Sound direction: Eriko Azuma
― First of all, please tell us what led to you being involved in the game.
Azuma: I joined SCEJ about 4 or 5 years ago. I first got into this line of work when I was assistant on Boku no Natsu Yasumi. After that I worked on things like Toro to Kyujitsu and PoPoLoCrois, and was in charge of sound direction on Yoake no Mariko... Then when I was working on Mariko I met the current Siren team, and ended up working on this game as well.
― Looking at your past work, it looks like there's a huge amount of comedic ones...
Azuma: Yeah. Lots of them are cutesy ones. This is the first time I've done something realistic, so it feels a bit like something that was latent inside me that I got out (laughs).
― Do you mean that you're quite a fan of horror, then?
Azuma: It's not so much that, rather than the fact that I have an academic background of composing since I studied music at university, so I usually focus on ethnic music. I've always liked it, too, and I went to India for about two months on a working holiday before joining SCEJ. Maybe those experiences caused it.
― So you've had training (laughs). Lots of different composers participated in Siren, but you were in charge of sound direction - precisely what did that entail? It looks to have been quite a long-term project.
Azuma: For this game we wanted to make it ethnic-sounding and use a lot of background noise, so I began by searching for composers who could get that balance right. I ended up receiving demo tapes from a few people, and out of those I chose the ones who seemed to have unique ideas about sound-making, asking Hitomi Shimizu and Gary Ashiya to do the main parts. Gary is quite familiar with horror, having worked on things like Juon, so he knows the best way to get the scariness out of things. Ms. Shimizu has done quite a bit of research into the microtonal organ, and had really unique and intuitive ideas.
― The microtonal organ?
Azuma: Yeah, one that uses microtones. For example, you can break down the sounds between C and D into even more specific sounds, and sounds of unequal temperament sometimes feel unpleasant (to humans). She'd made her own organ to try to produce that, which I found really interesting. The religious-feeling parts were done by Ms. Shimizu, and Gary did most of the field noises that make it sound like a horror. For the rock songs, one (from the genocide ending) was done by Noriyuki Asakura, a veteran, and around two songs to increase the tension of the end stages were done by Kimitaka Matsumae.
Khöömii (Khoomei): A traditional style of throat singing from Mongolia that began in the 10th century. Low, hoarse sounds are produced by the vocal cords, the trachea and throat harmonising so that a single person can simultaneously sing in two to three timbres. This creates a unique kind of harmony. Kargyraa is a style of singing used by the Tuvans.
The Buster!: The rockish BGM that plays during the Genocide Ending, used in a display of youthful resolution by Kyoya to rouse himself for the final battle by playing it loudly through his headphones. "I wanted some explosive music that would be befitting of a massacre, so of course I discussed it with Toyama. At first the idea was to have t.A.T.u." (Azuma)
― So you recorded all of the sounds as you were gathering the composers together.
― I think the general theme of the game had been decided before work began, but what kind of experience was it to work on?
Azuma: Like I said earlier, the themes for the tracks were divided into a few different kinds and shown to Toyama's group from the start. Overall, we wanted to express the oppressive, isolated feeling of Hanuda, so we started by filling it with low sounds, then adding sounds with an effect on-screen at the same time over the drone, for a sort of minimalistic approach... I went looking for data to solidify the image, combining things I thought sounded good, and went through a period of trial and error to complete it to what it is now.
― One of the things that stands out in Siren, a game based on environmental BGM, is the main theme, Hoshin Goeika. It sounds magical, like an incantation, but also solemn.
Azuma: For that hymn, the first thing we decided was that we wanted it to be primarily created around vocals. The first thing Toyama told me was that he wanted something like Bulgarian Voices, but I thought it sounded kind of cliche (laughs). When I was deciding what to try next, I searched around listening to methods of vocalisation from all kinds of countries, and what could be done with them. What we settled upon in the end was Mongolian Khöömii, but for the game we used a hoarse voice that was an octave lower, a style called Kargyraa.
― A voice that sounds like a deep growling?
Azuma: Yeah. Then I combined that with Hisako's song of prayer in Bulgarian style that Toyama mentioned... Then came the issue of whether there was someone (in Japan) who could sing it (laughs). Ms. Shimizu has a friend called Koichi Makigami who studies Khöömii, who introduced us to the members of Baions, the only group of performers in the country who could do it, and finally finished the song. The hymn was the one that took the longest time both to find someone to create it and to decide what to actually create, and everyone feels like they put their souls into it. On the other hand, though, it took so long that it made the rest trickier, causing a string of late nights...
― Before you mentioned looking into choral styles from all different places. What was there besides Khöömii?
Azuma: Starting with Japanese Buddhist hymns and folk songs, I also listened to Chinese songs and Gregorian hymns - basically, anything that would be called orthodox. I also looked into Arabian-style Khaliji, but since the game's theme was "Made in Japan" I wanted to keep it as close to Japan as possible. Eventually I was left with two choices - an orthodox choir style or an ethnic style - and it became what it is.
― When the hymn stops and the siren sounds solemnly, it sounds like it's full of pure madness and it's scary (laughs).
― Aside from songs, oppressive sound effects are also relatively important in Siren.
Azuma: What I felt when I was gathering materials was that the thing that was the best fit for an abandoned village was not normal sounds, but something that sounded old and had been recorded long ago. When I listened to things recorded on a phonograph at the library or somewhere, it was really crackly and so scary. Abandoned villages, too... Even when there's nothing there, they're still spooky, right? Like that, I wanted to put in something where the sound sounded degraded in quality. I thought that alone would be too much for the players, though, so I also put in high-pitched sounds, like in the Annunciation scene (Day 2 6:00, Hisako Yao's level) that sound like they're coming from high up in the sky. There's not much in the mid-range; the story itself is full of extremes.
― It's interesting the way the tunes change slowly along with the progression of the story.
Azuma: That bit was down to Gary. I told him of my idea to gradually increase the background noise from the present day into the Other World, and he gave me something like that straight away. Also, this was something I spoke with Gary about a lot, but in the initial stages of development we took out anything that had a steady beat to it, aiming for just environmental sounds and ambience. I did think, though, that if it ended up like Kintarō-ame, so wherever you cut it it would look just the same, it would get boring. You can't know what players will do and when they will do it usually, so I remember we spent a lot of time fine-tuning the balance of tracks, like how to make things more tense.
Many of the sound effects used in Siren are sounds that were recorded during on-site research.
"I don't think it's impossible to create the dustiness of old houses and things like that, but recording is a lot faster. And, of course, we recorded the sound of things falling, breaking etc. and applied effects to them. During research we had Toyama and the young staff members purposely fall over and recorded those sounds, but I would say, "No... I didn't like that much, fall again," over and over (laughs). At first I also thought we could use the environmental sounds, but midway through recording a dog in the area began barking (with perfect timing), so we couldn't use it." (Azuma)
One of the tales of troubles during development is that they appear to have had hallucinations.
"At the start, I assumed that we would record the lines normally and add effects in real-time, but due to memory limitations we couldn't do it. In the end we created separate files for every single line, both normal and the ones that are unintelligible during sightjacking, then steadily made all kinds of changes using effects. Because everything was done with such attention to detail, it ended up being a lot of heavy work." (Azuma)
Refer to archive No.085: Eri Azuma's Record. Azuma is the one playing Eri Azuma.
― Aside from that careful fine-tuning, is there anything else that you remember?
Azuma: It has to be the siren, doesn't it? At first I tossed out a few dozen proposals for siren sounds, and Toyama chose from those. We began with wanting to feel a sense of "water" from Datatsushi, so the first idea for the siren sounded more like a whale. But when we finally actually tried it out, it was buried by the environmental sounds and didn't really represent anything... From there we decided, in order to show that the sound of the siren is Datatsushi's scream, to leave it and put it together with the original (warning sound) siren.
― Is there something like the voice of an animal mixed in with the sound you ended up using? For example, you mentioned a whale just now.
Azuma: There's quite a mix in there. Whales in nature have all different kinds of voices, too. I was imagining an old air raid siren when I was making it, but if that was highlighted it would end up being nothing more than an air raid siren, so at that point I mixed in a bunch of different kinds of sounds.
― I'll have a good listen to it. How about the Shibito the players see and hear sometimes? It looks like a lot of effort was put into making them seem Shibito-ish.
Azuma: I spoke about this at the start with Toyama and Sato - the Shibito don't see or hear very well. Since you see things through their eyes, I tried to use effects to show this properly. I made it hard to hear, cut off high and low frequencies, things like that. Also, cheap-sounding effects from old films are weirdly scary, right? Toyama told me he wanted something like that, so I consciously tried to make the scuffling sound you hear when spider Shibito get close sound vaguely unpleasant.
― They sound like they could be an animal out of a disaster movie.
Azuma: Yeah, they do.
― What about the words the Shibito say?
Azuma: Sato told me that the Shibito do communicate in some way, but that she wanted it to sound like something unintelligible to humans. At the time of recording, we purposely recorded a lot of nonsensical things that sounded somehow divine. Like with the siren sound, the sound itself has had noises applied to it and mixed further, the volume balanced and then flipped around and distorted... Lots of things were done to it, like specifically warping and messing around with the sounds.
― The other day, Toyama and Sato told me about an actor who was like a Shibito, and did the hoarse-sounding feel well - it sounds like you had a good basis as well.
Azuma: Of course. With people like voice actors who have lovely voices, whatever effects you apply they still sound lovely (laughs). So, though it sounds kind of rude, in the end we called on the staff and recorded one with a rough-sounding voice, applied effects and altered the pitch, which sounded really great, and so we swapped that in (laughs). Things like that did happen.
― I see. This question is from a player's point of view: I think there will be lots of people who pay attention as they play the game who will already know your face from one of the in-game archives. I know you're actually on the record jacket as "Eri Azuma", but how did...
Azuma: It just happened (firmly).
― O-oh... I thought you created the song, and then wanted to actually appear on the jacket yourself.
Azuma: Not at all (laughs). First of all she was supposed to be an old idol so we made it sound like a pop song, but then Sato just said, "Hey, why don't we just have it (the photo of the idol) be you?" That's how I ended up on the jacket. Then she bought some stuff at a second-hand clothes shop and assembled a costume bit by bit, and then told me that they'd decided on a day for the shoot.
― So you were dragged into it (laughs). Finally, is there anything you want to say now you've finished work?
Azuma: This is a personal story, but when development was finishing up I was quite heavily pregnant at 8 months... The there was still a lot I had to do at the end, and at last Fujisawa ended up quickly settling all of the things I'd been worrying about, helping me out with so much. I don't think it would ever have been finished without Fujisawa.
― Are you sure it's okay to be letting your child hear such scary noises at 8 months pregnant, though?
Azuma: The game's sounds served as a kind of prenatal training... So sometimes I try singing "uuyaamaaaiiii" to my child to see if there's any reaction, but there's nothing really... Kids prefer the sound toys make, don't they? I think it's fine, but I do pray that it doesn't manifest itself somehow in the future (laughs).