C.W.Holeman III

• Author  ⥂ Personal
•   ☰  
  • Books
  • Current Projects
  • Setting: The Paradigm Codex
  • Back Burner Projects
  • Me
  • About
  • Contact
  • Support Me
  • Resources
  • Writing Tools
  • My Editing Process
  • Writing Resources
• Δ Fellow Authors
  • Carrina Nieman
  • Iver Gullixson
  • Jack Gallegos
  • Jen Fir

Right Bar

HST Clock

LOADING HST TIME

Books in Progress

Capitol Crimes & Ghastly Galas

First Draft

A Dakota Entendre

First Draft Second Draft

Wizard Tales & Lizard Beards

First Draft
Second Draft 10%
Third Draft 10% Querying 9%

Random Quote

Joke of the Week

There are only 10 kinds of people in the world: Those who know binary, and those who don't.

Writing Tools

Depending on which project I'm working on, I use a wide variety of tools.

List of Tools

Software

• Obsidian.md (Free) - World Bibles, Drafting, & Editing.
• Scrivener ($60) - Organizing, drafting, & editing. No mobile option for Android.
• Easy Voice Recorder Pro ($5/Free) - Recording audio for transcription. The ad supported version works just as well.
• Whisper AI (Free) - Voice-to-text AI, runs locally on personal computer. (Install Guide)
• Ollama & Llava (Free) - LLM (Large Language Model), runs locally on personal computer. (Install Guide)
• Resilio Sync Pro ($60/Free) - Syncing files between phone, computers, NAS. (The free version has everything needed but lacks File Versioning, which is nice to have.)
• Custom PowerShell script (Download ).
• Custom Bash script (Download ).

Hardware

• My local computer. Specs:
  

  CPU	AMD Ryzen 7 5700G: Cores/Threads: 8/16
  GPU	RTX 3070 with 8GB RAM
  RAM	32GB (2x16)
  Storage	512GB NVME SSD, 2TB HDD (x2)
  Motherboard	ASRock B450M PRO4
  OS	Windows 10 (yuck!) Windows Subsystem for Linux (yay!)

• OnePlus 8T (Android phone)
• SHOKZ OpenComm2 ($160) - Handy for on the go (driving, shopping, hiking), but not strictly necessary.
  

Step-by-Step Drafting Workflow

Current Workflow

1. I use EVRP (Easy Voice Recorder) to make a recording of a chapter or section of a story.
2. EVRP saves the recording into a specified folder on my (Android) phone.
3. Resilio Sync then syncs the .m4a audio file to my desktop computer and NAS.
4. My computer has a PowerShell script (which is auto-launched on bootup by Windows Task Scheduler) which watches the directory that the files are synced to. When it sees a new .m4a file, it uses Whisper AI to convert the raw audio file to an ugly text file.
5. Using a custom Bash script, I kick off a run of Ollama with Llava to convert the ugly text file to something approximating clean, formatted, text.
6. Then I copy & paste that clean text into Scrivener/Obsidian (depending on the project) for further refining & editing.

The Next (Upcoming) Evolution

• More streamlined steps, with 0-touches to go from the saved audio file to the clean text file.
• Fine-tuning the Llava model with my notes, outlines, and existing stories to help it aim at my exact wording better.

Aspirationally

• I'm hoping I can get the clean text directly into Scrivener (doing so for Obsidian is trivial).
• I also want to use Ollama to do much of the grunt work of editing; to use my previous writings to tighten up the text to match what I wanted to say while recording, rather than whatever I happened to ramble.

Still have questions? Ask me.