#tts | Anand S - Things I Learned

Mon, Nov 24, 2025. 1 second = 10 tokens for OpenAI Realtime APIs. 1 second = 25 tokens for Gemini Live API #pricing #speech-to-text #tts
- 39 cents / hour on GPT Realtime Mini = 36 cents audio input + 3 cents text output
- 139 cents / hour on GPT Realtime = 115 cents audio input + 15 cents text output
- 30 cents / hour on Gemini 2.5 Flash Native Audio (Live API) = 27 cents audio input + 3 cents text output

Sat, Nov 22, 2025. Models read pretty fast, consuming input tokens at ~4K-20K words per second. It's the "speaking" (output token rate) that is the bottleneck. So shortening input doesn't matter as much as shortening output for latence. ChatGPT #chatgpt #speech-to-text #tts #voice-cloning

Thu, Nov 20, 2025. TOON vs JSON. Early days, and TOON seems to be marketing a lot, so I'm wary, but for large tabular data where input tokens are crunched, it seems a readable alternative to multiple CSVs, but not worth the hype. #chatgpt #gpu #markdown #tts 0 19 Nov 2025. Always use GPT-5.1-Codex-Max instead of GPT-5.1-Codex. At every thinking level, it takes fewer tokens for similar or higher accuracy. Tibo

Sun, Nov 2, 2025. OpenAI TTS costs are confusing. But in short #speech-to-text #tts #voice-cloning
- TTS-1 costs $15 / MChars (max 4,096 chars per request), which ends up at ~86c / hour
- GPT-4o Mini TTS costs ~$16 / MChars (max 2K tokens which is ~7,000 chars per request), which ends up at ~88c / hour. Very similar cost, effectively
- TTS-1 HD is twice TTS-1.

Tue, Aug 19, 2025. In another example of vibe coding, an instructor for my TDS course vibe-coded most of an exam using Copilot and Sonnet. 6/8 questions worked one-shot. The two failures were interesting: #ai-coding #code-agents #tts
- One failed because of sample vs population stats. Copilot asked for sample variance but coded variance() instead of sampleVariance().
- Another failed because of rounding off. NumPy code rounds off differently from Python or JS code.

Fri, Aug 15, 2025. For live transcription, Gemini 2.5 Flash Live costs 0.6c/min of audio ($3/MTok x 32 tokens/second) while GPT 4o Mini Realtime costs ~2c/min and GPT 4o Realtime costs ~8c/min. ChatGPT #chatgpt #speech-to-text #tts

Tue, Jun 3, 2025. At the moment, the best speech to text for Android appears to be ChatGPT's transcription. The default Android text to speech (which I thought was good) no longer feels adequate. Gemini mis-hears and doesn't wait till I'm done. Whisper ASR has poor noise cancellation and a 30 second limit. #chatgpt #speech-to-text #tts #voice-cloning

Thu, May 29, 2025. Today, I'd go with Node's native test runner for backend JS testing. I used node-tap earlier. For front-end, I'd pick vitest. ChatGPT #chatgpt #github #tts

Sat, May 24, 2025. TTS typically costs $1/hour now. Gemini 2.5 Flash Preview TTS, Gemini 2.5 Pro Preview TTS, GPT 4o TTS, and GPT 4o Mini TTS are the current best-in-class text-to-speech models from the mainstream LLM providers. Assuming ~175 words per minute and 1 token ≈ ¾ words, 1 hour of speech ~ 10,300 words/hr ~ 13,800 input tokens ~ 75,000 audio tokens, it costs: #speech-to-text #tts #voice-cloning
- Gemini 2.5 Flash Preview TTS ($0.50/1 M input, $10.00/1 M output): ~$0.8 per hour
- GPT-4o-mini-TTS ($0.60/1 M input, $12.00/1 M output): ~$0.9/hour
- Gemini 2.5 Pro Preview TTS ($1.00/1 M input, $20.00/1 M output): ~$1.5 per hour
- GPT-4o-TTS (known as gpt-4o-audio-preview, $2.50/1 M input, $80/1 M output): ~$6.0/hour
- This is comparable to the earlier OpenAI Standard TTS ($0.75), OpenAI HD TTS ($1.5), Google Neural2 ($0.8). ElevenLabs Pro costs ~$6/hr.

Tue, Mar 25, 2025. The new GPT-4o mini Transcribe model is a bit better than Whisper and costs half: ~18 cents per hour. It includes background noise cancellation and semantic chunking, which is useful. #speech-to-text #tts #voice-cloning

Tue, Mar 25, 2025. The new GPT-4o mini TTS is about 3-4 times cheaper than TTS-1 since it's ~$12/MTok instead of $15/Mchar. It supports emotions with streaming. #speech-to-text #tts #voice-cloning

Sun, Mar 23, 2025. Phi-4 multimodal procehttps://huggingface.co/microsoft/Phi-4-multimodal-instructsses speech better than Whisper V3 on HuggingFace OpenASR, and images better than Gemini Flash Lite #future #huggingface #speech-to-text #tts #voice-cloning

Fri, Mar 14, 2025. Twine (docs) is an open source interactive fiction / story writing tool. #github #markdown #tts #write
- Snowman is a browser-based Twine 2 story template format.
- These enable behavioural experimentation. Cheaper than using tools like Gorilla.sc and Pavlovia for behavioral experiments
- For example, you can present a social or political issue and see if people change their opinions more or less depending on the content/path they see. Or, if it varies by demographics. Or, check if repeated mentions or emotional hooks improve memory / retention. More research ideas

Wed, Mar 5, 2025. Reliably creating interactive tutorials is hard today. Claude 3.7 Sonnet ran out of tokens when I tried creating an interactive tutorial on diffraction. Cursor got the tokens but failed to get the application right after 3 attempts. This is not yet reliable, and when it does become reliable, education will change a fair bit. #tts #impossible

Sun, Mar 2, 2025. YayText converts text to Unicode that has strikethrough, bold, italics, alternate fonts, and other interesting features. So does #speech-to-text #tts Unitextify, ConvertCase, and LingoJam.

Sun, Jan 12, 2025. TTS Arena is a benchmark of text-to-speech models. Kokoro-TTS is the current leader. It's just 82M, runs on Google Colab, and sounds slightly better than OpenAI TTS. #speech-to-text #tts #voice-cloning

Wed, Jan 8, 2025. whisper-flow does real-time speech transcription! #future #speech-to-text #tts #voice-cloning

Fri, Jan 3, 2025. Assembly AI offers speech to text with diarization at 12c/hour. Good diarization, average transcription quality. #speech-to-text #tts #voice-cloning In comparison, WhisperX (with GPU) was much slower, had slightly poorer diarization, and slightly better transcription.
```
uvx --python 3.9 --index https://download.pytorch.org/whl/cu121 whisperx --diarize --lang en --hf_token $HUGGINGFACE_TOKEN
```

Wed, Dec 18, 2024. GPT 4o Mini Realtime was released. A realtime conversation will cost ~50c/hr. About 36c for input, 72c for output. (I extrapolated from the 6c/min audio input cost for GPT 4o Realtime when it was $100/MTok. GPT 4o Mini Realtime is $10/MTok input and $20/MTok output.) #chatgpt #pricing #speech-to-text #tts

Fri, Nov 29, 2024. GPT-4o Audio supports tone control via XML tags like <cough>..., <laugh>..., etc. But at ~$15/hr of output, it's too expensive. Ref #speech-to-text #tts #voice-cloning

Mon, Nov 11, 2024. Gemini transcription does not give accurate timestamps. Whisper does. But the quality of transcription is similar. #speech-to-text #tts

Mon, Nov 4, 2024. Artificial Analysis has a bunch of new leaderboards and arenas. #speech-to-text #tts
- Open AI TTS leads the TTS Leaderboard. ElevenLabs is a bit behind.
- Recraft V3 > Flux 1.1 leads Text to Image Leaderboard

Mon, Nov 4, 2024. Hertz-Dev is an open source realtime voice chat model. But it doesn't fit in Google Colab T4's RAM #github #tts #voice-cloning

Tue, Oct 29, 2024. F5-TTS clones voices with just 15-second samples. #future #speech-to-text #tts #voice-cloning

Sun, Oct 27, 2024. Elevenlabs lets you create voices with a prompt. No need to even clone one! #speech-to-text #tts #voice-cloning

Wed, Oct 9, 2024. Reverb ASR does diarration as well as transcription. It seems the state of art right now. #speech-to-text #tts #voice-cloning

Tue, Oct 8, 2024. Revisiting text to speech models. Nothing much has changed since July 2024. #speech-to-text #tts #voice-cloning
- OpenAI TTS: $15/1M chars Ref
- Deepgram Aura: $15/1M chars Ref
- Azure AI Speech: $15/1M chars Ref
- Google TTS Neural2: $16/1M chars Ref
- AWS Polly Neural TTS: $16/1M chars Ref
- Cartesia Pro: $50/1M chars Ref
- Elevenlabs Scale: $300/1M chars Ref

Thu, Oct 3, 2024. OpenAI's Realtime API can be used in a text-to-text chat mode without needing to send the entire context. If the pricing works out right, this can be far cheaper than sending the entire conversation context. Ref #speech-to-text #tts

Thu, Oct 3, 2024. ChatGPT's advanced mode includes: "...you can use various regional accents and dialects." Ref Source #chatgpt #speech-to-text #tts #voice-cloning
- But the API can "laugh, whisper, and adhere to tone direction." Ref

Sat, Sep 21, 2024. Sarvam.ai offers Indic text to speech #speech-to-text #tts #voice-cloning

Mon, Jul 8, 2024. A quick check on the pricing of text to speech models #speech-to-text #tts #voice-cloning
- OpenAI TTS: $15/1M chars Ref
- Deepgram Aura: $15/1M chars Ref
- Elevenlabs Scale: $165/1M chars Ref
- Google TTS Neural2: $16/1M chars Ref
- Azure AI Speech: $15/1M chars Ref
- AWS Polly Neural TTS: $16/1M chars Ref

Fri, Feb 23, 2024. Teknoturf is using Gen AI to #ai-coding #automation #prompt-engineering #speech-to-text #tts
- Improve prompts when teaching prompt engineering.
- Pronounce languages better, identifying which words Tamilians and Malayalis will mis-pronounce.

Thu, Dec 21, 2023. I tried running Mixtral-8x7b locally (via Llamafile) and on together.ai. It's good, but far from GPT 4. #chatgpt #tts

Fri, Dec 15, 2023. Mixtral-8x7b-Instruct "... really does seem to be equivalent in quality to ChatGPT 3.5." Ref #chatgpt #speech-to-text #tts

Sun, Dec 3, 2023. Meta released SeamlessExpressive which preserves emotions in speech-to-speech translations #embeddings #future #speech-to-text #tts #voice-cloning

Sat, Nov 4, 2023. 04 Nov 2023: Seamless4MT does language and speech inter-conversion. CC-NC-BY license #document-conversion #speech-to-text #tts #voice-cloning

Thu, Oct 26, 2023. 26 Oct 2023: ChatGPT AutoExpert is a great prompt mechanism for higher accuracy, context and control. Example #automation #chatgpt #code-agents #prompt-engineering #speech-to-text #tts