TTS on Mia Heidenstedthttps://heidenstedt.org/tags/tts/Recent content in TTS on Mia Heidenstedt HugoenSun, 06 Jul 2025 15:49:51 +0000TTS on my Bloghttps://heidenstedt.org/posts/2025/tts-on-my-blog/Fri, 17 Jan 2025 16:06:00 +0200https://heidenstedt.org/posts/2025/tts-on-my-blog/ Best viewed on the original page, where extended functionality like the footnote helper is available.

For specific posts, I added an audio version to enhance accessability and maybe it is just nice to have something read it to you.
I am using Kokoro-TTS which is the first open weight TTS model that has sufficient quality for me to use it on my blog.

It is so nice to be finally able to use this technology, truly a dream come true.
Also i am looking into the possibility to use it for books that have no audio version yet. There is a good blogpost about how to convert an epub to an audiobook using this model and audiblez. But i have found that the emotional range and pauses of the result is not authentic enough and this kinda hurts understandability in situations where you are not focused on the audio. I hope future models will be able to bring a bit more life into the audio.

]]> Best viewed on the original page, where extended functionality like the footnote helper is available.

For specific posts, I added an audio version to enhance accessability and maybe it is just nice to have something read it to you.
I am using Kokoro-TTS which is the first open weight TTS model that has sufficient quality for me to use it on my blog.

It is so nice to be finally able to use this technology, truly a dream come true.
Also i am looking into the possibility to use it for books that have no audio version yet. There is a good blogpost about how to convert an epub to an audiobook using this model and audiblez. But i have found that the emotional range and pauses of the result is not authentic enough and this kinda hurts understandability in situations where you are not focused on the audio. I hope future models will be able to bring a bit more life into the audio.

]]>