Few games are as iconic as Capcom’s world-beating Street Fighter II. Miles Bowe follows indie label Brave Wave’s mission to release the definitive version of its unmistakable soundtrack.
Video game music is hard enough to get right, but fighting games are even tougher. A fighting game’s music needs to be memorable, but not too distracting. It needs to find a way to tell you about a character even when plot and dialogue are far from central. Most of all it needs to provide that shot of adrenaline to get the blood flowing before it’s actually spilled. Street Fighter II set the gold standard for music in a fighting game with each of Yoko Shimomura’s character themes. Its quality put it high on our list of the 100 greatest video game soundtracks, and it’s one of the few game soundtracks that has become permanently embedded in popular culture (after all, Guile’s theme does go with everything.)
This year we got word that game music label Brave Wave were cooking up what they called “the definitive Street Fighter II soundtrack”, and from the start it was clear how much respect, dedication and time the team was willing to spend making sure there was no hyperbole in that claim. The score was remastered directly from the raw sound pulled from the arcade game sound boards, with experts and Shimomura herself providing guidance. This edition even offered both of the distinctly different soundtracks split over earlier and updated versions of the game. Both arcade system boards, CPS-1 and CPS-2, were given equal attention to highlight their unique strengths.
Street Fighter II: The Definitive Soundtrack is now out, and currently the best-selling album on Bandcamp (every vinyl edition sold out almost immediately, but fear not, the label is pressing more.) To get a deeper look inside this labor of love, we reached out to Brave Wave’s PCB (printed circuit board) recording director Haruhisa ‘hally’ Tanaka, restoration engineer Marco Guardia and creative director Mohammed Taher, who provided us with in-depth answers, photographs of the soundboards themselves, and comparison tracks to hear the result of their hard work. Read our interview below and head to Brave Wave’s store to order.
What drew you to Street Fighter II? Why do you think this music is so timeless?
Mohammed Taher, creative director: Yoko Shimomura’s music does an excellent job at personalizing the fighters. ‘Ryu’s Theme’ has this wistful quality, him being the lone warrior and all; Blanka’s is appropriately heavy on percussion and has a sort of jungle feel to it; Balrog’s is a melancholic tune that fits the unpredictable nature of boxing. She essentially crafted short stories for each theme, breathing life into all of the characters. It was revolutionary at the time, and I think that struck a chord with gamers. And when it comes down to it, the songs themselves are simply memorable. Beautiful melodies through and through.
How did you decide on which actual arcade game soundboard you’d eventually record the audio from? How many did you look at, and how many did you go through?
Haruhisa “hally” Tanaka, PCB recording director: We didn’t need to be concerned with the condition of the PCB (including the soundboard) as long as it worked correctly. This is because we had to modify the sound output anyway. We saw just one Street Fighter II: Hyper Fighting PCB and two Super Street Fighter II: Turbo PCBs (the first one didn’t work at all) which we bought from used arcade PCB stores in Japan.
This past May, we started to modify Hyper Fighting. We had to convert the FM sound chip’s analog sound output to digital for the best sound quality. There were no critical problems because we had some experience from a similar scenario before. Unlike Hyper Fighting (CPS-1), Turbo’s PCB (CPS-2) brought us lots of problems. In July, after finishing Hyper Fighting’s recording, we had to start researching the sound output to see if the digital-to-analogue sound conversion was also available for Turbo’s PCB. Turbo’s sound system is based on wavetable synthesis, which is completely different from Hyper Fighting’s FM sound. The wavetable synthesis music is played through the QSound audio system, which has still not been fully reverse-engineered. After one month of research we decided that digital output was impossible, at least with today’s knowledge and technology. Then we found the best analog sound output instead within a few days.
The process for CPS-1 took around a month, and two months for CPS-2. Most of our time was consumed by researching, though. The recording itself was finished within a few days in both cases.
There’s such a distinct different between the CPS1 and CPS2 versions, a lot of people might not realize there are these two very unique versions. Did each one present different challenges in remastering?
Marco Guardia, restoration engineer: The CPS-1 used a combination of an FM synth chip and a primitive sampler (mostly used for percussion). The CPS-2 used wavetable synthesis, which is entirely sample based.
When you first started working with the material in the soundboards, what were some of the immediate qualities of the music you knew needed to be approached in the remastering process?
Guardia: The two boards are quite different, so both the sound and the way we captured it posed unique challenges. With CPS-1, even though it was an older board, we found a way to capture the audio digitally. This gave us two separate streams, one for the FM chip and one for the sample channel. Essentially I had two stems that needed to be mixed back together first.
For CPS-2, it was an altogether different challenge. First, we were working with an analog sound capture. Despite our best efforts and a lot of research it turned out to be impossible to get the sound off the board in any other way. CPS-2 also uses a patented QSound chip which adds a post-processing pseudo-surround effect to the audio. It wasn’t possible to bypass this. Some aspects of this surround effect were pleasing, others problematic, so this needed to be addressed. The general quality of the CPS-2 capture was simply not that great. The analog signal path gave us a fair amount of distortion and “corrupted” higher frequencies. It was clear that CPS-2 would be the bigger challenge.
What is it about the arcade board sound that makes the original music so raw? Switching between the two versions of ‘Guile’s Theme’ for example, the original sounds so limited and held back while the new one sounds like it’s been “unlocked” from those limitations, like how it was intended to be heard.
Guardia: For the music directly off the board it was the low sampling rate and bit depth, and also the analog circuitry itself. This was all pretty standard for the time — it was a matter of memory and economic concerns. High-end sampling chips were incredibly expensive or simply not available and there was just not a lot of room for high quality samples.
There are several methods to capture audio from the CPS-1 board, and as far as I know not a lot of people are aware of how to get the clearest and cleanest audio even when accessing the board digitally. Plus, much of what people are familiar with now in terms of the sound is the emulated version of CPS-2. If you go on YouTube and look for the music you’ll find that almost all of the videos contain emulated sound. This sound is taken from a ROM image being played back on a computer, either in an emulator or a dedicated music player that reads these ROMs. It sounds alright, but it’s far from 100% accurate.
What was your process for approaching all of the clicks and noise in the original? What do you think makes the biggest difference in the remasters?
Guardia: The entire remastering was a three-step process: sound capture, restoration, mastering. Each step was crucial to how the final result sounded. For CPS-1 I think the biggest difference is the added clarity, warmth, punch and thickness of the sound. For CPS-2 it’s the much improved stereo image, the cleaned up, more defined low-end, and, again, the added clarity and essentially distortion-free sound.
There was plenty of recorded silence from the board to get a clear reading for a noise profile. This profile’s application is then tweaked with some parameters. You listen to the removed part of the sound and dial in the noise removal only to a point, just before it “grabs” anything in your signal that it shouldn’t remove. I don’t remove the noise floor completely this way. Noise removal can get overly destructive. The rest of the noise is removed through simple noise gating: basically, as soon as the signal drops below a certain point, it is muted completely. This is useful for beginnings and endings of songs or brief silent periods where the faint remaining noise might still be audible.
For CPS-2 I had to work against the distortion and the lost high frequency content. It took a while to come up with a process that worked here. Even though the emulated CPS-2 soundtrack wasn’t 100% accurate, it gave us a good idea of how the higher frequencies of the samples were supposed to sound when played back digitally, so that was our reference point. Another aspect of the CPS-2 restoration was the QSound surround effect. While pleasing for the high mid and upper frequency spectrum, and certainly authentic to how the music sounded in the arcades, this effect had a very negative, destructive impact on the low end, robbing it of punch and giving it a very hollow sound. After some research and better understanding of how the surround effect worked, I was successful in reducing its intensity smoothly and gradually toward the lower end. The resulting sound is much more pleasing on modern stereo speakers and headphones.
What was your favorite part of the soundtrack to work on?
Guardia: For me personally, I liked working on the tracks that had the most “problems” out of the gate. I remember Fei Long’s CPS-2 theme having issues with very audible distortion and unfortunate side effects from the surround effect initially. It was great to hear that come together eventually. ‘Guile’s Theme’ (CPS-2) gained a beautiful, bigger stereo image and a punchier sound after restoration and mastering. And I’m also particularly fond of ‘Zangief’s Theme’ (CPS-1), which had a lot of distracting clicks all the way through and sounded very muffled initially.
What was it like working with Yoko Shimomura and being trusted with these recordings by the original composer?
Taher: Street Fighter II is arguably the most important game in the genre, so it’s a great privilege to have been given this opportunity from Capcom. Working with Yoko Shimomura was both an honor and a pleasure. The entire process was streamlined and uncomplicated for us. She was very open about restoring and modernizing the soundtrack during the mastering process and actively encouraged it. She approved our work at every step, and gave her input and advice on how she envisioned the final sound. Shimomura is a true pioneer of the 8 and 16-bit era, and having her approval of our new remaster means a lot to the integrity of our work.