VA-11 Hall-A Voice Over YouTube Series

It’s been an absolute blast and honor, and also a lot of sweat and tears (luckily no blood!), to have worked on this fanmade VA-11 Hall-A voiceover project, especially considering the very talented cast we ended up with!

But, what is VA-11 Hall-A you ask? VA-11 Hall-A: Cyberpunk Bartender Action, or more easily pronounced, Valhalla, is a visual novel (game) centered around a bar in a dystopian cyberpunk alternate future, where we get to meet and talk to people trying their best to survive a harsh landscape of corporate greed and big economic inflation. Gameplay consists of reading. A lot. And then mixing drinks to the orders of the customers that you get to interact with. While the only choices you have as a player is what you serve and what items you buy for the main character it may sound lacking, but it’s the story, characters and how it’s written that is the biggest attraction for this visual novel. Though, playing it, while the music is outstanding, you really do wish the characters could be heard speaking too. And that’s what we wanted to fix!

Watch on YouTube now!

Check out the playlist on our B.A.S.E. (Balancing Act Studio Entertainment) YouTube channel, and continue reading below if you’re interested in the making of this project and my involvement in it.

Note though, offensive language and themes are present, so it’s not for everyone!

If you’re interested in future projects of this group make sure to follow our Twitter account and subscribe to our YouTube channel! As of this writing we’re looking into making a voice over mod (and possibly videos) for another visual novel game called Coffee Talk. More on this in the near future!

The Beginning

It all started with a tweet from a figurine and cosplay photography friend, Balance (whom I met with back in 2018), that I happened to stumble upon in my timeline:

While the tweet mostly seemed geared towards finding interested voice actors, and I’m certainly no voice actor but rather a hobby audio engineer, I figured he could perhaps use some help with audio. And, a voice over project sounded like fun to try my hand at. At the time I read it I felt like doing more with audio (I had just done some remasters of some old music me and my brother had recorded), so this seemed like a perfectly timed opportunity. I asked Balance if I could join as an audio engineer and he invited me in a week later.

Introductory message in the B.A.S.E Discord.

Initially I was supposed to be the support engineer for the project while Ritzy was the main audio engineer. However, due to some circumstances I ended up doing all the audio engineering work and with Ritzy as a quality checker.

I helped early on with reviewing auditions and cast decisions, setting up audio recording guidelines and experimenting with the sound processing of the Lilim characters (androids essentially). Later once the project started rolling for real in March/April I also attended most recording sessions to assist Balance in audio checks, directing and double checking the script.

Setting a Standard

Audio processing wise it wasn’t anything too special, though my background in mixing music, and especially with low quality audio sources, taught me a lot of the workflows and tricks that I used for this project.

For the Lilim it was fun trying out different plugins to find something that worked well. We already knew about some other fandubs attempting to add voice processing to the characters in VA-11 Hall-A, but I was never really impressed with them as they not only had technical errors, but also really strange choices of processing. I knew I wanted to try do better.

While VA-11 Hall-A is a game running on a game engine there is a possibility to modify (modding, or mod for short) the game and insert the voice lines to the text lines as they appear, but this required skills that we didn’t have in the team, so it was decided from the start to make videos for a YouTube channel instead. This has the added benefit of total control of video, audio and timing, which I believe made the experience even better.

Even though this was a video project it is a very unusual video project where we essentially made an audio book first and then timed the video to the speech. Because the point with the project are the voices, having good natural flow and timing in the conversations were more important than making sure speech lined up with the video. I believe this is one of the mistakes some fandubs make, and for me it simply makes it awkward trying to watch those videos.

Melda Production’s free plugins came to great use for some voices, like for Dorothy. And yes, that is “Auto Tune” (real time pitch correction) that is used for both Dorothy and Deal. I’m somewhat surprised viewers caught onto that!

The Big Work

Once I had enough recordings of the actors/characters for each episode I went to work. The biggest time consuming task for me was to find the lines in the recordings, choose the best takes and time them to each other. Since actors were not always recording the episodes in the same session, some solo, and some in pick-up sessions, not to forget the multitude of retakes that could happen, there were just no way of easily trimming all recordings in sync at once, so going through every recording individually to find the best take was a time consuming necessity.

Initially I used Cubase 5 for all editing and mixing, but up till episode 4 it became quite tedious as the editing tools weren’t very fitting for quickly adjusting timing. Balance was using Davinci Resolve for video editing, and having used Resolve at work for some video editing I knew Fairlight in Resolve (the audio mixing part of the program) was quite versatile, and after looking into it more closely I decided to switch to Fairlight. This helped speed up my editing process, as well as my mixing process making my mixes sound even better, somehow.

One mistake I made early on was to use one and the same project in Cubase for all episodes, thinking it would speed up mixing by not having to redo any mixing, but I quickly learned that no recording environment or setup ever sound the same between recording sessions, so adjusting the mixes per episode is a necessity anyways, so for the Fairlight workflow I switched to per episode projects instead.

My view when editing dialog. On the left the script, on the right the Davinci Resolve Fairlight tab.

When I made the mix changed during the project too. At first I did mixing after finishing dialogue editing, but I’m someone who can much better spot issues if I come back to something I’m creating at a later time, so creating a mix before I started editing really helped me fine tune the mix during dialogue editing instead of waiting with it. Plus, it was more pleasant to listen to while editing.

(Warning, next paragraph is technical stuff.)

This project finally got me to invest in newer and better audio equipment, and what a difference it made. Ever since I moved to a different country I never invested in a new ASIO capable sound card/hardware (ASIO enables software to talk directly to audio hardware). I had managed to get away with it for a while, but when working more seriously with audio in this project the windows drivers just didn’t cut it anymore, so I invested in a new audio interface. That sped up my workflow quite a bit. But the biggest difference was definitely the new headphones I got. I had been using the same old studio monitor headphones I bought probably 15 years ago and somehow I hadn’t noticed they were lacking in treble and bass a bit too much. I love a “flat” frequency response, but these old headphones were too “dull” now. (And probably always have been.) I finally noticed their poor quality, cause it was actually after I came back from visiting my parents in my home country where I noticed a newer version of the headphone I was using at their house for my electronic drum set just sounded… better. So half way through the project I bought and started using newer headphones. If you noticed an increase in mixing quality watching through the series, this is part of the reason.

Handing it Over for Video

Once I were done editing, adding sound effects, mixing and checking panning (characters voices to the left or right in and off the screen) I sent it off to Balance who had the task of recording the gameplay footage to the audio. Like mentioned earlier, it is an unusual project in that regard, but he found a way to minimize amount of editing he had to do to time the video to the voices.

During the first test round of putting together a video to the audio I noticed right away that the connection between the spoken words and the character on screen weren’t there. The character’s lips only move when the text is being written in the text box, which means the audio simply didn’t match the length of the character’s lip movements 90% of the time. This is when I suggested with video editing we resync all the lip movements so they only happen during audible speech. Yes, this sound like a lot of work, and it was, Balance can attest to that, but the end results were so much better that it was worth doing it for every character that had moving lips. This eventually did also create even MORE work as some characters were changing places or even having text flying in front of them (in the cases of Streaming-Chan showing up, in these scenes the flyby comments were replaced with our own made up ones, all just because of the lip resyncing needed to be done). I don’t envy Balance having to go through all of this, but like I said, the end results were made much better thanks to it. (Thanks for your hard work, Balance!)

As for the rest on the video editing side of this project, as well as pre-production and other stuff that Balance was handling, you can read Balance’s own post about it over at his blog here! Also, I have to commend Balance for always having great choice of music in his video edits. He even made his own challenge of trying to include all songs from the game’s discography at least once!

The Mishaps

It really goes without saying that there’s always going to be some slip ups or mistakes when it comes to big projects like these, especially for us as amateurs. In the first episode the wrong audio mix got used, so some dialog might be harder to hear. And in the first through fourth episodes the mix also ended up mono after some video editing mishaps, resulting in left and right pannings of the characters getting lost (all just centered). Luckily there are no other big issues with the audio after episode 4. So even if you didn’t notice it while watching, now you know (why it might sound off)!

The Result

The end result is 19 episodes of fully voice acted, audio edited and video edited intriguing cyberpunk bartending action, all 11.5 hours in total. It is currently as of this writing also the only fandub of this game that includes a full playthrough of the game. This video series is totally free, but if you end up enjoying our series I highly recommend buying the game and playing it yourself as there are a lot of story lines and endings that weren’t shown in this playthrough!

The Thanks

Thanks to everyone involved in this project making it so much better than I could have anticipated. All the voice actors, the majority of them being amateurs, blew me away with their amazing performances. Balance for not only letting me join the project, but also manage the project of so many people, directing and getting the best performance out of everyone and also his sleepless nights editing all the video on a very tight schedule.

Head over to YouTube to watch the whole series here!

1 Comment on “VA-11 Hall-A Voice Over YouTube Series

  1. Congrats on the release! I have a couple random thoughts… I have an old-ish pair of headphones (DT 770 Pro) that sounded like it was going bad; turned out that replacing the padding fixed it because the old ones were getting flat, and distance of the drivers to the ears had a big effect on freq. response.
    I also have a ridiculously old RME Fireface 400 audio interface; so old that it’s firewire-based, and THEY’RE STILL SUPPORTING IT with driver updates. I’m an RME customer for life now.

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.