This review page is supported in part by the sponsors whose ad banners are displayed below

With the current-mode BNC-terminated Satri-Link connection even 100-meter long interconnects are claimed to suffer no signal loss. This overshadows usual advantages of balanced-line transmission, hence no balanced inputs. The absence of remote control reflects firm headfi focus where no wand is required. The lack of line-level outputs underscores once more that the HPA-21 was never meant to double as preamp and is a very narrowly focused extreme machine.
Soo In specified a fine Tokyo Ko-On Denpa 2CP Series 10kΩ attenuator with a resistive ladder circuit. He enjoyed not having to worry about internal EMI from a motor controller IC. Bakoon's advanced bias circuit is said to create thermal readiness in 10 minutes. This lays to rest concerns over wasting battery power for lengthy warm-up calisthenics.

With my first headphone on the desktop—Beyerdynamic's sealed 102dB/32Ω T5p—even at full left the machine still leaked sound. The next few millimeters on the volume control betrayed imbalance before both channels locked in. Then things got too bleeding loud a bare crack beyond. Low-gain current mode delayed this by a few more millimeters. Yet useful range with the Wyred4Sound mINT's fixed output was intolerably narrow. This couldn't be true. I wondered about the chosen taper. To see whether I'd chanced across an utter fluke, a pair of Sennheiser HD-800 off AURALiC's Vega were next. Those reacted just the same.

Soo In Chae: "I can see why this would be an issue. Your T5P seems to be the most sensitive headphone I've ever seen due to its low impedance and high sensitivity. I understand why they call it portable. As you already know, the Satri circuit takes a full 2V input signal and and controls gain in the output section. This is one reason why it's better than conventional designs which reduce gain at the beginning to sacrifice signal purity. But with our circuit, gain could be too high for headphones like yours. Please listen to the HPA-21 with your LCD-2 first. In the meantime I'll discuss with Nagai-san if there might be a fix for high-sensitivity headphones. Our Satri circuit determines gain by the ratio of input resistor at the voltage/current conversion just prior to the Satri IC to the output resistor at the volume pot right after the IC.

"The pot's resistance at minimum is small but it's still 2-3Ω. That's why there's leakage even at the lowest level. This is not audible with less sensitive headphones or speakers in the case of the AMP-11R. But it's there and you can hear it with highly sensitive headphones like your T5P. I wonder how many high-end users will use the HPA-21 with such headphones. If this is the case, we'd need to think about a design change. Otherwise the current gain level seems appropriate especially since you can adjust the output gain on the current output. And this we could easily change by simply switching the current output's gain-setting resistors. The voltage output gain level would still require a design change if at present it's too high." The likeliest fix for high-efficiency cans and/or high-output sources—4V outputs after all are quite common—will be to reduce lo-gain's amplification factor. My question was whether the attenuator's current taper might not still come on too quick across its first quadrant. This could be addressed with a 5kΩ pot. Here Soo In cited 8 to 10-week lead times for TKD to make this an option for the next production run. The gain resistors for lo-gain current mode could be changed right away.

Swapping in my favorite Audez'e from the nightstand rig hit true mute at the left stop in lo-gain current mode though hi-gain and voltage mode still leaked. But now the taper was more gradual and channel tracking perfect in all modes. The white power LED of my silver loaner confirmed a precharged battery. An hour later a red light appeared beneath the central toggle. This probably warned of impending power loss. True enough, a bit later the white light extinguished. Time for a full charge with the included SMPS wart and its figure-8 two-prong power input. Without listen-while-you-charge function, when the power goes out, it's lights out and concert over.

The four thin cork wavers mirror what goes for footers beneath the AMP-11R too. It suggests that like the latter's RCK-11, a proper decoupling undercarriage could be in the works. If a DAC-21 is planned to complete a current-mode Satri-Link'd headfi system—hello please!—one expects a similarly styled two-tier stand.

Time to take the measure on sonics. Retreating upstairs into my dedicated headfi rig—160GB AIFF-loaded iPod Classic, Pure digital dock, Burson Conductor as DAC, Bakoon AMP-11R as headphone amp, Audez'e LCD-2—the superiority of current drive was plain. It was hard fact, not marketing fiction. The effects were super resolution and ultra clarity. This manifested as heightened simultaneity of intelligible detail. That's a mouthful. It means that by a simple act of attention, the tiniest feathered cymbal tick was as completely drawn out in its harmonic halo surrounding a clear physical action as were the forward-mixed main instruments and everything between. To illustrate, let's go into the kitchen. Cooking a stew we expect our individual ingredients to blend. This 'mushification' gives up the discrete separation of flavors we began with. We end up with a new group flavor. Perhaps it retains some layering which allows a professional chef to identify everything (or most) of what's in our stew. But in general it's a changed thing of considerable complexity.

This was very different. Very fine brush work on a cymbal's edge so minuscule compared to the louder bass runs and various other instruments was perfectly unchanged; just as though it'd been cooked (recorded) separately all by itself. This was true for each and every detail (microphone). By simply shifting attention, I could zero in and drill down on what each had captured as though in isolation and as a single track in the sound booth. All the ingredients were still neatly laid out like on a TV chef's kitchen counter. I could inspect each in all its complexity of harmonic shifts, string/wood components, air/finger pressure modulations surrounded by finger slides, key clacks or other giveaways. It didn't matter how far down the mix anything was, how subdued it was versus louder foreground stuff. It all was one massively paralleled event of discrete overlaid slides. I could decide which to inspect or whether to look through the entire stack at once.

In 'holistic' listening mode this was richer, fuller, denser and more colorful. In 'analytical' mode it was the ability to lock down on the tiniest item and see it clearly, then follow it unwavering and constant no matter what happened around it. Analytical mode on everything at once relies on a quasi-meditative state. Your mind can't get stuck to anything or you'll be out of the flow in that moment and as such out of the oneness that sees/hear everything. Should you be capable of such fugue vision, the HPA-21's astonishing degree of resolution supports it ideally.

Bakoon would tell us that the cause was significantly lower distortion. That's about less interference. What's there as recorded substance is undisturbed, laid bare and made obvious. Easy. That's true and not manipulated resolution. With photography true resolution begins with a superior lens, proper lighting and a tripod. Manipulated resolution is corrective. It happens in Photoshop. It's artificial and routinely creates pixilation, unnatural saturation, color shifts and more. The HPA-21's higher current-mode resolution didn't sound like a trick, after-the-fact correction or enhancement. It sounded like a byproduct of less interference. Some disturbance or filter had been removed. And that made what was recorded sound better because I heard more of it. Despite the wordy abstractions, this was a simple outcome of lower distortion = higher resolution = better sound. And it begged a question. Why had Bakoon even bothered with their voltage output when current drive was so clearly superior? Was it as proof by contrast? Or perhaps not all headphones respond similarly well to current drive? Time to play musical chairs.

But first more background on 30-year old Soo In Chae. "My college major was in industrial engineering. My last job was at one of the most advanced inspection equipment manufacturers in the nano tech sector. We customized and developed products to meet client specifications. I sold one to a well-known German company which cost over €1.000.000. This had a lot of engineering challenges which I as product manager had to overcome with some of the very best engineers. During this time I learnt how to make a good product and how to collaborate with top personnel. This affected Bakoon Products International where I again work with possibly the best engineers to create something perfect or at least superior to the usual high-end hifi products. About four years ago I quit my last job. I was about to pursue a Ph.D in the U.S. Then I was fortunate to meet Nagai-san (I still was a customer) whom I considered to be the top audio/electrical engineer in the field. After I got to know him, I realized that I could become part of Bakoon. He felt the same and so I started my company.

"Along the way I found and was introduced to top engineers and suppliers with whom I now collaborate. We do not simply 'repackage' Nagai-San's Japanese products. Each time we release a new model, its circuit is more advanced than the original's. The HPA-21 vs. HDA-5210MK3 for example had the HPA-21's current output completely redesigned to perform better. Sometimes I do this also by upgrading parts. But I'm no electrical engineer. I don't get into too much electrical engineering myself. I understand enough to discuss with Nagai-san how to improve things. Then we have a Korean master engineer who develops measuring/inspection systems for Samsung factories all around the world. He develops or customizes auxiliary circuits to my needs. With him we successfully designed the perfect battery power supply for example."

My currently favorite speaker in Soo In's system. soundkaos and Bakoon would co exhibit at Munich HighEnd 2013.

A few days after my email about the gain, he had this: "I just wanted to update you that we made a quick fix which has lowered the low-gain level of the current output. With my Grado SR225 for example—not the most sensitive design but quite so at 98dB/32Ω—my 'normal' listening level with a standard 2V source now is around 12:00 to 1:00. This falls into the gradual range of the Tokyo Ko On and I believe clears your concern about the T5P or similarly sensitive 'phones. The fix was applied to all units we sent off today and I hope we've met more customer demands/situations with it."