takehika
@takehika.bsky.social
Specializing in audio tech development, leveraging AI/ML and data analysis. Freelancer in Japan.
Why are there three?
November 10, 2025 at 4:44 AM
Why are there three?
Took a break and stopped by the YAMAHA cafe in Yokohama Minato Mirai.
It was fun to get hands-on with various instruments like guitars, drums, and violins in their trial space!
retailing.jp.yamaha.com/shop/yokoham...
It was fun to get hands-on with various instruments like guitars, drums, and violins in their trial space!
retailing.jp.yamaha.com/shop/yokoham...
November 4, 2025 at 12:42 AM
Took a break and stopped by the YAMAHA cafe in Yokohama Minato Mirai.
It was fun to get hands-on with various instruments like guitars, drums, and violins in their trial space!
retailing.jp.yamaha.com/shop/yokoham...
It was fun to get hands-on with various instruments like guitars, drums, and violins in their trial space!
retailing.jp.yamaha.com/shop/yokoham...
Digital sound from Sora 2.
Theme: "motion against nature’s vastness"
Theme: "motion against nature’s vastness"
October 31, 2025 at 4:48 AM
Digital sound from Sora 2.
Theme: "motion against nature’s vastness"
Theme: "motion against nature’s vastness"
Digital audio from Sora 2.
It feels like there's still plenty of room for improvement after some testing.
It feels like there's still plenty of room for improvement after some testing.
October 31, 2025 at 4:40 AM
Digital audio from Sora 2.
It feels like there's still plenty of room for improvement after some testing.
It feels like there's still plenty of room for improvement after some testing.
Just got the update notification for ChatGPT.
Is this readable for you?
Is this readable for you?
October 30, 2025 at 10:56 PM
Just got the update notification for ChatGPT.
Is this readable for you?
Is this readable for you?
OpenAI released a new "gpt-4o-transcribe-diarize" model in the Transcription API.
October 22, 2025 at 4:38 AM
OpenAI released a new "gpt-4o-transcribe-diarize" model in the Transcription API.
in 4 days 21 hours 59 minutes....
September 13, 2025 at 7:11 AM
in 4 days 21 hours 59 minutes....
microsoft/VibeVoice-1.5B on Hugging Face.
huggingface.co/microsoft/Vi...
It can generate multi-speaker conversational audio from text, supporting up to 4 speakers and 90 minutes.
You can even try it yourself with the demo page on GitHub.
huggingface.co/microsoft/Vi...
It can generate multi-speaker conversational audio from text, supporting up to 4 speakers and 90 minutes.
You can even try it yourself with the demo page on GitHub.
August 27, 2025 at 5:07 AM
microsoft/VibeVoice-1.5B on Hugging Face.
huggingface.co/microsoft/Vi...
It can generate multi-speaker conversational audio from text, supporting up to 4 speakers and 90 minutes.
You can even try it yourself with the demo page on GitHub.
huggingface.co/microsoft/Vi...
It can generate multi-speaker conversational audio from text, supporting up to 4 speakers and 90 minutes.
You can even try it yourself with the demo page on GitHub.
Voice-first desktop now has a chat panel on the left and a content panel on the right. I'm aiming for a dynamic UI that displays content as you speak.
I'm currently using both ChatGPT and Gemini. ChatGPT's API is costly but very fast, so it might become the primary engine.
I'm currently using both ChatGPT and Gemini. ChatGPT's API is costly but very fast, so it might become the primary engine.
August 20, 2025 at 6:56 AM
Voice-first desktop now has a chat panel on the left and a content panel on the right. I'm aiming for a dynamic UI that displays content as you speak.
I'm currently using both ChatGPT and Gemini. ChatGPT's API is costly but very fast, so it might become the primary engine.
I'm currently using both ChatGPT and Gemini. ChatGPT's API is costly but very fast, so it might become the primary engine.
Audio Analyzer can now do transcription, speaker diarization, emotion and acoustic analysis, and sound classification for a single audio file. Next up: implementing a model for audio separation and improving the UI.
Wow, what an uninspired UI...
Wow, what an uninspired UI...
August 20, 2025 at 6:31 AM
Audio Analyzer can now do transcription, speaker diarization, emotion and acoustic analysis, and sound classification for a single audio file. Next up: implementing a model for audio separation and improving the UI.
Wow, what an uninspired UI...
Wow, what an uninspired UI...
Curious why Onitsuka Tiger is a hit with tourists?
This picture's from Ginza back in April, but the same shop had a massive line again just yesterday!
This picture's from Ginza back in April, but the same shop had a massive line again just yesterday!
June 23, 2025 at 1:03 AM
Curious why Onitsuka Tiger is a hit with tourists?
This picture's from Ginza back in April, but the same shop had a massive line again just yesterday!
This picture's from Ginza back in April, but the same shop had a massive line again just yesterday!
Wanna grab a drink again tonight?
June 17, 2025 at 7:50 AM
Wanna grab a drink again tonight?
I threw out my back just from sneezing, and this is the situation now.
My doctor ordered me not to sit down at all until it's healed, but you know, I'm finding that working on my feet isn't so bad after all.
My doctor ordered me not to sit down at all until it's healed, but you know, I'm finding that working on my feet isn't so bad after all.
May 26, 2025 at 5:16 AM
I threw out my back just from sneezing, and this is the situation now.
My doctor ordered me not to sit down at all until it's healed, but you know, I'm finding that working on my feet isn't so bad after all.
My doctor ordered me not to sit down at all until it's healed, but you know, I'm finding that working on my feet isn't so bad after all.
Ebiya(海老屋), antique shop at Nihonbashi, Muromachi (Tokyo).
April 23, 2025 at 12:21 PM
Ebiya(海老屋), antique shop at Nihonbashi, Muromachi (Tokyo).
You can eat delicious miso ramen near Tokyo Station for under 6 USD (780 yen).
I don't want visitors to have the idea that ramen is expensive.
I don't want visitors to have the idea that ramen is expensive.
April 22, 2025 at 1:05 PM
You can eat delicious miso ramen near Tokyo Station for under 6 USD (780 yen).
I don't want visitors to have the idea that ramen is expensive.
I don't want visitors to have the idea that ramen is expensive.
This is “Muromachi” area, traditional shops and modern surroundings.
15 minutes from Tokyo station on foot.
15 minutes from Tokyo station on foot.
April 21, 2025 at 8:28 PM
This is “Muromachi” area, traditional shops and modern surroundings.
15 minutes from Tokyo station on foot.
15 minutes from Tokyo station on foot.
Tokyo station. Lots of visitors from around the world.
April 21, 2025 at 12:47 PM
Tokyo station. Lots of visitors from around the world.
First, create a primary image as the source. Then, make changes to it.
April 15, 2025 at 11:21 PM
First, create a primary image as the source. Then, make changes to it.
I can't correct the squares and the placement of the pieces on the Shogi board.
April 14, 2025 at 11:50 PM
I can't correct the squares and the placement of the pieces on the Shogi board.
"OpenAI Agents SDK" is good to create a chain voice reaction (speech-to-text → LLM → text-to-speech), I guess...
April 13, 2025 at 12:53 AM
"OpenAI Agents SDK" is good to create a chain voice reaction (speech-to-text → LLM → text-to-speech), I guess...
The best cherry blossom season is here.
This is the Odawara Castle in Kanagawa.
This is the Odawara Castle in Kanagawa.
April 6, 2025 at 12:06 PM
The best cherry blossom season is here.
This is the Odawara Castle in Kanagawa.
This is the Odawara Castle in Kanagawa.
Mt. Fuji from a plane (to Taipei).
March 21, 2025 at 12:00 PM
Mt. Fuji from a plane (to Taipei).
I'm not sure this structure is really a good one for web scraping (generated by ChatGPT), but let's try.
March 8, 2025 at 11:44 PM
I'm not sure this structure is really a good one for web scraping (generated by ChatGPT), but let's try.
Apple Intelligence says this is me... wow
February 12, 2025 at 5:43 AM
Apple Intelligence says this is me... wow