No.626:Sony "High Sound Quality" MicroSD Card Developer Asked a Question
Details of "graphs with no numerical values" are also included. Stability/quality maintenance efforts
(2015/3/9 12:01)
The SR-64HXA microSDXC card was recently released by Sony on March 5 as a "microSDXC card with a focus on sound quality," and AV Watch immediately published a news article about it on the day it was announced, resulting in a great deal of spreading via Twitter and Facebook, However, most people's reaction was that it was an "audio occultist's story. The majority of people's reaction was that it was an "audio occultist's story," that "digital media can't possibly change the sound," and "is Sony really okay with releasing such an occultist product? .......
In fact, the author became aware of this article late at night, almost half a day after it was announced, but in the meantime, several people had sent me messages via Twitter and other means asking me to check whether Sony's high-quality sound memory card was genuine. We heard that it is very difficult to measure and verify the audio characteristics of these cards, but it is hard to believe that Sony would commercialize something that is completely ineffective. We thought it would be a good idea to talk to Sony to get to the bottom of the matter, so we contacted them via the editorial department, and they immediately responded with an OK for an interview.
We asked four people involved in the development of the SR-64HXA why they had developed such a product and whether there was any evidence that it would change the sound. The four people who responded were Shingo Aso, who is responsible for the SR-64HXA project; Yozo Goto, who is in charge of product planning; Kenichi Satori, who is the project leader for development and design; and Hiroaki Sato, an engineer in charge of sound quality for the high-resolution Walkman ZX series and other products. The three, with the exception of Mr. Sato, belong to the Memory Media Product Department of the Storage Media Division of the Device Solution Business Group, a department that usually pursues "higher capacity, faster speed, and more reliable memory media," regardless of sound. We immediately asked them about various topics.
From left, Yozo Goto, Storage Media Division, Device Solutions Group; Kenichi Satori, Storage Media Division, Device Solutions Group; Hiroaki Sato, V&S Division, Video & Sound Division; Shingo Aso, Storage Media Division, Device Solutions Group
How did "high sound quality on memory media" get started?
--First of all, please tell us how and why Sony came up with the idea of developing a "high-quality sound memory card.
Mr. Yozo Goto, in charge of product planning
Goto: "Proposing a new memory card that only Sony can offer" is a theme that we always consider when planning our products, and the idea for this product came out of a discussion at ...... about whether we could add value other than simply increasing the memory capacity. The idea for this product came out of that. The idea for this new product came out of the question of whether it would be possible to add value other than simply increasing memory capacity. If we could make a difference in sound quality by doing so, we thought this was a product that only Sony could make, and that no other manufacturer that only deals with memory could do.
Shingo Aso, head of the SR-64HXA project
Aso: Every year, each division submits ideas for new product planning, and "storage with good sound" and "media with good sound" are themes that have been frequently raised. We also hold cross-divisional discussions on these topics, and around last summer, we came to realize that we could really change the sound, so we decided to work on this.
An approach to improvement from memory cards. Electrical noise refers to noise generated outside the audible range, such as unwanted radiation
--With old analog media such as cassette tapes, it was obvious that the sound quality would change depending on the material, and everyone could recognize the difference. But isn't the idea of "good sound" with silicon media, which is digital and does not undergo any data-like changes, outlandish?
Aso: Certainly, in the past, we have focused on copyright management and other issues in silicon media, and we have been thinking about sound quality based on the assumption that it would not change because it is digital. In that sense, we did not even have the idea of "creating a memory card with good sound. I felt exactly the same way as you have been saying on Twitter, etc., and that was my original position.
Kenichi Satori, project leader of development and design
Satori:All the engineers in the division were of that opinion, so I think it is natural that many people are wondering about this announcement. However, when I first heard from Mr. Sato, who is in the Walkman ZX unit, that "the sound differs depending on the media" during the communication between the divisions, I honestly thought, "Really? I honestly thought, "Really?
--I see, so even though we are part of the same Sony company, the memory media division, which handles semiconductors, did not originally have the idea of "high sound quality on memory media.
Aso: As a division, it is a department that handles storage in general, and in terms of optical discs, it is not the first time that we are pursuing sound, since we have also produced "sound masters" and the like. But, as one might expect, there was no such idea in silicon media.
--But from Mr. Sato's point of view, who is looking at the sound quality of Walkman, there is a difference in sound depending on the memory media.
Hiroaki Sato is in charge of sound quality for the high-resolution Walkman ZX series and other models
Sato: In February of this year, we released the NW-ZX2, a high-resolution Walkman model. Also, at the end of last year, we released the NW-A10, a high-resolution Walkman, each of which is equipped with a microSD slot to enable playback using external media. During the development process, we were conducting sound checks, and it was clear that the sound was different when played back using the internal memory and when played back using microSD. Since there was so much difference, I thought it might be possible to visualize the difference as data (......) and measured the frequency response, SN, dynamic range characteristics, THD+N characteristics, etc. However, it was still difficult to find any difference. However, when I listened to the data, there was a clear difference, so I approached the Memory Media Products Department for further verification.
Satori: Mr. Sato ordered us to provide him with a variety of products, which we categorized into different parts used inside, different firmware, etc. Some of the products were the same inside, but had different colors. Some of them had the same contents but different colors. There were probably more than 30 different types as we sorted them in detail. ......
Some of the microSD cards used in the listening test. In fact, they listened to and tested over 30 cards.
--Sato: Yes, there was a significant difference in sound.
Sato: Yes, there was quite a difference in sound. Of course, I was not the only one who listened to the cards, but many members of my department listened and checked, and it was the same. What surprised me was that there was a difference depending on the color of the card.
--I was surprised that there was a difference depending on the color of the microSD card. ......
Sato: Of course, this also does not show up in the measurement results, which is very frustrating for us as electricians, but since the members around me also gave the same impression, I think it is certain. However, when I looked into it, I found that different colors have different distributions of metal powder mixed into the paint, so I thought that these things might have some influence on the results. ......
There was a lot of back and forth between Mr. Satori (left) and Mr. Sato (right).
Satori: To be honest, I was skeptical at first, but when I interviewed everyone, not just Mr. Sato, they all said the same thing, so I was convinced that this was the case. I myself am not an expert in audio, but after Mr. Sato pointed this out to me, even I could tell the difference when I listened to it, and I came to believe that there was a possibility of commercializing this product.
--I think that this could be a possibility for commercialization. Is it possible that it is simply a variation from one individual product to another?
Sato: At first, I thought so too. However, we were provided with several pieces of media of the same configuration, and when we compared them, we found that the same type of media sounded the same, while different types sounded different. We selected the ones with the best sound.
What do you mean here by "good sound" and "bad sound"?
Sato: There is no difference in SN in terms of measurement, but there are hard sounds and soft and gentle sounds, and I feel that soft and gentle sounds are closer to the original sound. This was partly because the technology was still in its infancy, but as technology has evolved, the sound quality of CDs has also improved. With the advent of high-resolution recording, however, it has become possible to recognize differences in sound to an even finer level, and I believe that these minor differences are now becoming more noticeable.
What does the much-discussed "graph with no numerical values" indicate?
--What I would like to ask you about is the graph that was mentioned in the article. If it is impossible to measure, I would like to know why there is a graph like that, and since there are no numerical values on that graph to begin with, I think many people on the Internet have called it "fishy. Could you explain a little?
Graph shown by Sony at the time of the product announcement
Aso: For a variety of reasons, we did not present the graphs in a proper manner, which may have been a factor that aroused suspicion. However, these are the results of measurements taken to an in-house testing facility. Specifically, Sony EMCS. At first, there was some discussion about not disclosing the results because it might cause misunderstanding, but we thought it would be an appealing point, so we disclosed the results in a slightly blurred form. Incidentally, the frequencies above this level are in the realm of cellular phones, so we stopped.
Satori: As you can see, the horizontal axis shows frequency, but the unit is MHz. In other words, it is not in the audible range to begin with. Therefore, the noise shown here is not directly audible to the ear.
--I see that the horizontal axis is an exponential graph. If we look at the spacing between the lines on the horizontal axis, we can see that it is an exponential graph, and if we look at ......, we can see that the top of the graph is 800 MHz, which is below the cell phone range. What does the vertical axis mean in terms of radiation intensity?
Satori: Radiant intensity is a physical quantity that expresses the radiant energy emitted from a point-like radiation source in a certain direction per time. We measured this using an actual Walkman. First, we took out the circuit board of the Walkman in the form of three pieces, and inserted the memory card to be measured into the microSD slot on the circuit board. The shape of the microSD slot is exactly as shown in the figure. In other words, I put a probe on the microSD slot and measured the magnetic field in the vicinity of the microSD slot. The reason why we did not show the numbers on the vertical axis in the graph is, of course, an internal confidentiality issue, and we did not reveal this information to other companies so as not to leak our know-how to them.
Radiation measurement is illustrated: microSD card is inserted into the slot and the nearby magnetic field on the slot is measured with a probe
A different technical approach from conventional "high-speed memory cards" is also used.
--So you say that these measurements made a difference in the sound. But if the noise is not in the audible range, it should not affect the sound, right?
Sato: That is correct. However, in the case of audio, the areas around the power supply and analog circuits are very sensitive, and if they are affected in any way, there may be a difference in sound. In other words, it is highly possible that the noise shown in this graph is not what you are hearing, but that it is affecting the surrounding circuits.
--it seems likely that this is the case. When this Digital Audio Laboratory started its serialization 14 years ago, it was under the theme of "Does the sound change when a CD is copied to CD-R? We conducted various experiments and interviews under the theme of "Does the sound change when a CD is copied to CD-R? As a result, we concluded that although we could not measure the sound in terms of data, there was indeed a difference in sound, and that the factor was probably the electromagnetic waves generated by the movement of the servo motor when reading the CD, which adversely affect the surrounding analog circuitry. ...... This story is very similar to that. This story is very similar to that.
Sato: I also examined CDs extensively about 15 years ago. I knew from that time how the servo current affects the surroundings, and that if the RF here is clean, the sound will be better as a result. It is a very classic method to strengthen the power supply with noise suppression ......, but it works for the sound as a result. If I had not had that experience, I probably would not have come up with this idea.
--When I was verifying CDs at the time, I was convinced that the sound could change even if the data itself did not change, depending on the media conditions, but I wondered if isolating the power supply circuits and audio analog circuits would not cause any problems. However, I thought that if we isolated the power supply circuit and the analog circuit of the audio system, the problem would not occur. At the time, critics often said things like, "This media extends high frequencies well, but this one makes vocals clearer," but I wondered if this was a story specific to individual devices and not applicable to media in general. So I would like to ask you, what did you use to verify the sound this time?
Sato: We used the NW-ZX2 Walkman and NW-A10 series, which were prototypes at the time. Both of these devices had a similar sound when the media was changed.
--I see. If that is the case, isn't it possible that the good sound of the SR-64HXA is specific to the Walkman ZX2 and A10? For example, if you use a PC's microSD slot for music playback, wouldn't that have a much smaller effect? It seems a bit unreasonable to say "this memory card has high sound quality".
Walkman NW-ZX2 on the left and NW-A10 on the right
MicroSD slot of Walkman NW-ZX2
NW-A10 microSD slot
Sato: We did not conduct rigorous tests on other companies' products, so we cannot say that the effect is the same for all products, but it seems that the effect has a similar trend, including playback on smartphones. However, it is difficult to say because we do not understand what other companies' products are trying to achieve in their sound production. In some cases, the sound that was dimly audible became clearly audible, and as a result, the characteristics of the audio equipment were clearly conveyed, giving a bad impression. ...... It is true that pulling the analog circuitry away from the microSD slot is an option, but I think it is inevitable that it will be affected if the ground is connected.
--I think I have a general idea of what is going on. From what you said earlier, you chose a combination of various components to create a memory card with high sound quality.
Goto: Until now, the technological approach to high-speed memory cards has been to increase the bus clock speed by increasing the bus size for exchanging the controller and NAND flash, and to increase the speed of the bus between the host and controller. In contrast, the new high sound quality approach is designed to optimize the strength of the controller and NAND driver by lowering the bus clock as much as possible to the extent that the host device does not run out of sound. In the same way, overclocking a PC can cause instability and noise, but if the clock is lowered, the PC will stabilize, although it will not achieve full performance. It is a way of increasing stability by providing a margin of performance.
Satori: Then, it is difficult to say whether it is enough to lower the clock anyway. The only way to check this is by listening with the ears, so we actually use prototypes and conduct auditory tests over and over again before commercializing the product.
Technical Approach to Realization of High-Speed Memory Cards
Technical Approach to High Sound Quality
--Goto: So you are not just commercializing existing memory media that have been evaluated as having good sound quality as high-quality sound memory, but rather designing them to suit the needs of the market?
Goto: That is correct. We tuned it to be optimal through repeated listening tests. In addition, as a manufacturer, we have a responsibility to continue to guarantee this quality as long as we claim "high sound quality. In reality, memory media frequently changes NAND and controllers, and firmware and parameters are replaced with new ones. Therefore, even if the part number does not change, it is often the case that the contents of the media are slightly different. However, in the case of high quality sound models, we cannot change components without permission, so if there are any design changes in the future, we will be sure to check the sound quality to maintain the current quality.
In response to the question, "If you accumulate buffers, won't microSD have any effect? Answer to the question
--I would like to ask one last question about the design of Walkman. If this buffer size is large, I think the microSD card operation will be unaffected.
Sato: Yes, we do have a buffer, but it is not that large since this is a consumer device. We have a certain amount of leeway to avoid sound interruptions, but even so, we do not have the capacity to load an entire song. Also, it doesn't turn off immediately after reading, so I think that the electromagnetic radiation from here will inevitably have an effect.
--I understand very well. Thank you very much.
After this exchange, we were allowed to actually listen to the sound using Walkman ZX2 and Walkman A10. I will not give a sensory evaluation of the sound here, but it is true that the sound did indeed sound different. This difference was felt not only when playing back high-resolution sound but also when playing back CD sound; there is not so much difference as the difference between the MP3 192kbps sound and the CD sound, but anyone can recognize the difference if they listen with attention to certain instruments or the vocal quality of some of the vocals, can't they? However, the sound was confirmed only with the Walkman; there may be room for further verification of playback using a PC or other devices.
