OVERVIEW
Eiri Sanada is a pen name that I began using in 2018 to make my online activities more distinct. Before 2014, I was using my real name but eventually changed my mind. After a few years of using variations of Eiri×× due to name reservations, I mostly settled on eirifu as a username for the majority of places, and Eiri Sanada as a nom de plume.
「Skills, Specialties, Interests」
Video Games: Design, Reviews, QA, Historical Studies, Debugging, UX
Translation: Written Japanese to English in limited quantities, IPA/linguistics, subtitles and Closed Captions
Video Production: Youtube, both Livestream and Standard
「Software Experience」
Audio Engineering: Audacity
Music: Studio One, Piapro Studio (Vocaloid in JP and EN)
Image Editing & Painting: Krita, GIMP
Animation: Pyxel Edit, Live2D Cubism
Video: DaVinci Resolve, OBS Studio
Programming: Lua
Game SDK: Pixel Game Maker MV
「Languages」
English (Native Australian), some Japanese and Vietnamese
For confirmed places where I’m active online, see my homepage and linktree.
Project History
2023 Q3
Reporting on LK-99
When a scientific announcement went viral on the internet, I decided to compile the key information in an article here, and the post itself quickly went viral.
The post has been read and commented upon by people around the globe, consisting of tech enthusiasts, journalists, investors, scholars, engineers, and company executives. I have recieved and seen comments, interview requests and messages from all of these types of people that directly praise my information gathering and summarising skills.
As of the end of 2023, the post has had over 100,000 local views and has been openly copied and plagiarised by several dozen websites, including Wikipedia. It has also been independently translated into Chinese and Korean, and posted to their social media networks.
2023 Q2
Father and Daughter
Just for fun, I translated a Japanese advertisement that went viral. It appears to have since become the primary reference version for the majority of non-Japanese speakers, and is praised for the quality of the editing and translation.
The video has appeared frequently as an embed on sites about creativity, advertising, and copywriting, such as LoveTheWorkMore, and is linked as a reference in interviews and articles from creative agencies, directors, and copywriters.
The translation uses a combination of manual translation for the dialogue and titles, and AI-assistance via ChatGPT for the song lyrics.
As of the end of 2023, the translated video has had approximately 200,000 local impressions and 25,000 views, and continues to grow. (It is unclear if embedded video is consistently tracked.) It has also been copied onto Facebook with several million views.
2023 Q2
Review Essays
I started a highly-edited form of video game review based on my core idea of crossing the game design mindset with the review process.
Instead of plainly reviewing the game, I also think about the game’s idea with a specific game development concept in mind. For example, I reviewed Zelda TotK while cognizant of Nintendo’s evolving definition of Zelda. Then for the next ones, I considered the problem of game genre naming, and inaccurate retro nostalgia.
2023 Q1
Self-Published eBook
As part of a large-scale experiment with the then-new ChatGPT, I used AI to translate a public domain French novel into English and published the results. The intent was to be an optimistic take on AI generation, though I doubt it turned any heads.
Since this was developed in December 2022/January 2023, this may possibly be the world’s first ChatGPT-translated novel.
Non-AI parts include well-researched footnotes that were made in the process of confirming the quality of the translation, and serves as a good reading guide due to the novel’s age. The cover is also completely hand-drawn, with the exception of the ocean pattern.
Due to the nature and age of the story, the non-AI work and footnotes includes research into 19th century scientific terms, 17th century Indian history, 19th century Japanese language and romanisation, 19th century US history for several US states, Native Americans and wildlife, and the history of measurement units, world currency, and monetary inflation.
I self-published the results as a Kindle book, to also serve as formatting/publishing practice. A printed version is also available. A free copy is available on Patreon or by request.
2022 Q3
Compose Man – Level 1
I reduced my streaming activity due to a changing internet landscape, and instead developed a game prototype in Pixel Game Maker MV, which eventually turned into a detailed, fully featured demo. (I’m not sure if there’s much of a difference between prototype and demo at the level of detail I made it in.)
The concept game consists of one level that is equal in content to one full Megaman 11 level, or approx 2.5-3 NES Megaman levels. The level design is reverse-engineered from Megaman 11’s level design principles, while the visual design is an extension of the NES-inspired concepts formed by Shovel Knight and Retro City Rampage, with additional research.
The music and core concept comes from a fan album, but all of the pixel art and sfx (including one of the fonts) was made from scratch for the project. The game is available for free on itch.io.
2022 Q1
Live2D Variations
Due to the “paper-doll”-like structural design of my Live2D model, I was able to create dozens of clothing variations with a very low amount of re-rigging in Live2D Cubism. To the left is a reel of all the clothing designs, ranging from hasty-and-scuffed for the early items, to colourfully painted and full of depth for my final pieces.
It’s more-or-less a condensed timeline of my improvement with hand-painting with a drawing tablet in Krita over time.
2021
Music Video
With Japanese idol culture becoming highly dominant in the Vtuber sphere, the first generation of Vtubers began to retire, reduce activity, or move to other careers during this year, due to increasing irrelevance. I wanted to make a music video with who was left or about to leave.
Participants include a Japanese tourism mascot, a manager, scientific researcher, social VR users, artists, technology commentators, and singers. I produced the video and music mix myself. This is also my first use of DaVinci Resolve for video editing instead of Blender, which was highly outdated.
2020 Q3
Transition to Live2D
Looking for a creative reboot, I retired my 3D models and made a Live2D model using the art skills I was rebuilding over the prior year. The art file itself was hand drawn in Krita and was designed from the ground-up for modularity and rigging, instead of being a conversion.
I then proceeded to document this methodology by creating a tutorial video, which has been praised for being highly accessible and beginner-friendly, yet highly comprehensive. The video has over 100,000 views.
I also created this website at roughly around this time, to retain a written history of early Vtuber software and a written version of the tutorial.
2020 Q1
Original 3D Vtuber Models
My 2019 activities culminated in making a more detailed original model that was suitable for full motion capture – though still using a VRoid base. This iteration, internally labelled as “version 11” would be used in a Let’s Play of Half-Life: Alyx, but not much further than that.
The final design is themed after Japanese idols. After a few more months, I realised that it wasn’t practical for me to keep up with the increasing hardware expectations of 3D Vtubers. At some point or other, I also realised that idol culture didn’t suit me.
2019
Twitch Streaming
First complete year of streaming on Twitch, though at this point it’s still periodic and not great. After trying both styles, I settled on using VRoid and 3D as my Vtuber base instead of Live2D, and continued to try developing a distinct brand and identity throughout the year.
I tried a lot of styles using assets from a new, growing Vtuber asset store on booth.pm, but I never made sellable assets of my own – the last time I did any texturing was in 2010, so the original assets I made weren’t very good. By the end of the year I made a fully original model, though it was quite scuffed.
Throughout my time on Twitch I also made many, many iterations of stream overlays and Lua scripts for OBS, entirely from scratch.
2018 Q4
Let’s Try! Reviews
After the death of Totalbiscuit, I began formulating a video game review methodology based on my game design background and Totalbiscuit’s “WTF is” and “15 minutes” series. It resulted in a short review series I call “Let’s Try!”, which I have since revisited annually.
The core theory behind my review approach is to identify the length of time it takes to reach the end of one game loop (not including the tutorial), and see whether it fits in the average length of a game session based on statistical data. In lieu of an identifiable loop, a hook or incentive to continue the game within the timeframe is also noted.
In the following years, I would try to reapproach this review format from different angles, though I still consider the first series to best represent the original concept.
2018 Q3
The First Virtual Youtuber outside Japan?
Japanese internet communities began developing homebrew Vtuber programs during the year. After studying the software, I become one of the first – if not the very first – non-Japanese person to identify as a Vtuber on Twitch, using dedicated Vtuber software. (Earlier avatar streamers using VR or Facerig retroactively labeled themselves as Vtubers in 2019, so the distinction is fuzzy.)
I ended up teaching myself how to use OBS Studio, Live2D Cubism, and the then-new VRoid Studio to create my first livestream. My first two dozen or so streams all ended with post-mortems and self-improvements.
2017
Vocaloid in English
Composed and tuned several Vocaloid songs in English. I decided to study and experiment with Vocaloid in response to a general community sentiment that “Vocaloid is only good in Japanese, and English will always sound bad”.
The resulting series of songs used an application of linguistic concepts associated with singing accents and phonetic drift, with each song increasing in technical and creative complexity. The final song in the series features fluently spoken English.
2016
Translation Placeholder (KR to JA to EN)
Created placeholder translations for an MMORPG project by machine translating from Korean to Japanese, then translating normally from Japanese to English. The use of machine translation was justified by citing several research papers indicating that contemporary KR-JA machine translators had an accuracy of 95-99%.
I also provided basic advice to the company on the legality of community translation, and left the project when a proper support structure was established.
BEFORE 2014
For a few years before 2014 I was using my real name as a screen name. Most of the data is lost or incompatible with Windows 10 and current web browser standards, and what remains won’t be linked on this page for privacy. My older software experience is still written here for the sake of completion, since I’m open to picking it up again.
Pre-2014 Software Experience
Famitracker (Music), Reason (Music), Unreal Engine 3 (level design), Unity 3 (dev), 3DS Max (3D modelling), GraphicsGale (pixel art), Audacity (Audio), Aegisub (Closed Captions), Blender (3D modelling and video)
Languages: C#, UnityScript, Python, OBScript (Oblivion, Fallout 3, New Vegas Engine)
Eiri Sanada 