Mr. Hisaaki Hijikata (Hereinafter all indicated by surnames only): First of all, could you introduce yourself to the readers?
Mr. Gilad Tiefenbrun: I am a Managing Director of LINN. LINN’s general manager of business.
Mr. Keith Robertson: I’m at the position of the chief technical officer of LINN.
Mr. Tomonori Sato: I am developing high-end ADC’s and DAC’s in AKM, and making sounds color of those products.
Hijikata: So, I would like to ask you about the background of how this remarkable collaboration took place with a technical explanation of, “KATALYST.”
Hijikata: The “KLIMAX DS/3,” I have listened earlier had a wonderful sound far beyond my imagination.
Gilad: Thank you very much. The, “KLIMAX DS/3,” has become the biggest upgrade so far, renovation of the DAC section. The biggest feature is the adoption of the new generation DAC architecture, “KATALYST,” which is the fourth generation.
Keith: LINN has a long history in understanding the D to A process in its entirety. Our 1st generation DAC architecture was in 1991 with the, “NUMERIK,” product. There was a D to A character numeric； it was a D to A process for CD players. The 2nd was in 1998 with the very famous “CD12.” The 3rd was the product that Gilard and I started, which was in 2007, which is the “KLIMAX DS.” Today, I would like to introduce to the situation LINN’s 4th generation DAC architecture, which is featured in the new LINN KLIMAX DS and KLIMAX DSM.
Hijikata: About the new KLIMAX DS, the most interesting point is the innovated DAC block. Could you explain the new technology?
Keith: It’s important to understand that the DAC means many things to many people - the word DAC. To some people it means a chip, to other people it means a product. And to other people like LINN, it means the entire process of converting anything that’s involved in the conversion of digital, and then from that creating an analog signal - the entire thing. “KATALYST” is a completely new look at how to do the DAC architecture having independent power supplies for the master clock and each stage.
You can see in this picture here that there are fundamentally three things that are required to make an analog signal. Data, Clock, and Reference Level. From those inputs you can actually create the analog signal. The output is the music signal that you will eventually hear.
Keith: First, data needs to be prepared for creation. You can’t just input and convert the data. You need to prepare the data for the creation process. Any inaccuracies in the preparation of the data, if you introduce inaccuracies, they will be there as you create the signal. When we talk about these very minor errors that you get in there, there is no way to take them out later, so it’s irreversible damage to the signal, and that’s why it is so critically important to get precision at exactly this point of creation.
Sato: AKM believes that, “high-bit,” and, “high-rate,” are essential factors to reproduce, “real live sound.” This idea of AKM will match with the LINN’s thoughts.
Keith: The second is the clock, and the clock controls the precision of the timing of the music that you hear. And every clock - tick, tick, tick - you create a new signal level. So you can really think about the clock as controlling the horizontal accuracy of the end music signal.
And the last is the reference level. The reference level is the actual signal prior to being modified, before it is created into a new form. And what happens is that every clock, you take the reference level, and multiply it by the data. So, you multiply the reference level by data, data, and data. The clock controls this timing and the reference data controls this amplitude.
The middle part of the signal flow, that is called creation, LINN gave this process a new name as we are innovating the new DAC circuits instead of the word; conversion, that’s the technical name.
Hijikata: Then, AKM’s flagship DAC chip, the “AK4497” is adopted.
Keith: We have been very pleased to work with AKM on this product because AKM is the best at the actual act of the creation process itself in the center. We’ve been able to use AKM and what they are best at and we’ve been able to use LINN for all the things that LINN is best at.
Hijikata: That’s Great !
Hijikata: Later, I would like to ask you more about the development and collaboration style, but now, could you tell us how this collaboration was actually realized? and Why it was AKM‘s AK4497?
Gilad: Let us go through the process. Why we chose AKM and what we did.
Keith: So there were three aspects through why we chose AKM. The first aspect is characteristics of the chip. In late 2015, (about this time last year), we were looking at all the many options, a number of new options and new ideas. We had a bunch of ideas. What we did is we’ve listed up and got all of the available high-quality creation chips.
Sato: We have been contacting LINN for about 10 years introducing our products including demo boards. Of course we approached LINN when we developed the AK4497. It was a good timing I suppose when we introduced the AK4497 to LINN.
Keith: And we created a special KLIMAX DS. The special KLIMAX DS took plug-in cards that we were able to create a D to A process, a LINN D to A process, but based on about six different concepts. And this allowed us to plug them into the board very quickly, so you can put A or B or C, all ways, all six. So it allowed for a level playing ground, a fair comparison between all of them. AKM performed very well, the best in measurement.
I would like to show one or two measurements that we have been able to make of the product. So, the first measurement is a total harmonic distortion. The blank graph along the bottom is the frequency spectrum, and along the left is magnitude. The idea in a total harmonic distortion is that you request that the DS plays a 1 kHz (0 dBFS) tone, and what you would expect to see is a 1 kHz response and ideally nothing else.
In reality there will be some harmonics, but if you have harmonics that are at two, three, four, five, all the way through, you wish them as low as possible, and you would prefer lower order harmonics to higher order harmonics. I’ll show now the measurements for the KLIMAX DS/2, the one before the new KATALYST-based KLIMAX DS on the left (red) and measurements for the new KILIMAX DS using KATALYST on the right (blue).
Hijikata: These results are quite different. I can tell that there are only few distortions after 5kHz in the right graph. In addition to that the output frequency is very low comparing to the DS/2. I guess even sound professionals will be surprised by seeing this result.
Keith: Although DS/2 performance is quite good, the DS/3 has excellent result especially around 2kHz and 4kHz, greatly reducing harmonic distortion. I’d like to show you a second measurement now, which is a very unusual measurement. We have never done it before, and as far as we are aware, nobody has done it before. The total harmonic distortion measurement shows the aggregate measurement of the product from input to output. So this new measurement is a measurement of the reference level stability.
What we have never been able to do before is to take control of that to provide an independent, isolated and tailored reference level to the creation process that is immune to the processes going on around it. The same idea, frequency on the bottom, magnitude on the left. We are going to again request that the DS plays a 1 kHz tone. What you would expect to see is a 1 kHz response and ideally nothing else.
Keith: Measurement result for the DS/2 is in red and measurement result for the DS/3 is in blue. Here are the measurements for the KLIMAX DS with KATALYST. You can see that there is at least a 10dB drop across the board. In the noise floor, in the 2 kHz disturbance have dropped by about 10dB, and on top of that, all of the other small disturbances are completely gone.
Any inaccuracies in the preparation of the data will have big affection to the signal creation in DAC. It does not prove audio quality - but proves precision engineering. As you listen earlier we would see performance potential of the “KATAKYST.”
Sato: DAC chip has the VREF pin that supplies reference voltage. The reference voltage is often thought as, “just connecting the power,” since it has the same level as the power supply voltage. However, AKM designs a chip with a strong philosophy that “VREF is a signal, it is not a power supply.”
Keith: These are just measurement results, but as we know, came up to great outcomes. As the second aspect of choosing chip, we evaluated the architecture and how much access we could get to all off the different parts of the D to A process. And as we have already described, LINN and AKM working together has allowed LINN to take control of all the aspects of the input, and for AKM to focus on the creation process.
Sato: AKM’s chip is designed with high flexibility to satisfy customer’s different demands. However, without understanding the architecture correctly, this high flexibility might cause unexpected results. To prevent this situation, I went to the Scotland to meet Keith and other key engineers including Ian.
Keith: This diagram shows LINN’s DAC operation concept. First, data is input, and output finally after several processes. Gray blocks are the parts that are processed by AKM’s DAC chip. Blue blocks are the parts that are processed by LINN. Red lines describe power supply concepts.
Sato: Mr. Robertson mentioned that the clock is the one of most important factor. In fact, AKM has the same philosophy and our device has a dedicated power supply for clock, not only the AK4497 but for other devices since 10 years ago. This might suite the “KATALYST” system.
Sato: And our chip has greatly improved distortion by new, “VELVET SOUND,” architecture that is AKM’s innovated technology. This time, we also reviewed the LSI production process for the AK4497. We developed a new LSI production process only for premium audio devices. With this new process, the sound quality characteristics have hugely improved, achieving the best sound quality ever.
Keith: So LINN has always had an interest in all aspects of the D to A conversion process. We have been very pleased to work with AKM on this product because AKM is the best at the actual act of the creation process itself in the center.
The result that you hear on the KLIMAX DS is the combination of AKM’s ability at the creation, but also because of AKM’s wisdom in allowing us to get access to all of the elements of the creation of the music signal. So all of the inputs are able to be controlled by LINN. LINN has been able to take control of an unprecedented number of different aspects of the input to the creation process.
Sato: There was no hesitation about how much we could disclose the architecture of the chip. What we did was to work together and to cooperate as much as we could to realize technical requests of LINN.
Keith: And then the third aspect was the most important; both first and second matter, but then the 3rd is to listen to all of these. So the result of what we were able to do because of one and two and how LINN was able to use the AKM creation meant that the result of three was also the best.
Hijikata: People working at LINN are known as they have good ears. It is very remarkable that they did accept AKM’s sound quality. I guess this is a very good collaboration example not just providing a chip but also cooperating each other.
Gilad: Can I tell a little story? We had as Keith said a number of different comparisons we were making, and so AKM was here and then the next one, (I am not going to say who it was), was here, and the CTO from that company came to LINN to listen and we made him the comparison between his board versus the AKM’s board, and even he had to admit, yeah…..So we played the first track and he said: “Mmm, OK,” then we played the second track and he said “Mmm” and then when we played the third track he said “OK.”
Sato: In the end of July, Mr. Robertson came to Japan with a prototype of DS/3 to demonstrate the sound for us.
Keith: It was a prototype…at that time there was only one piece; I had to bring it with me in my luggage, and take it back.
Sato: I was very pleased to see the actual set that includes our technology and proud of being involved with the update of “KLIMAX” this time. When I listen to both the DS/2 and the DS/3 to compare, I realized how much improvement in sound quality has been made and I felt a relief that we could achieve the sound we were expected. At the same time, I was also thankful that he came to Japan to show the prototype, thinking us as a sound creation partner.
Hijikata: So now we have the KLIMAX KATALYST, DS, and DSM. Do you have plans to put KATALYST on other products?
Keith: Yes. We will focus on growing KATALYST, providing KATALYST to the KLIMAX range of products. So, KLIMAX 350 and KLIMAX EXAKTBOX.
Hijikata: Wow, my, “KLIMAX EXAKTBOX,” will integrate, “KATALYST,” in the future! I cannot resist the urge to update as I listened to the sound of, “KATALYST,” today.
Off the subject, the KLIMAX EXAKTBOX has high performance in digital domain area such as band division and each compensation. It is drawing attention and many people in the industry come and listen to the system.
Keith: So we have seen a very big improvement with KATALYST in the DS and I expect a bigger improvement in the EXAKT KATALYST.
Sato: Because there are two KATALYST engines in the DS, there are six KATALYST engines inside of a KLIMAX EXAKTBOX.
Keith: LINN has a policy that is to develop technologies ourselves as many as we can to not have untouchable parts for audio products. In this sense, it was very good that we could collaborate with AKM scrupulously and openly in technological exchanges.
Hijikata: So using the best chip in characteristics freely based on collaborative relationship brings on results in sound quality.
Sato: Our philosophy of the sound could take shape via LINN’s product. The direction that LINN is aiming to achieve is very interesting, and it would be great if we could keep helping it by developing new technologies.
Hijikata: Through this interview, I could see this collaboration was realized with great relationship based on trust between LINN and AKM. As an audio fun, I am very pleased to see that an audio manufacturer and a device maker have designed a DAC circuit together based on a great relationship. I appreciate all of your careful explanations. Our readers will be able to understand well. Thank you very much taking time today.