Hacker News Reader

I don't know anything about your specific situation, needs, and preferences, but GPUs are not self evidently useless for blind users. In some systems, they can contribute to higher quality speech synthesis. They could also have a role to play in vision support tasks (e.g. determining whether you're properly positioned in front of your webcam in a video chat, or recognizing and describing something you're holding up to your camera).

View on HN

Can you post info about them contributing to speech synthesis? I use NVDA as my screen reader with the built-in espeak synthesizer. I've been using synthetic speech so long I find I don't want natural sounding voices, I prefer more robotic sounding ones since I can listen to them at a faster rate. Vision tasks is a good point but not something that applies to me. I almost never use a webcam on my personal computer. I use my iPone for OCR although I could see a webcam being useful for OCR in some cases.

I don't have information about specific speech synthesis systems, but contemporary speech synthesis often relies on neural networks for best quality, and those can benefit from GPUs. If you truly prefer robotic sounding, that may not be relevant to you. But if it's all about synthesis speed, that's where GPUs may bridge the performance gap.

Reply on news.ycombinator.com