[PR Project] Listening to Anime Songs in the Rumored Music Format "MQA"!　Part 4: Chiyomaru Shikura, the representative of MAGES.(5pb.), takes a sharp look at "MQA"!

Navigator: Kenji Nomura

Kenji Nomura is an AV writer active in a wide range of genres, including headphones. He has a deep knowledge of high-resolution sound sources, and is a supervisor for an ani-song label. He also appears regularly as an advisor on the "KAIUN Hi-Res" corner of TBS TV's Kaiun Ongakudo.

Guest: Chiyomaru Shikura

A multi-creator with many facets, including lyricist, composer, novelist, programmer, guitarist, and music producer, Chiyomaru Shikura launched the "5pb." label in 2005. This led to the "MAGES." label, which has released numerous anime, game, and novel titles. He is currently a director of Dwango and chairman of MAGES.

Mr. Shikura's 25 years of fighting against the "waveform" of sound

Nomura The label "5pb." established by Mr. Shikura has released high-resolution versions of almost all anime tie-in titles, and you have already been very active in the high-resolution field for the past three years. The other day, Yukari Tamura's new album "Princess ♡ Limited" was released in high-resolution (flac/MQA) and attracted a lot of attention.

Shikura: When I first heard about high-resolution distribution from the staff, I of course understood what high-resolution was and thought, "High-resolution?　Distribution?　Let's do it," and we started immediately. I was personally interested in high-resolution distribution, and at first I started working on it as an "experiment. I was also genuinely interested in how the highly conscious fans who dared to listen to high-resolution music would react when they heard a high-quality sound source being distributed, and whether the evolution of high-resolution sound could be properly experienced within the audible range of human hearing.

Shikura When was MQA first introduced?

Nomura The technology itself has been around for some time, but it wasn't until the year before last (2016) that e-onkyo music began distributing music in the "MQA" format. MQA differs from previous compression methods such as MP3 and AAC, in that it allows for highly efficient compression while preserving the quality of the sound, no matter how high the bit rate of the music. I think it is similar to the "Dolby Digital" and "DTS" compression methods used for video.

Shikura: That sounds wonderful. However, when "MP3" first appeared on the market, it was a huge shock. We had been sampling and programming music for game creation even before that, and since we started out by creating tools for music production, we had been "battling with waveforms" for a long time. That was when I was making games for the Super Nintendo, so it was about 25 years ago. From that time on, I was looking at sampling waveforms and studying things like points to get loops, tricks to avoid noise, and so on. In those days, even game music had to be run through a workstation once, and then the data had to be burned onto a chip before you could see what the final sound was like. You didn't want to go through such a tedious process, did you?　So I created a tool that would tell you what the sound would probably be like by just simulating it on a computer, once you set some parameters first.

So, I have been dealing with music waveforms for about 25 years now, and in the midst of such a long "battle with waveforms," when "MP3" was introduced to the world, I was at a loss to understand what it meant!　What the heck is MP3? It's not possible that standard compression would reduce the size of a file by a factor of 10! (laughs). So I did a lot of research on music compression at the time. I was thinking, "I see, they cut out the parts outside the audible range. That meant that a lot of the original data was lost, and I didn't want to do that.

Later, when "PlayStation" came out, it was also a very special music format, and as I worked on it, I began to get the hang of it. (Laughs.) It was already in the realm of the quantum realm, and we were doing calculations like that with a calculator.

Nomura: It was a world of craftsmanship (laughs).

Shikura When you import music data into a computer, you see the waveforms visualized, right?　That is a lie, strictly speaking. Digital waveforms are ultimately nothing more than on/off values of 0 or 1, so they should be rectangles. That is how digital data is supposed to look like, even if it is PCM, but when we pursued the connection of sound, we eventually came to the conclusion that some measures had to be taken at the level of this digital data in order to achieve a good sound. So, I have spent the past 25 years struggling with the digitization of music. If I were to talk about all of this, I would have to talk about the evolution of game consoles, the evolution of computers, and the evolution of "Pro Tools" (* music production software created by Avid Technology of the US) as a representative tool for handling waveforms, and I would not be complete without talking about all of this (laughs).

Nomura In the field of music production, I hear that the evolution of "Pro Tools" has changed the sound of the latest version of the software just by track down.

Shikura The evolution of "Pro Tools" itself is one thing, but the evolution of plug-ins is probably more significant. Some plug-ins are quite heavy, such as spatial plug-ins like reverbs and delays, and dynamics plug-ins like compressors and limiters, which are several times heavier on the CPU than the plug-ins included from the beginning. For example, if you were to start up a high-resolution 96 kHz session for an anime song with over 100 tracks, it would be very difficult even with a modern PC and an expansion card. Therefore, I have no choice but to produce a 48 kHz session. We have always wanted to produce sound sources in the "true sense of high-resolution," but we are still struggling to take one more step.

However, with the evolution of technology this year or next year, we may be able to solve this problem. As you can imagine, we too are already resistant to production in 48 kHz sessions, so we really want to start from the beginning with 96 kHz. But there are practically no studio environments that can produce true high-resolution sound sources. Anime songs, in particular, are the only genre with that many notes and tracks. The number of tracks is double that of ordinary pop songs, so it is a flood of sound. If we were to do it all at 96 kHz, I would be a little hesitant. However, I would like to challenge the limits, and as for how to deal with the evolution of hardware and software, it's more of a ...... hobby (laughs).

Chiyomaru Shikura talks about the doubts he had about "MQA" to a great extent

Shikura: I have heard that anime songs have been the driving force behind high-resolution audio. I think it is really difficult to distinguish between high-resolution and non-hi-resolution sound sources, especially with ani-song as a source. Well, I still prefer to listen to 96kHz. If a 96 kHz high-resolution source can be compressed, like MQA, then there is no choice left.

Nomura Yes, that's what it means.

Shikura When I first heard about MQA on the Internet about a year ago, I was half skeptical and did a lot of research on the Internet. Then I read something about "folding theory," and I thought, "What? I thought, "What? I thought, "You can't fold it. I still don't understand what it means.

Nomura To be honest, that explanation is a bit confusing (laughs). I interpret this as meaning that the sound information in the 96kHz/24bit audible bandwidth was "folded" into the sound. Therefore, for someone like Mr. Shikura, who has been looking at waveforms in detail for a long time, it is natural that he would feel "Then, there must be some missing information, right? It is natural that you would feel that there is some missing information. However, the original data is only 24-bit, not 16-bit information. Therefore, the amount of information in the sound is clearly different from that of 16-bit sound in high-resolution.

Shikura: So even if it is 24-bit sound, it is not lossless, is it?

Nomura It is not lossless.

Shikura So you are saying that because the original sound is good, the sound is good even if it is lossy (lossless compression), is that correct?

Nomura Yes, it can be played back as a 44.1kHz/16bit CD sound source, but when played on a compatible player, the sound is almost the same as 96kHz/24bit. It is a format with such versatility.

Shikura: I see. But that's a great thing, isn't it?

Nomura Another interesting thing about MQA is that it cuts out unnecessary noisy sounds on the time axis. By doing this, the sound is organized, and for example, the alignment of the rising sound is improved, so the focus itself is improved. Therefore, the impression of MQA may be that of a sound with a light filter or effector applied. The resolution is slightly lower than that of ordinary Hi-Res, but vocals and other sounds are rather clear, so if you are listening to it on the go on a smartphone, for example, it may be easier to listen to than ordinary Hi-Res. At the same time, the file size can be compressed to a small size, which I think is a big advantage for the listener.

High-resolution music, including MQA, is suitable for "singing songs" rather than anime songs

Nomura: At first, the world of high-resolution audio, including MQA, seemed like it would be for audiophiles only, but now that the ani-song genre has taken off, it is finally becoming more common. In the past, the spread of VHS video was driven by adult videos, but high-resolution music is now showing signs of spreading with ani-songs.

Shikura: Then, I guess it is adult. High-resolution 18-rated drama CDs, for example (laughs).

Nomura I think it would definitely sell, especially for girls (laughs). They usually use their imagination to compensate for various things, but with high-resolution music, they don't have to compensate for it (laughs).

Shikura No, I honestly think that genre is the most suitable for high-resolution. There are not too many tracks, and if you want, you can even produce a session at 192 kHz. The high resolution gives you a realistic feeling of being really there, depending on the listening environment, so I think it would be quite something.

Nomura It seems that recording in 24-bit gives a sense of depth of space, which makes a big difference when the recording is mastered.

Shikura Well, that may be true, but the time axis is really important in sound creation, and when you sample a sound and look at the waveform, you may find that the waveform is something you cannot imagine from the actual sound. For example, when making a professional wrestling game, the sound of a wrestler hitting the ring with a "bang" is very different from the sound of a wrestler hitting the ring with a "bang," and there is some kind of noise before and after the "bang. And if you remove that noise sound in the editing process, naturally, the grain of the sound will be much better. So, some of these editing operations that I have learned from experience can be substituted to some extent for the original presets that I have created in the compressor. I make sure to run that filter through it, and when I run it through, the sound grain sounds better this way, no matter who hears it. So we have been fighting with the sound source itself, not just the sampling rate or bit rate, and I feel that MQA also contains such a filtering element.

Nomura: I don't think that higher bit rate is good enough, and I don't reject 16-bit CD sound sources at all. What I want to deny is compressed sound sources such as MP3. I have my doubts about whether the data that the creators have painstakingly created can be so easily lost.

Shikura: Well, it's not good at all. The biggest problem is that, for example, with music CDs, when we listen to them in a wonderful studio environment, we think they are really great and release them to the world, but when we actually listen to the pressed CDs, we find that they are so different that we say, "What the heck! When we listen to the actual pressed CD, we sometimes find that the sound is so different from what we heard in the studio. Compared to the studio, the listening environment is totally different, the rate is different, the data is already smaller, and if it is pressed into something like 44.1 kHz/16 bit (the bit rate of a music CD), it becomes something totally different. If I could do it, I would say, "Come to the studio and listen to it! I would like to say, "Come to the studio and listen to it! That dilemma has always been there.

Nomura In that respect, high-resolution sound sources can be listened to with satisfaction by the listener, although it is necessary to have the right environment to a certain extent. Of course, it may not be as good as listening in a studio, but it can reproduce a sound very close to that of a studio, and MQA is almost the same way.

Shikura: I have been doing this job for many years, and I am particular about tone selection and other aspects of sound quality, even when I am mixing music. If I could, I would like to produce everything from the beginning in a 96 kHz session, but the sampled sound sources that are now considered the standard probably don't have bit rates higher than 48 kHz. The reason for this is that if you install only such heavy sound sources, the load time when calling up tones will be too long and difficult, and the tool itself will not work in the first place. Of course, there are some string instruments and piano instruments that have very high sampling rates. But that's only a small percentage, and in most cases, the main sound source is 48 kHz, or sometimes lower, and is brought into the studio. The amount of data is certainly there for 96 kHz, but that's just for show. I wouldn't go so far as to call it "niserezo," but in essence, unless the entire process from start to finish is 96kHz, it can't be called high-resolution in the strict sense of the word. That is a dilemma, and I have been thinking about it for a long time.

However, the "songs" are overwhelmingly different. The session itself is made at 96 kHz, and the orchestra is at 48 kHz. But only the songs are live, so recording them at 96 kHz makes all the difference. For instrumental music without singing, if it is made by inputting data, there is almost no difference between 96kHz and 48kHz. However, of course we record live guitars, basses, and in some cases violins and other stringed instruments, and to be more precise, we have "songs," which are truly emotional and changeable instruments. Since our main focus is on singing, we are trying to use 96 kHz just for that purpose.

Nomura I appreciate that, and I think the listeners expect that too. After all, many people buy music because they want to hear the songs. In the case of pop music, the three sections of singing, drums, and bass are the most likely to benefit from high-resolution, so the extension of the high frequency range of vocals is often mentioned as a benefit of high-resolution. I think this is why vocal voices are often said to have a high resolution. I think that is why vocal voices sound completely different at 96kHz and 48kHz.

Shikura: That's right. For example, the sound of a violin is the noise of the bow rubbing the strings, so there are noises that you can hear, and there are noises that are actually coming out, but you cannot hear them, but you can hear that "sense of noise" in high-resolution. It is because of these live instruments that high-resolution has its merits, so I think it is a misnomer to say that ani-songs with a high ratio of synthesized sounds are suitable for high-resolution. It is not suitable because of the large number of notes. However, the unique personalities and charms of the voice actors and anime song artists, and the way their breath comes through more clearly, are completely different from the sound on a CD, and I would like people to be able to hear that alone.

Nomura: What I like about high-resolution is that I want to hear the "masters" as a simple fan psychology. It's not that I want to hear high-resolution, but rather, if I have a 44.1 kHz/16-bit music CD, I want to hear the original sound that was created before mastering, such as 48 kHz/24-bit. That is basically what I am talking about, so I am recommending MQA because it is a sound closer to the master.

Shikura I see. I simply don't want to lie. I am a technophile, so I can explain things scientifically, and I can say, "This is why this part is the way it is, but this part is really good. And since this part is a song or other live sound, if I don't explain the reason for the song or sound, the audience will feel cheated. To achieve this, I think the production side needs to further develop the high-resolution sound to the point where it can be properly called high-resolution. However, it is only a matter of time. Human beings have not evolved at all, but digital technology is evolving very quickly. Once the resolution of not only sound but also images exceeds a certain level, it is an area beyond human control.

Nomura: I have high expectations for you. Now, it seems that it is time for us to leave.

Shikura What?　No, it's not time!　I haven't talked enough about this yet (laughs).

Nomura Then, let's talk about it at another time in another project.