Jun Sasaki / Log

2026-03-25

Rt2112 Development Log | March 22 - March 25, 2026


Latest channel data (as of March 25)

  • 28-day views: 1,260 (previous 1,084)

  • 28-day impressions: 18.3K (previous 17K)

  • Watch time: 68.8 hours

  • CTR: 2.7%

  • Average view duration: 4:54

  • Subscribers: +10

  • 48-hour real-time: 55 views

I changed the policy last time, but I'm rolling it back again.
The reason is one thing: "Auto-dubbing for major languages has arrived."
I'm not changing the focus on the STORY BGM genre, but the accuracy of music separation after auto-dubbing is poor. (The music sounds choppy.)
Since it's a functional issue, I have no choice. In the previous policy, I planned to make the BGM louder during the reading, but this time I decided to make it quieter and separate it into the first half (reading) and the second half (ambient).
I am optimizing the means without changing the goal of the estimated demand.


New unified format (finalized)

Thumbnail

  • Include the text "COZY STORY"

  • A single image that conveys the world of the story

Title

  • 4 Cozy Stories from a Gentle Future 📖 [Story Title]... | Road to 2112 vol.##

  • Remove "Cozy Sci-Fi," "STORY BGM," and "Relax" from the title

  • Keep #cozyscifi in the tags

Intro

  • White text on a black background: "Road to 2112 / Stories from a Gentle Future"

  • Voice: "Road to 2112... four stories from a gentle future"

  • Fixed for all videos

The optimal solution for auto-dubbing and BGM dB

  • BGM -15dB to -18dB: Auto-dubbing triggers (Confirmed all 18 languages triggered in Vol. 15)

  • BGM -10dB: Auto-dubbing does not trigger (Confirmed in Vol. 19, 20, 21)

  • If the BGM is too loud, YouTube's audio separation fails and auto-dubbing is rejected

  • Auto-dubbing trigger = YouTube distributes to 20 language regions = Impressions skyrocket = Directly linked to view counts

  • Future production will be based on -15dB, adjusted to -18dB if it sounds too loud

→ This is just a baseline, as it depends on the volume of the fish. Adjust based on the feeling of whether it's too loud.

Full replacement of videos

  • All 2-hour videos (v5) will be set to private

  • Renumbering Vol. numbers starting from 1-hour videos (Currently published up to Vol. 11)

  • Standardize all videos based on the success pattern of the old Vol. 15 (BGM -15dB, all auto-dubbing triggered)

  • Remake old Vol. 19, 20, and 21, which have not seen an increase in views

  • New Vol. 11 and 12 are currently being tested for auto-dubbing in private.

→ Materials used in the 2-hour videos will be separated, and the BGM and loop animations will be moved to the sub-channel specializing in BGM (Lo-fi 2112). Reading materials will be created from Vol. 11 onwards with new animations.

14-day post-publication schedule strategy

  • Keep as private for 14 days after uploading

  • Set the publication schedule after confirming that auto-dubbing has triggered

  • Maximize impressions by starting the first 48 hours in EN + 18 languages

  • There is a track record of Vol. 15's auto-dubbing arriving after 7 days, and the second one after 10 days

→ This is a point of concern, but even though I replaced them and impressions dropped, I prioritized stocking to maximize the initial response. As a result, I can stock 25 videos in 14 days, so I will just wait for the auto-dubbing to hit. Increasing the stock also provides peace of mind.

Things to do during the 2-week stocking period

  • Published 10 videos to Lo-fi 2112 (ported materials from 2-hour videos)

  • Stocked voice materials for the main channel from Vol. 13 onwards

  • Stocked approximately 25 videos' worth of materials from Kindle volumes 006, 007, 001, and 002

  • Sequential publication starting from those with auto-dubbing activated after 14 days

Auto-dubbing quality check

  • Watched Vol. 15 in Spanish, etc. The quality is quite high

  • There are some parts where the BGM cuts out, but it is generally within an acceptable range at -15dB to -18dB

  • Sound quality for auto-dubbed languages is awaiting improvements from YouTube's side. The only thing we can control is the BGM dB level

Current channel structure

Main Channel (Road to 2112)

  • Position: Cozy Stories reading channel

  • Format: 58 minutes, 4 stories, auto-dubbing in 20 languages

  • BGM: -15dB to -18dB (auto-dubbing activation condition)

  • Thumbnail: 'COZY STORY' + a single image representing the story's world

Sub-channel (Lo-fi 2112)

  • Position: Pure BGM channel

  • Format: Natural length of BGM (1.5 to 2 hours)

  • Materials: Reused from the main channel (loop animation + BGM)

  • A catch-all for the BGM layer. Monetizing the audience that leaves the main channel here

Next steps

  • Fully focused on creating a 14-day stock

  • Added 10 videos to Lo-fi 2112

  • Sequential release on the main channel after confirming auto-dubbing activation

  • Restarting in two weeks with a 25 + 10 video system

Advertisement