The effect of break-in: Vsonic VC02

Introduction

It is generally known that a new pair of headphones must go through a "warming up(aka break-in)" phase for certain period of time before it opens up its true sonic potential. Previously, both electroacoustic and psychoacoustic aspects of a 100 hours-long headphone break-in have been presented with SONY MDR-EX1000, Sennheiser HD650, and Fostex T50RP; The physical transformation is evident, yet its degree is not of a night and day difference, as audiophiles normally describe it.

On Vsonic VC02's cable, a tag, which is written in Chinese, is attached by the manufacturer, openly claiming the dramatic sonic change. The manufacturer states 100 hours of break-in improves the sound quality of VC02.


Test methods

A brand-new Vsonic VC02 is to be broken-in for 100 hours straight with XLO's break-in sample (100 dB SPL @ peak), driven with an Objective 2 headphone amplifier. Once the headphone is *placed on the occluded ear simulator, its physical placement must remain untouched for the next 100 hours in order to prevent any placement-related deviations. Each 10 times averaged pre break-in & post break-in measurement data are to be compared, and should any type of change occur, they are to be reproduced back utilizing a binaural recording technique to be ABX-compared.

Objective Assessment

Impedance

Overall impedance is slightly decreased after the break-in. This is the similar behavior observed from previous break-in experiments.

Frequency response

The resonances in the mid frequency range shifted. Interestingly, this resonance shift has been reverted once the voicecoil cools down.

Phase response 

Harmonic distortion

Although its degree is rather minimal, the amount of lower frequency distortion has been lowered. Again, such distortion decrement has been seen from previous experiments as well.

Time-domain characteristics 

Impulse

CSD

ETC: pre-break in / post-break in

Group delay

Subjective Assessment

ATTENTION: In order to accurately reproduce these binaural recordings, listeners must use a flat diffuse-field equalized headphone, such as Etymotic Research ER-4B and STAX Lambda PRO with ED-1. The accuracy of the reproduced result can not be guaranteed otherwise. The reference sample is an excerpt from the free hi-resolution sample provided by 2L - the Nordic sound. Downloading below samples is highly recommended for critical listening experience.

Reference sample @ 24-bit / 96 kHz:

A brand-new Vsonic VC02 @ 96 kHz:

A 100 hour broken in Vsonic VC02 @ 96 kHz:

Conclusion

As previously confirmed by others, 1 2 3 4 5 6 , and by myself, 1 2 3 , the physical effect of break-in is quite evident. Moreover, the psychoacoustic aspect of Vsonic VC02's break-in has been demonstrated for ABX comparison so that listeners could verify the audibility of 100 hours of what-so-called 'warming-up'. 

On the tonality modification: Ultimate Ears Triple.Fi 10 Pro

개요

2004년에 무시무시한 출시 가격과 함께 첫 선을 보인, 現 Logitec社(이후 로지텍)의 Ultimate Ears(이후 UE) 브랜드 (舊 Ultimate Ears社) generic 라인업의 옛 플래그쉽 모델, Triple.Fi 10 Pro(이하 트파)은 그 특유의 박력있는 음색으로 많은 사랑을 받아왔습니다. 브랜드 자체가 로지텍으로 넘어간 이후에도 그 인기는 끊이지 않았으며, 특히 미국의 Black friday에 벌어진 3년 연속 $99의 특별 할인가는 트파를 더욱 많은 사용자층에 보급시키는 기폭제가 되었고 이로서 고가 IEM을 트파로 처음 접한 사용자들은 트파의 소리를 레퍼런스로 오해할 정도로 큰 반향을 일으켰습니다.

아쉽게도 현재는 제조사나 정식 딜러 모두 남은 재고가 소진된 상태이며, 사용되고 있는 트파 역시 거의 대부분이 보증 기간이 만료되어 문제 발생 시엔 사설 수리나 몰딩을 맡길 수 밖에 없습니다. 말 그대로 현재 사용되는 트파는 '고장나면 끝'인 시한폭탄이지요. 그렇다면 이런 보증 기간이 지난 트파를 조심조심 수명이 다 할 때 까지 사용하느냐, 혹은 요긴하게 재활용하느냐.. 고민을 해 봤습니다. 그러다 어떠한 계기로 인해, 트파도 얼마든지 유저 튜닝이 가능하다는 것을 알게 되었습니다.

그 계기는 바로 제가 노즐의 뚜껑(?) 부분을 아주 손쉽게 뽑아낸 일이었죠(;;;) 그리고 그 안에는 놀랍게도 Acoustic damper(이하 필터)를 간단하게 교체할 수 있도록 설계되어 있었습니다. 이런 뚜껑의 파손은 케이블 분리 시의 내장 적출과 더불어 발생하는 트파의 고질적 내구성 문제입니다만, 접착제만 가지고도 손쉽게 수리가 가능하므로 약점을 장점으로 바꿀 수 있는 수단으로 활용될 수 있을 것입니다. 물론 필터 오염 시 교체도 타 제품들처럼 손쉽게 할 수 있겠지요.


트파의 내부 구조

트파는 이와 같은 구성으로 이루어져 있습니다:

[고역 3300옴, 저역 2200옴이 맞습니다]

발음체는 고역/저역이 나뉘어진 2-way, 3 driver 이며 UEHI는 고역을 담당, 저역은 UEXXX 로서 이중체로 되어 있습니다. 


[공진이 제어되지 않은 트파]

UEHI와 UEXXX의 상세 스펙은 알려진 바가 없으며, 아마도 Knowles社로부터 주문 제작된 것이 아닐까 추측을 할 따름입니다. 저역은 시간차로 인한 위상 중첩을 보상하기 위해 후방으로 밀려나있지요. 각 대역을 담당하는 bore는 필터를 통해 음향적으로 제어되어 있습니다. 아쉽게도 bore는 알려진대로 나뉘어진 것이 아니고 필터를 지나면 노즐 뚜껑에서 바로 소리가 합쳐지게 됩니다. 우리가 흔히 일컫는 '돼지코 노즐"은 사실상 아무 기능도 없었습니다.


조정 방식

조정은 크게 세가지로 분류가 가능합니다:

1. 크로스오버 네트워크 회로 조정: 전기 회로에 대한 전문적인 지식과 더불어 IC소자 납땜 기술이 상당한 수준에 도달해야 가능한 방식입니다. 아무래도 일반적인 사용자들에겐 무리가 있겠지요.

2. 음향적 댐핑을 통한 조정: 노즐의 뚜껑만 따면 바로 행할 수 있는 방식입니다. 흔히 구할 수 있는 필터로는 Etymotic Research社가 판매하는 MC5의 680옴, ER4의 1500옴, Ety8의 2200옴 짜리가 있고, 또한 필터를 틀어막고 구멍을 내어 4700옴 이상의 음향 임피던스를 인가할 수도 있습니다. 또한 필터의 중첩을 통한 음향 임피던스의 증강도 가능합니다.

3. 전기적 댐핑을 통한 조정: 단순한 저항 어댑터만으로 손쉽게 접근 가능합니다. 기존의 인식으로는 출력 임피던스의 상승은 트파에 부정적인 영향을 미친다고 알려져있습니다만, 이를 역이용할 수도 있습니다. Etymotic Research社의 플래그쉽인 ER4만 해도 전기적 댐핑을 통한 조정으로 만들어진 제품이지요.

여기서는 2번과 3번을 집중적으로 살펴보도록 하겠습니다. 참고로 현재 모 처에서 판매하고 있는 소프트웨어 EQ App을 통해서도 음질의 조정이 가능하긴 합니다만, 소프트웨어가 가지는 한계는 아무리 고해상도로 처리한다 해도 극복할 수 없다고 봅니다.


음향적 댐핑을 통한 조정


UEHI 의 특성 변화입니다. 조정을 해도 대역폭이 크게 변하지 않는 점이 최대 단점이나, balanced armature형의 특징인 높은 음향 출력 임피던스로 인해 330 옴만 물려도 피크가 간단히 잡히는 모습을 보여줍니다.


UEXXX 의 특성 변화입니다. 2 kHz 에서의 공진을 이용할 수 있습니다. 


전기적 댐핑을 통한 조정


33옴만 물려도 고역이 크게 떨어지면서 중역이 보강됩니다. 감쇄폭을 충분히 상쇄할 수 있는 고역 필터의 사용이 필수.


몇가지 예제

각 필터의 음향 임피던스와 함께 전기적 댐핑이 가져오는 경우의 수만 따져도 수많은 가능성이 있습니다만, 여기서는 몇가지만 살펴보도록 하겠습니다. 각 데이터 하단에 사용된 필터 및 저항값이 기재되어 있으므로 추가적인 설명은 생략합니다.

Ultimate Ears UE900을 기준으로 한 조정


AKG K3003을 기준으로 한 조정



결론

지금까지 트파가 가지는 무궁무진한 잠재력에 대해 알아보았습니다. 간단한 조건의 변경 만으로도 사용자 개개인의 입맞에 맞는 음색으로 튜닝이 가능하며, 이는 결과적으로 단종된 제품의 희소 가치를 더욱 향상시키지 않을까 생각 해 봅니다. 여기서 나아간다면 회로의 개조 및 고역 발음체의 교체를 시도 해 볼 수 있겠으나, 그 정도의 전문적 지식을 가진 사용자라면 아마 이러한 클리닉은 필요가 없으리라 사료됩니다.

[혹시 특정 필터와 임피던스 추가 조합의 결과가 궁금하신 분들은 댓글로 문의를 해 주시면 제가 바로 측정 해 드리겠습니다]

Special thanks goes to my fiend, Inks

Vsonic VC02: The wonder of ⌀ 3 mm

Released in December of 2012, VC02 is an iem made in the mainland China, packed with innovations. First of all, VC02 comes with an electrodynamic transducer of 3mm in diameter, which is the smallest driver available in the industry to date, equipped. In addition, The cable is made of silver, and it is interchangeable with aftermarket IEM cables. Reasonably price at ~40 USD, Vsonic VC02 opens up a new scene of budget-fi.

PRO: Very nice channel matching for a budget pair. High frequency extension is excellent, resembling that of Ultimate Ears UE700. Due to its linear impedance characteristic, it is almost impossible to alter the sound of VC02 electrically. It is important to note the mid-frequency dip is neutralized once group delay kicks in.

CON: Inverted polarity. Slightly sibilant above 10 kHz, as seen on the CSD plot.

ON SECOND THOUGHT #1: It is quite interesting that the polarity of VC02 is inverted. All of electrodynamic IEMs I've measured so far have never been inverted! It is fortunate this can easily be fixed by simply replugging the driver unit upside down, of which the logo 'VC02' is flipped vertically.

ON SECOND THOUGHT #2: Insertion depth is critical for VC02, since the ultra-micro dynamic transducer used here has very high acoustic output impedance. Keep this in mind: the shallower the insertion gets, the peakier the tonality becomes.

ON SECOND THOUGHT #3: Although there's a vent at the bottom of the unit, it is more of a depressurizer than an equalizer. Advised to keep it open.

I usually don't measure anything inserted deeper than where the reference plane is located, but it will be an exception this time only.

ON SECOND THOUGHT #4: Once VC02 has been inserted past the reference plane, which is approximately located at the 2nd-bend of the external ear canal, VC02's frequency response evens out even further. However, it is critical for users to note that this insertion depth is deeper than that of Etymotic Research ER-4.

ON SECOND THOUGHT #5: This is the tag attached on VC02's cable, recommending 100 hours of break-in to users. It seems the manufacturer, Vsonic, openly & blatantly advocates the effect of break-in. Thus, Vsonic VC02 will be broken-in for 100 hrs at 100 dB SPL @ 1 kHz, and its pre/post results will be comprehensively analyzed.

Sony MH1C: Is a bad excuse better than none? [UPDATED]

[UPDATED ON 2/2/2012] A brand-new analysis has been carried out, and it simply settles everything. I will still keep this article for reference purposes. Please keep in mind all the information below have been fully revised!

Sennheiser HD800 part3: In-depth analysis #2

Previously, by fully analyzing the European patent DE102007005620, a theoretical concept of Sennheiser HD800's driver has been discussed. The driver is in an annular configuration, metal coated with capricious crassitude, acoustically vented in the rear, and enclosed in a non-reflective baffle. By utilizing all the above, high-frequency phase intereference, oscillation, distortion are minimized, while the reflection-related contortion can be effectively controlled. The general analysis surely confirms Sennheiser's claim, but one factor still remains equivocal: a peak at 5~6 kHz. According to a Head-Fi user, arnaud, his FEM simulation indicates that HD800's asymmetrical geometry is the cause of such anomaly.

Above polar pattern represents HD800's positioning pattern on my dummy-head, EURI. It is clear that the peak is present at all positions except at the front(90°), where the peak is slighly reduced.

( v / frequency ) = wavelength for fundamental modes

(340,000 mm/s) / 5,000 Hz = 68 mm

(340,000 mm/s) / 6,000 Hz = 57 mm

And considering the role of a quarter-wavelengh resonance, the reflection has to be occuring approximately 14 mm to 17 mm away from the driver in the housing. However, even if the housing is acousically treated with modification, the peak is more or less still present. Then what is the ultimate cause of the peak? Is it possible that the source is from the center hole?

A ring radiator transducer simulated

In order to figure out the effect of a center hole, a Kobitone 25CE500-RO driver has been used for experiment. Its datasheet is here.

Frequency response

A calibrated free-field microphone placed at 10 mm distance

First of all, as a center hole is created in the middle of the diaphragm, the motion break-up of the cone is gone. And the consequential frequency response change is as seen above. The bandwidth remains unaltered, but the high frequency resonances are shifted downwards. And surely enough, a familiar pattern emerges:

Relative difference with the stock as a reference

Impedance

The mechanical resonance has been increased by a great degree with an upward shift.

Thiele-Small parameters

		Stock	Radiator
DC R	Res (Ω)	550	550
Free Air Resonance	Fs (Hz)	121.01	139.4
L of coil	Le (uH)	2342.59	2255.72
L due to ind coupl of eddy current	L2 (uH)	3427.82	1952.33
R due to eddy currents	R2 (Ω)	44.08	52.11
Total Q	Qts	0.62	1.09
Electrical Q	Qes	2.19	1.89
Mechanical Q	Qms	0.87	2.58
Effective mass	Mms (g)	0.18	0.12
Mechanical R of driver loss	Rms (kg/s)	0.16	0.04
Compliance	Cms (m/N)	0.0096	0.011
Air volume equivalent	Vas (liters)	2.13	2.45
Effective area of cone	Sd (cm^2)	12.57	12.57
Force factor	Bl (Tm)	5.76	5.33
Reference efficiency	ETA (%)	0.17	0.34
SPL @ 1 m for 2.83 V	Lp(2.83V/1m) (dB)	66.1	69.27

According to Dr.Wolfgang Klippel, mechancal & acoustic losses are dominant source for microspeakers's nonlinear distortion. Thus, the most importance parameters in this analysis are Bl, Cms, and Rms. Compliance of the suspension and mechanical resistance have definitely improved, while the force factor slightly decreased.

Total harmonic distortion

Interestingly, while the overall harmonic distortion figures are pumped up, distortions at 3~4 kHz, at which the human hearing is the most sensitive, have been reduced.

Directivity index

0-degree measurement data as reference

As a ring driver radiates soundwaves more uniformly than a dome, although its perceptual effect is not subjectively evaluated yet, it can be expected that the spatial distribution of a ring driver should be more like that of a electrostatic driver.


The role of a concha

Apart from electroacoustic properties of the driver, it is also vital to look into the physioacoustic properties of our body parts. Naturally, a concha boosts input signal up to 10 dB at 5~6 kHz due to Helmholtz resonance of its own cavity volume.

And of course, if the input signal has a peak at such frequency range, the result can be expected to be even more amplified, just like the case of HD800.


Conclusion

By simulating a ring radiator driver by making a hole in the middle of a Kobitone unit, its electroacoustic effect has been predicted. Thus, it can be assumed that the peak at 5~6 kHz of Sennheiser HD800 is deriving from these causes:

1. Resonant housing
2. Resonance shift due to center hole
3. Concha resonance

While the overall sound signature can be altered with modifications, it is virtually impossible to totally eliminate the peak.

(The deviation between Sennheiser's official frequency response data and conventional HATS measurement data shall be discussed in the next entry, as Gennadiy is expecting to receive the frequency response chart from Sennheiser soon)

HiFiMAN RE-ZERO

HiFiMAN RE-ZERO is a limited version of RE-0. Compared to the original, RE-ZERO comes with a silver voicecoil and 16-Ohm impedance value. It can be balanced-driven as well.

PRO: Overall very linear. Distortion-free. Due to its linear impedance characteristic, it is almost impossible to alter the sound of RE-0 electrically.

CON: Slightly more peaky than the original version. The residual resonance at 2 kHz may or may not be of a problem, depending on how users take it.

ON SECOND THOUGHT #1: Just like the original version, RE-ZERO is also enginneered to be shallow-inserted. Once it is inserted too deep(at the reference plane) or too shallow(just blocking the ear canal), the response above 10 kHz loses its linearity.

ON SECOND THOUGHT #2: The effect of acoustic damping of RE-ZERO is same as the original. Refer to RE-0.

ON SECOND THOUGHT #3: Stock single sleeves yield the most linear response.

ON SECOND THOUGHT #4: Just like the original version, the left vent works as a depressurizer, while the other one works as an equalizer.

Ultimate Ears UE700 rev.1

This UE700 is made before Logitech revised it to rev.2.

PRO: Well-extended low frequency decay. Reverberant characteristic is appropriate, resembling that of a RRF room condition until 6~7 ms.

CON: Slight channel mismatch in the lower frequency, inverted polarity. The high frequency peak can be bothersome to some users.

ON SECOND THOUGHT #1: UE700 rev.1 is designed to be inserted slightly shallow. The deeper the insertion depth, the less emphasis on the peak. Fortunately, the form factor of UE700 rev.1 is extremely small.

ON SECOND THOUGHT #2: As expected with Knowles TWFK driver, UE700 rev.1 can become a lot more linear with the help of a serial resistance. A 33-Ohm resistor will do the trick, but higher value further tames the peak.

ON SECOND THOUGHT #3: This is the flattest TWFK-equipped IEM I've measured, on par with UE900 modified. I hope I could get a hold on to a pair, so that I can do some modification!

HiFiMAN RE-0

I thank my dear friend, Bryan, aka Gnarlsagan, for providing me with an opportunity to analyze this nice piece of IEM.

Originally priced at $79, HiFiMAN RE-0 is a budget IEM, known to outperform some of the top-notch IEMs in the market. While other high-end products of the brand, such as RE-262 and RE-272, are finely engineered to meet the general criteria of Hi-Fidelity in headphones, it is considered that RE-0 also comes very close.

PRO: Overall very linear. Distortion-free. Left & right channel matching is excellent. Due to its linear, yet high impedance value, it is almost impossible to alter the sound of RE-0 electrically.

CON: Slightly peaky in the upper frequency range due to resonances.

ON SECOND THOUGHT #1: Surprisingly, RE-0 is precisely engineered to be shallow-inserted. Once it is inserted too deep(at the reference plane) or too shallow(just blocking the ear canal), the response above 10 kHz loses linearity.

ON SECOND THOUGHT #2: Due to resonant peaks in the high frequency range, it is a good idea to apply multiple layers of the stock damper, preferably more than two plies. It is, however, important to note that the treble response drops gradually as the number of dampers increases.

ON SECOND THOUGHT #3: Stock single sleeves yield the most linear response.

ON SECOND THOUGHT #4: There are two acoustic vents on RE-0. Considering the left channel unit only, while the vent on the left only works as a depressurizer, the right one equalizes the mid frequency range up to 4 dB. I would only use the right one if I were to make a tonal change.