I didn’t listen to Tanner, since I know nothing about regional British accents. I did listen to both Henry and Oliver, and I thought their voices were clear enough, although Henry’s seemed rather … cheerful? … for a tense opening scene. Which, of course, a text to voice program wouldn’t be expected to recognize and adjust for.
I’ve tried the text to voice program on my iPad for “reading” fanfic. It’s definitely easier on the eyes, which I appreciate since I have vision issues. The program is okay, and some voices are better than others, but in the iPad program at least there were a lot of odd pronunciations and a general lack of expression.
I also found it hard to get past the fact that the voices aren’t the right voices for my characters, though. In dialogue, especially banter, I hear the characters’ voices in my head when I read well written dialogue, but I wasn’t able to do that with voice to text. Do you get used to it eventually and hear your characters’ voices?
In the beginning, I sometimes felt that Henry's delivery clashed with the mood of a scene. As I listened to him more, I think I got used to it, though. It no longer sounds cheerful to me, but just like the way he talks.
I read mostly on an iPad. Tom, the voice I linked at the end of the post, is an Apple voice. Apple's accessibility settings include an interface for correcting its voices' pronunciations, which is a good feature that Speechify lacks. (If you search Apple's Settings for "pronunciation" you should find it.) The Speechify voices are definitely more natural, though, and understand punctuation better. There are also some Speechify voices labeled "AI Enhanced" that try to adjust their delivery to the meaning of the text. I don't like those because they sometimes get it jarringly wrong, but I expect they will improve with time.
After a lot of time spent listening to computers read, my mind does supply a lot of the missing expression and sometimes even character voices. If you think about it, when you read visually
( ... )
I didn’t find the “pronunciation” setting on the iPad program all that helpful. I was able to teach it to pronounce “Waverly” but I never was successful in getting it to pronounce “Illya” correctly. And a lot of the mispronounced words are just because the selected voice runs words together or cuts them off oddly, which varies from voice to voice and isn’t really a problem that can be fixed by “correcting” the pronunciation.
I also got really frustrated when I spent a lot of time and hard drive space downloading the enhanced version of a voice I liked (it was called Nathan). But every time I select it - snd despite the fact it’s downloaded and selecfed - when I try to use it, the program substitutes a mechanical female voice.
Yes, the highest quality version of Nathan was buggy. I had the same issue. They have since removed it altogether.
As far as the pronunciation feature goes, you might try again. The tool that allows you to speak the pronunciation you want and have it transcribed phonetically has finally been fixed. (Due partly to my own complaints, I suspect.)
In the end, though, TTS tools only work for people for whom visual reading is hard enough to make the various frustrations worth it. I’ve been in that group for a long time. As the technology improves and the frustrations shrink, the group is growing to include more and more people. Maybe it doesn’t yet include you, though.
Yes, the highest quality version of Nathan was buggy. I had the same issue. They have since removed it altogether.
Unfortunately, they won’t allow me to delete it off my device, even though it’s completely non-functional. Using “edit” I can delete the smaller unenhanced Nathan file but not the big enhanced one.
I tried downloading the enhanced Tom, but when I select it, it sounds like the clunky unenhanced Tom. The only enhanced voice I’ve been able to get to work properly is the UK Daniel … which isn’t a bad voice, honestly, but which sounds kind of odd to me when playing MFU. But maybe I’d get used to it if I used if long enough.
Hm. That does sound like a high dose of frustration.
It sounds like, like me, you prefer male voices. I’ve just been experimenting with the Siri voices. I’d discounted them in the past, but they are actually quite natural, if maybe a bit perky. You might want to try out US Siri Voice 1 and see what you think.
I probably wouldn’t care whether the voice was male or female for listening to something like a news article. What I’ve mostly tried to listen to with the text to voice program, though, is MFU fanfiction, and that does sound better to me in a male voice, probably because it’s almost always written in the point of view of a male character. But I suppose if I found a really good female voice it might be better than a poor male one.
And maybe the technology will continue to improve as time goes on. I get frustrated by the voices not pausing at punctuation points or paragraph breaks and the lack of inflection in dialogue, but for the most part the technology isn’t “there” yet.
I’ve tried the text to voice program on my iPad for “reading” fanfic. It’s definitely easier on the eyes, which I appreciate since I have vision issues. The program is okay, and some voices are better than others, but in the iPad program at least there were a lot of odd pronunciations and a general lack of expression.
I also found it hard to get past the fact that the voices aren’t the right voices for my characters, though. In dialogue, especially banter, I hear the characters’ voices in my head when I read well written dialogue, but I wasn’t able to do that with voice to text. Do you get used to it eventually and hear your characters’ voices?
Reply
In the beginning, I sometimes felt that Henry's delivery clashed with the mood of a scene. As I listened to him more, I think I got used to it, though. It no longer sounds cheerful to me, but just like the way he talks.
I read mostly on an iPad. Tom, the voice I linked at the end of the post, is an Apple voice. Apple's accessibility settings include an interface for correcting its voices' pronunciations, which is a good feature that Speechify lacks. (If you search Apple's Settings for "pronunciation" you should find it.) The Speechify voices are definitely more natural, though, and understand punctuation better. There are also some Speechify voices labeled "AI Enhanced" that try to adjust their delivery to the meaning of the text. I don't like those because they sometimes get it jarringly wrong, but I expect they will improve with time.
After a lot of time spent listening to computers read, my mind does supply a lot of the missing expression and sometimes even character voices. If you think about it, when you read visually ( ... )
Reply
I also got really frustrated when I spent a lot of time and hard drive space downloading the enhanced version of a voice I liked (it was called Nathan). But every time I select it - snd despite the fact it’s downloaded and selecfed - when I try to use it, the program substitutes a mechanical female voice.
Reply
Yes, the highest quality version of Nathan was buggy. I had the same issue. They have since removed it altogether.
As far as the pronunciation feature goes, you might try again. The tool that allows you to speak the pronunciation you want and have it transcribed phonetically has finally been fixed. (Due partly to my own complaints, I suspect.)
In the end, though, TTS tools only work for people for whom visual reading is hard enough to make the various frustrations worth it. I’ve been in that group for a long time. As the technology improves and the frustrations shrink, the group is growing to include more and more people. Maybe it doesn’t yet include you, though.
Reply
Unfortunately, they won’t allow me to delete it off my device, even though it’s completely non-functional. Using “edit” I can delete the smaller unenhanced Nathan file but not the big enhanced one.
I tried downloading the enhanced Tom, but when I select it, it sounds like the clunky unenhanced Tom. The only enhanced voice I’ve been able to get to work properly is the UK Daniel … which isn’t a bad voice, honestly, but which sounds kind of odd to me when playing MFU. But maybe I’d get used to it if I used if long enough.
Reply
Hm. That does sound like a high dose of frustration.
It sounds like, like me, you prefer male voices. I’ve just been experimenting with the Siri voices. I’d discounted them in the past, but they are actually quite natural, if maybe a bit perky. You might want to try out US Siri Voice 1 and see what you think.
Reply
I probably wouldn’t care whether the voice was male or female for listening to something like a news article. What I’ve mostly tried to listen to with the text to voice program, though, is MFU fanfiction, and that does sound better to me in a male voice, probably because it’s almost always written in the point of view of a male character. But I suppose if I found a really good female voice it might be better than a poor male one.
And maybe the technology will continue to improve as time goes on. I get frustrated by the voices not pausing at punctuation points or paragraph breaks and the lack of inflection in dialogue, but for the most part the technology isn’t “there” yet.
Reply
Leave a comment