why you made the choice to stick with a filtered voice? thank you again for all of your hard work, you have a very nice voice to listen to!
I do still read these and I’m glad you’re enjoying the audiobook!
I think I tried a TTS for her voice and really didn’t like how it sounded. TTS tech has come a LONG way since I made this so nowadays it would be a no-brainer, but back then it was not a lot better than Microsoft Sam lol.