At CEDEC2024 held at Pacifico Yokohama North
The session featuring AZSTOKE has been published!
Materials during the performance You can check it out at CEDiL .
The materials uploaded to CEDiL include links within the explanatory materials, so you can check the details of the content.
AZSTOKE Co., Ltd. gave a presentation about RIGDOCKS, a technology that was developed over a period of five years.
It is a direct embodiment of what you want in game sound development, and by making it customizable,
It has become a technology that can contribute to a wide variety of titles.
Below is a selection of what was presented at the event.
I hope you will take a look.
table of contents
◆ Five issues
Youtube URL: 05:40
When working on game sounds, one of the biggest challenges is the number of sounds, and we wanted to find a way to solve this problem by automating it, which led us to develop RIGDOCKS, thereby contributing to reducing labor costs.
Pain points are areas where many developers currently have problems and are looking for solutions for a lot of their work.
◆RIGDOCKS plan "BRONZE/SILVER/GOLD"
Youtube URL: 13:28
One of the issues with RIGDOCKS was the documentation, which has been thoroughly improved in Japanese, so that functions can be resolved simply by looking at the documentation.
▽APIDOCK is a document that allows you to check the functions of individual APIs. https://www.azstoke.jp/rigdocks/apidock
▽REAPDOCK is a document that explains how to create practical scripts using APIs.
We have provided extensive documentation to enable even the most experienced sound creators to create their own sounds.
We provide explanations of the genres of the approximately 300+ APIs currently available for each plan: BRONZE/SILVER/GOLD.
◆ Patented Technology 001 "HANDAUTOMER"
Youtube URL: 17:32
Patent Technology 001, "HANDAUTOMER," explains a technology that can automate voice volume adjustment.
When recording voices by voice actors, the volume changes greatly because the way the voice is raised changes depending on the emotion of the lines. To adjust the volume, it is not possible to use a compressor or other processing method as it will degrade the sound.
Currently, the curve is adjusted manually to avoid compromising sound quality.
However, there are a considerable number of voices for game sounds.
Sound creators are handling tens of thousands, hundreds of thousands, and in some cases millions of items one by one, and they need dedicated personnel to make large-scale volume adjustments.
We developed a dedicated volume adjustment algorithm and applied it to each waveform data to automatically generate a curve. This allows us to detect differences in each waveform data and generate a dedicated curve even for waveform data with completely different volumes.
[HADAUTOMER detailed explanation page] https://www.azstoke.jp/handautomer
◆ Patented Technology 002 "TALKTRON"
Youtube URL: 20:13
Patent Technology 002, "TALKTRON," describes a technology that can automate the task of adjusting voices in multiple languages.
Since multilingual voices are produced based on the main language voice, it is generally best to adjust them to a similar volume.
Multilingual support includes Japanese, English, French, Italian, German, and Spanish via JEFIGS, with many more languages being added in the future.
For example, if English is produced as the main voice and has 20,000 voices, the voice volumes for all other languages will increase multiply, and if the number of languages mentioned above is included, the total will swell to 100,000 voices.
The technology that automates this process is "TALKTRON."
It not only automatically generates tracks and arranges waveform data, but also arranges voice data with the same name in the main language, and then automatically arranges the multilingual voice data you want to adjust alternately.
"HANDAUTOMER" is run on the specified language recorded during placement, and after adjusting the volume, the MaxMomentary of the main voice is obtained and applied, all in a fully automatic manner.
This has evolved into professional-grade automated technology that allows you to focus fully on creativity by allowing humans to check and fine-tune after all adjustments have been made.
[TALKTRON detailed explanation page]
◆ Patented technology 003 "LOUDMATCHIFY"
Youtube URL: 23:20
Patent technology 003, "LOUDMATCHIFY," describes a technology that detects specified strings in file names through partial matches and automatically changes them to a dedicated loudness value.
When adjusting the volume, adjustments are made for each character.
Each action has a different volume. For example, the volume of a special move shout and the breathing of someone walking will be very different.
In such a case, if you set different volume values for each character,
Different people have different volume levels, which can be uncomfortable for users.
In the end, the sound creator spent a lot of time fine-tuning it by ear to ensure the quality.
"LOUDMATCHIFY" has developed technology to adjust the volume level fully automatically.
To match file names, we create a list of the matching strings and their corresponding loudness values. By sharing this list within the team, the volume values will not change even if the person working on the project changes.
This helps automate various tasks that previously required individual volume adjustments.
[LOUDMATCHIFY detailed explanation page]
◆ Patented technology 004 "MIDDLESCANNER"
Youtube URL: 25:18
Patented technology 004, "MIDDLESCANNER," is capable of obtaining the volume change value of the middleware and measuring the amount of change.
If you adjust the volume on a DAW and then insert the data into middleware, the volume adjustments will be applied in various places, which can result in the data adjusted on the DAW coming out at a completely different volume value.
The problem is deep-rooted, and it becomes increasingly difficult to adjust in the DAW.
In the end, it takes an enormous amount of time to output the volume on the middleware, visually check the volume value, and then readjust it.
To solve this problem, if you can obtain the volume change value on the DAW, it is possible to automate the final volume adjustment for that specified volume change amount.
It is currently applied to Audiokinetic's "Wwise".
[MIDDLESCANNER detailed explanation page] https://www.azstoke.jp/middlescanner
◆ "Demonstration" I want to edit and export a large amount of audio
Youtube URL: 30:22
The content will include demonstrations of the functions required for actual production and how to create them.
You can see an explanation of each API used and the results of executing the script.
Because it is highly customizable, it will be implemented by assembling parts.
[Link]REAPDOCK
We will explain the production flow for actually creating a script using RIGDOCKS.
*The session was created using Lua .
[Link] APIDOCK
▽Lua
[Similar] REAPDOCK
▽Lua
▽Python
◆ I want to replace materials in the "demonstration" cutscenes.
Youtube URL: 36:17
This section demonstrates how RIGDOCKS can be used to line up in-game footstep sounds and play them in sequence for videos such as cutscenes, and then automatically place a specified number of similar footsteps all at once.
In line with this production, not only footsteps but also action sounds, weapon changes, character changes, etc.
It is effective in a variety of places.
[Link]REAPDOCK
We will explain the production flow for actually creating a script using RIGDOCKS.
*The session was created using Python .
[Link] APIDOCK
▽Python
AZ_GetPathInfo
[Similar] REAPDOCK
▽Lua
▽Python
◆BIG SALE
Youtube URL: 42:56
We distributed flyers and gachapon discount tickets to those who visited our booth.
The distribution period has ended.
The discount ticket is about to expire.
The deadline is September 30th , so please consider purchasing now.
◆30-day free trial version
Youtube URL: 44:24
All plans offer a 30-day free trial.
All features are available for 30 days only, so please consider using this. 👇 For details on how to set up the trial version, please see the following URL.
◆SKYSQUARE Launched
Youtube URL: 46:12
SKYSQUARE has finally launched.
We have purchased a mountain in Yamanashi and are currently making preparations to be able to record a variety of sounds.
SKYSQUARE is working to achieve two different uses.
1. Environmental sounds recorded 24 hours a day, 365 days a year
2. Large-scale Foley recording using heavy machinery
We are currently in discussions with various companies to achieve these two goals.
We are making progress, little by little.
We hope to keep you updated on our progress via our blog, social media, etc.
Please look forward to future developments.
X (old: Twitter) https://x.com/A_Z_Stoke
Facebook https://www.facebook.com/azstoke/
Comments