User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
Admin
completely blind computer geek, lover of science fiction and fantasy (especially LitRPG). I work in accessibility, but my opinions are my own, not that of my employer. Fandoms: Harry Potter, Discworld, My Little Pony: Friendship is Magic, Buffy, Dead Like Me, Glee, and I'll read fanfic of pretty much anything that crosses over with one of those.
keyoxide: aspe:keyoxide.org:PFAQDLXSBNO7MZRNPUMWWKQ7TQ
Location Ottawa
Birthday 1987-12-20
Pronouns he/him (EN)
xmpp fastfinge@im.interfree.ca
keyoxide aspe:keyoxide.org:PFAQDLXSBNO7MZRNPUMWWKQ7TQ
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg Yeah true. I remember one time I was on the train, and a random sighted person said to me "Hey! There's another blind person just down the car! Don't you two know each other? Why aren't you sitting together? Here, I'll take you to him." I was like "No, I'm traveling alone. We're not together." Then the other blind person overheard the conversation, and it turned out we'd known each other for years. So we sat together and chatted for the rest of the train ride. I was so tempted to pretend I didn't know him at all, just so I didn't validate this random sighted person's stereotypes.
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg Yeah, that's the guy LOL. We're re-enforcing the stereotype that all blind people know each other!
2
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg Also, could we maybe get a neural model that just overrides espeaks stress marks? That seems to be the main place espeak falls down.
2
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg I feel like it's caching something, somewhere. If I delete the addon and reinstall it, it has all different problems each time.
2
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg Also, another stress test I've been using on text to speech systems lately is the name of my friend "Hrvoje". It's croatian, and pronounced "her voy yay". Every AI text to speech system does something new and awful with it. So does every klatt system haha.
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg I'm not actually having any of those issues. I'm having other ones, though. Like for me "notifications" is said "norifications". But "g" is said fine. As is ctrl and alt.
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg "Well they should just learn English. Idiots. They should stop being blind while they're at it." -- probably some big tech CEO, somewhere
3
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg Right, but they just used codex and probably included all of torch. There are onnxruntime versions available, and compiled versions in rust.
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg I've also been meaning to look into this. They advertise streaming TTS lag at faster than realtime on CPU, with 100 MS lag or less: github.com/kyutai-labs/pocket-tts?tab=readme-ov-file
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg It's possible to get Onnxruntime to be snappy with a correctly optimized model, though. Blastbay TTS uses it.
0
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg I suspect you're going to wind up training your own g2p model. Eloquence can already output its phonemes, so you'd just have to write a script to convert from eloquence to IPA, and then you could just make a bunch of training data.
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg Also some phoneme tuning: thread and threat sound almost identical.
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg The lack of prosody markers is a blocker, though.
2
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg Yeah, not quite what we need. But a step in the right direction...
2
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
Tagging @Tamasg odorediamanka600-source/FYLs-G2P: A lightweight hybrid G2P engine with less than 1.8M parameters and can be deployed on any devices (almost) github.com/odorediamanka600-source/FYLs-G2P
2
0
1
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
Openclaw Is More Fun to Fiddle With Than To Use: Of course I have to write about , the latest viral sensation: stuff.interfree.ca/2026/02/06/openclaw.html
1
9
4
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@kaveinthran @Tamasg @nick It might be because the newer versions of the driver had settings that the older one didn't, and it's causing a crash. I'm really not sure; everything is different about the driver since I wrote that.
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@kaveinthran @Tamasg @nick It only includes support for 64-bit NVDA, because gruut needs 64-bit.
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@Tamasg @kaveinthran @nick Here, the version that used an ancient version of speech player was still on my file host. I won't be updating it, but if you want to have a play, here you go: share.interfree.ca/app/open/AX2qBbXiJ8D-2wZ6znjs4bv-nZAAAyQ1zkB-9FHwonCxUDQ?view=1
1
1
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
4mo
@kaveinthran @Tamasg @nick Unlikely. It needed, like, 700 megs of just Python dependencies. And loading all that into NVDA slowed the entire thing down hugely.
1
0
0
0