Search This Blog

Monday, June 17, 2013

SONY PFR-V1 part3: In-depth analysis

Disclaimer: Thanx to my friend, vAsurada, I am having another shot at this funky headphone. 

Continued from Part 2.


Sony PFR-V1 features several brilliant innovations:

1. Free-field referenced localization with separated treble / bass transduction


2. Elongated bass port (extended bass reflex duct)


3. Minimized cabinet resonance


4. High-quality materials


While considering the others as commercially driven mumbo-jumbos, the first two factors must be considered in a technical manner, since they are the fundamental working principles of PFR-V1.


On the Free-field referenced localization 

According to Sony, PFR-V1 intended for professional monitoring. As the localization characteristic of an usual monitoring setup is of a free-field, it can be assumed that is what the manufacturer intended with this headphone.
The yellow curve is the target insertion gain

In order to identify PFR-V1's correct localization characteristic, free-field measurements have been carried out at 10 mm / 50 mm distance, using a precisely calibrated microphone. Whichever an orthotelephonic gain of the dummyhead measurement becomes close to the free-field data, it is then the headphone's spectral property. In addition, PFR-V1 has been measured at various angles in azimuth using EURI at 50 mm. Should V1's in-situ eardrum response match any of the angle, then the correlating HRTF is indeed that of the headphone.

The white curve is PFR-V1's HRTF characteristic measured at the eardrum, with the driver's characteristic compensated

And the HRTF somewhere in between 45° ~ 60° roughly correspond the in-situ response of PFR-V1, but there is still no perfect match. Is it of a random-incidence then? But the sound is definitely coming from the front. First, let us turn to what SONY has to say. The developers of PFR-V1, Yamagishi & Yamaguchi, tell an intriguing story behind the development of the headphone:

"Why such a bizarre design: In fact, the idea of placing speakers in front of ears has been around for 20 years. Although I have been involved with designing headphones/speakers only for about 7 years, even at the time when I first joined SONY, there was already an ongoing-research regarding the subjective tonal quality enhancement of a forward-placed speaker in proximity. Why was the technology not commercialized? In order to reproduce high-quality bass, the speaker had to get larger & heavier. The positioning of the speaker could not be balanced well on the head, because the unit itself was simply too bulky."

"I have developed many different speakers & headphones. And about 2 years ago, a brilliant idea suddenly struck my mind: 'If I port a duct to SRS-AX10, will there be more bass?' So I brought a plastic straw from a restaurant's vending machine, and equipped it onto the speaker. And then VOILA! That's the moment which jump-started the development of PFR-V1."

Prior to Theile's breakthrough with the diffuse-field target in 1986, an in-front localization has been the target reference model for headphones. As outlined in DIN 45500-10, headphones back then are calibrated in an anechoic environment with a mono loudspeaker configuration as a reference. Once the calibrated responses are averaged among listeners, they are considered to be flat in reference to the free-field.


Typical free-field referenced headphones


According to Poldy, a free-field referenced headphone should have a peak at 2-3 kHz and minima near 1.2 kHz and 8-10 kHz, with the peak deriving from a quarter-wavelength resonance of an ear canal, while the minima coming from a shoulder reflections & a half-wavelength resonance of a pinna.

A typical loudspeaker configuration                          from Shaw's 1985 research

However, this reference model brings up a fundamental problem with the accuracy in localization. Not even mentioning the absence of mono/binaural cues in headphones vital for creating a directional localization, according to Møller:
"Evidently anatomical differences may have a large effect on the HRTF for a single direction, but when these are averaged over the sphere, the effects of interindividual differences are reduced drastically."
Consequently, although the headphone was free-field equalized, listeners can not feel such sensation. The averaged calibration target is of no use due to much differences from person to person. Rather, listeners are experiencing coloration in tonality, or linear distortion, as Theile put it.

Even the headphone designers at that time thought something was definitely wrong with the free-field. The founder of Fontek Research, 丹羽久雄-san, who was a former engineer at STAX, also noted that the in-front placement is not a working solution: 



丹羽-san's note deliberately disregarding STAX sigma series

Then how do you make sense of V1's frequency response? In his article from 1995, Møller presented something called a 'combined-field' reference target, which is a compromise between a free-field & a diffuse-field. It would definitely make sense, if V1's directional characteristic is slightly mixed up with that of a diffuse-field reference.


As the insertion gain at 67.5°, which is an intermediate angle between 45° and 60°, definitely seems to be an adequate angle of incidence, PFR-V1 indeed shows a slight sign of diffusion in spectral balance, as asserted by Møller. This will effectively help eliminate any of the interpersonal deviation caused by the variance in pinna shape. Thus, it can be concluded that the PFR-V1's localization property is very close to ±67.5° incidence from both azimuth & elevation. While PFR-V1 is somewhat diffuse-field oriented, its localization property is still well-obliquely oriented.


On the bass extension

Although the elongated bass duct does what it's intended to do, the amount of bass is just too little, even when the headphone is properly equipped.

After its brief introduction to the market in 2008, a Japanese blogger finds out SONY has only partially utilized the bass duct. With the removal of a single screw, the blogger claims the lower frequency extends dramatically.

The frequency range below 1 kHz get boosted quite a bit, while evening out the peaks in the high frequency range along. The modification is definitely worth a shot, but the damping becomes overly slow.


Conclusion

Although there are some weaknesses, including the overly bright driver, lack of bass, and unusual placement angle of 67.5° in azimuth, Sony is innovative & daring enough to make such a cool-looking headphone, with the state-of-the-art localization tuning technique. It is truly one of a kind product, and even compared to much higher priced headphones, such as STAX sigma and AKG K-1000, PFR-V1 still finds a glory in its technical uniqueness.


References

C. A. Poldy, Chapt. 12 in Loudspeaker and headphone handbook, Borwick, J., ed., Oxford: Reed, 1994.

D. Gibbon, R. Moore, and R. Winski, Handbook of Standards and Resources for Spoken Language Systems. Berlin: Mouton de Gruyter, 1997.

D. Hammershøi and H. Møller, "Determination of Noise Emission from Sound Sources Close to the Ears," Acta Acustica, Vol.94 No.1 (January 2008).

DIN 45500-10, Heimstudio-Technik (Hi-Fi); Mindestanforderungen an dynamische Kopfhörer nach dem Tauchspulenprinzip, Beuth Verlag, Berlin, Cologne, (1975).

E.A.G. Shaw and M.M. Vaillancourt, "Transformation of Sound-Pressure Level from the Free Field to the Eardrum Presented in Numerical Form," Journal of the Acoustical Society of America, Vol.78 No.3 (September 1985).

H. Møller et al, "Design Criteria for Headphones," Journal of the Audio Engineering Society, Vol.43 No.4 (April 1995).

No comments:

Post a Comment