Head-Fi.org › Forums › Equipment Forums › Sound Science › Skeptico Saloon: An Objectivist Joint
New Posts  All Forums:Forum Nav:

Skeptico Saloon: An Objectivist Joint - Page 40

post #586 of 915
Yeah I am aware of the impedance differences. I have a FiiO brand headphone replacement cable and I tried my best to volume-match it with the stock headphone cable using a single headphone amplifier. It's hard to do when the amplifier has a digitally-controlled volume and it's hard enough to do it with an analog one. Despite turning the volume louder on the stock cable, I'm pretty certain I could hear a difference with the FiiO one, which is made of the Pure Copper Ohno Continuous Cast instead of the stock headphone cable's run of the mill 99.999% oxygen free copper (probably).
post #587 of 915
Switching gears, I just did a quick Google Scholar search of "high resolution audio" and this was the first result.
http://www.aes.org/e-lib/browse.cfm?elib=14186

Which used this as one of their references:
http://www.extra.research.philips.com/hera/people/aarts/RMA_papers/aar07pu4.pdf
Woszczyk, W., Engel, J., Usher, J., Aarts, R., Reefman, D., June 2007. Which of the two digital audio systems best matches the quality of the analog system?. Audio Engineering Society 31st International Conference.
Quote:
(overview of conclusions)
In the last few years, a number of complex listening tests were performed at the NHK Science and Technical Research Laboratories in Tokyo, Japan [1,2] to identify the importance of very high frequencies (over 21 kHz) in the discrimination of high sampling rate recordings. These tests were performed using two-channel stereo recordings of many different program types some containing large magnitudes of high- frequency energy. In double-blind tests, casual and professional listeners could not reliably identify high-bandwidth and high-resolution (192 kHz 24bit versus 48 kHz 24bit) conditions.
...
(audio preparation)
The tests consisted of a real-time comparison of an audio scene captured in an anechoic chamber and presented live to each subject in surround sound over six loudspeakers in a listening room. The subjects were comparing live analog reference with two digitized presentations of the scene made with two separate analog-to-digital and digital-to-analog conversion systems (X, Y) working simultaneously: (X) with high, and (Y) with low sampling rates.
Two identical high quality 8-channel Analog-to-Digital-to-Analog Converters (ADAC) were used in addition to a high quality 6-channel analog monitor selector and level controller having four groups of six-channel inputs and one six-channel output. The converters each generated a 5 bit delta-sigma modulated signal sampled at 128xFs (128 times 44.1 kHz) from which 24 bit resolution signals were down-sampled to DXD format of 352.8 kHz (8xFs) and to 44.1 kHz (1Fs) by decimation.
...
(audio recording)
The audio scene was captured in the anechoic chamber using high quality microphones and microphone preamplifiers. Each amplified microphone signal was delivered to two electronically balanced and isolated outputs (XLR and DB25, see Appendix 1) of Grace Design Model m802, 8- channel preamplifier, and each output was connected to two loads in parallel to feed the following devices equally:
...
(test procedures)
Listeners compared two digital versions of the scene with the reference analog version, considered to be a “live reference”. Their task was to specify which of the two digital versions had the most identical sound to the analog reference. The listeners were not to express preference for liking any particular version, they were simply asked to decide which of the two digital representations sounded most like the specified analog reference.
...
(test conditions)
There were two separate tests conducted, each providing a different bandwidth condition. Test Condition C1 applied a 100 kHz audio bandwidth, whereas Test Condition C2 had a bandwidth cut-off at 20 kHz by virtue of microphones and loudspeakers used that limit audio bandwidth to 20 kHz. In condition C1, Sanken CO-100k microphones (flat to 100 kHz) were used to capture the sources, and ribbon super-tweeters to reproduce them. In condition C2, measurement microphones limited to 20 kHz, and standard studio loudspeakers were used, with super-tweeters not active.
...
(test subjects)
Ten listeners, male and female, age between 20 and 30 years, took part in the test on a voluntary basis. All expressed strong interest in the test outcome, and all had prior listening experience making evaluations of sound in high-resolution or low-bit-rate systems. Some subjects were accomplished musicians, or had musical and technical training, and all were on staff in the Acoustics and Signal Processing group at Philips Research.

Each listener specified which one of the two digital presentations sounded most like the analog reference A. Each time a decision was made, the test facilitator would change the assignment of the sampling rate to each of the converters, and would indicate to the subject using an intercom that the next sound setting was ready for the evaluation.
...
(listening evaluation)
Listeners were able to choose any one or any number of sonic elements included in the scene in order to make decisions about the matching of sound quality between the three transmission systems. This approach allowed each listener to develop a personal point of interest focusing attention towards specific items within the complex spatial soundscape. Each listener could set the loudness level according to a personal preference, and could listen as long as needed in order to make judgment with a degree of certainty. Listeners had to make a choice (forced choice) even if they were uncertain that the choice was correct or appropriate.

In this test, we have made a general assumption that the high-sampling version was closer in quality to the reference system than the low-sampling version, simply on the basis of it having 8 times greater number of data points in the conversion from analog, and without regard for the perceived difference between these digital formats.
...
(conclusions)
The results of this test indicate that listeners more often than not identify high-resolution audio as being similar in quality to the unprocessed analog audio. This conclusion, based on listening to the audio scene captured and reproduced with microphones and loudspeakers limited to 20 kHz bandwidth, indicates that high-sampling conversion system seems to be more transparent and provides a higher degree of fidelity to the analog reference.

Our listeners also reported that when they were listening to the wider bandwidth, up to 100 kHz, of analog audio converted to digital, they would choose the low-sampling rate of digital audio as sounding more like the analog transmission.

These results seem to indicate that the ultra high-frequency content may not be necessary to reproduce audio that sounds more transparent. Supersonic content may contain noise-like artefacts that interfere with the perceived transparency of audio. However, to achieve a higher degree of fidelity to the live analog reference, we need to convert audio using high sampling rate even when we do not use microphones and loudspeakers having bandwidth extended far beyond 20 kHz. Listeners judge high sampling conversion as sounding more like the analog reference when listening to standard audio bandwidth.

These results suggest that the archiving community should consider using high- sampling conversion to ensure transparency even if the recording is made with standard audio-bandwidth transducers, and when digitizing older recordings made with bandwidth-limited analog systems.



I am thoroughly confused by the last 3 paragraphs. So audio extending beyond 20 kHz sounded less like the analog version to the subjects, yet the paper suggests converting audio to the higher bandwidths.

What the heck? *scratches head*
post #588 of 915

Ultrasonics can cause amps and dacs to oscillate which creates distortion in the audible range, which is why 192khz might sound different.

post #589 of 915
Quote:
Originally Posted by JRG1990 View Post

Ultrasonics can cause amps and dacs to oscillate which creates distortion in the audible range, which is why 192khz might sound different.
Parts within the amps and DACs have a natural frequency in that range? o.0


Also, the lack of any scientific publishing regarding 24 vs 16-bit audio concerns me; likewise for 44.1+ kHz audio and why such sampling rates might matter in the audio reproduction range. Sure, 192 kHz sampling rates have the Nyquist frequency of 96 kHz, but no one here ever mentions that such sampling rates may be useful for digital filters and whatnot; people only say "oh 192 kHz / 2 is out of the audible frequency range so it's useless."
post #590 of 915
It is out of the audible frequency range so it's useless. smily_headphones1.gif
post #591 of 915

Super audible frequencies are indeed the only way to discriminate between redbook and high bitrate audio. That is one of the few differences between the two. But that isn't important because studies by the AES have shown that super audible frequencies are completely irrelevant when it comes to sound quality in music.

 

Human beings can perceive super audible frequencies as sound pressure, but not sound. Music is sound.

post #592 of 915

I had the opportunity to check my ears when it came to frequency response recently. I've tuned the response of my speaker system by ear using good recordings of acoustic instruments. Oppo asked me to be a beta tester with their PM-1 headphones, so I shifted to cans, downloaded an equalizer for my iMac and started in balancing the cans the same way I did my speakers. It was a lot easier, because the deviation was within 3dB on the beta versions of the headphones they sent me. I jotted down my settings, then invited an engineer friend over to run tone sweeps on them. He found the exact same deviations from flat that I had, and some narrow range dips I had missed because my equalizer wasn't sharp enough to catch them. It was a pleasant surprise to find that I was so close to correct using my admittedly low tech technique. Oppo took the notes from me and the other beta testers and applied it to their next version. I just got the final retail version and it is the most balanced set of headphones I've ever heard.

post #593 of 915
Quote:
Originally Posted by bigshot View Post
 

I had the opportunity to check my ears when it came to frequency response recently. I've tuned the response of my speaker system by ear using good recordings of acoustic instruments. Oppo asked me to be a beta tester with their PM-1 headphones, so I shifted to cans, downloaded an equalizer for my iMac and started in balancing the cans the same way I did my speakers. It was a lot easier, because the deviation was within 3dB on the beta versions of the headphones they sent me. I jotted down my settings, then invited an engineer friend over to run tone sweeps on them. He found the exact same deviations from flat that I had, and some narrow range dips I had missed because my equalizer wasn't sharp enough to catch them. It was a pleasant surprise to find that I was so close to correct using my admittedly low tech technique. Oppo took the notes from me and the other beta testers and applied it to their next version. I just got the final retail version and it is the most balanced set of headphones I've ever heard.

can you somehow compare them to HiFiMan or Audeze?

post #594 of 915
I don't have those. I can only compare them to my Sennheiser HD-590s, which they are in a totally different universe from, and my speaker system, which for two channel they sound just as good, maybe a tiny bit better. (The speakers win hands down in 5:1).
post #595 of 915
Boy, it's been a while since I was out in the bizarro world on headfi. I laid out a pretty detailed breakdown of the differences in the response curves of the beta headphones I evaluated... including one derived from very careful two octave test tone sweeps... and the headphone junkies are telling me that they can't get any concept of what headphones sound like from info on frequency response. One of them told me that two headphones with the same frequency response can sound completely different because of "speed and PRAT" and another tried to tell me that Beats had a relatively flat frequency response!

It's interesting how much time people spend talking about headphones without actually learning how they work. I don't think that many of these people know what balanced response sounds like. They only know a million different shades of colored.
post #596 of 915
Quote:
Originally Posted by bigshot View Post

Boy, it's been a while since I was out in the bizarro world on headfi. I laid out a pretty detailed breakdown of the differences in the response curves of the beta headphones I evaluated... including one derived from very careful two octave test tone sweeps... and the headphone junkies are telling me that they can't get any concept of what headphones sound like from info on frequency response. One of them told me that two headphones with the same frequency response can sound completely different because of "speed and PRAT" and another tried to tell me that Beats had a relatively flat frequency response!

It's interesting how much time people spend talking about headphones without actually learning how they work. I don't think that many of these people know what balanced response sounds like. They only know a million different shades of colored.


I for one am very appreciative that you ventured into the bizarre world, especially since you knew what to expect and went there anyway. You gave a thoroughly convincing description of what you found with the headphones, and you did an outstanding job of communicating your findings.  And yes, they have no idea of what balanced sound is, and they probably never will.

 

anyway, thank-you very much for braving those waters. Hic sunt dracones

post #597 of 915
The thing that I'm really happy about was that I could help Oppo refine their headphones. The response of the first beta they sent me was very good, but it was obviously skewed to what audiophiles think sounds good... upper mid and mid bass boost, high end rolloff, etc. I used my EQ to give them specific suggestions about corrections listing exact specs for frequency and volume, and damn if they didn't incorporate them perfectly. If you want to hear what my dream headphones would sound like, check out the Oppos.
post #598 of 915
Quote:
Originally Posted by bigshot View Post

Boy, it's been a while since I was out in the bizarro world on headfi. I laid out a pretty detailed breakdown of the differences in the response curves of the beta headphones I evaluated... including one derived from very careful two octave test tone sweeps... and the headphone junkies are telling me that they can't get any concept of what headphones sound like from info on frequency response. One of them told me that two headphones with the same frequency response can sound completely different because of "speed and PRAT" and another tried to tell me that Beats had a relatively flat frequency response!

It's interesting how much time people spend talking about headphones without actually learning how they work. I don't think that many of these people know what balanced response sounds like. They only know a million different shades of colored.

lol this is why I took a chance in asking here about a comparison first rather elsewhere. There will be a lot of chippers of varying knowledge of headphones and it becomes chaotic to navigate through the thread for info.

 

A few questions if you don't mind, what EQ were you using and why did you use that one specifically? Are they at all affordable?

 

Were you able to test the velour vs the leather pads?

 

and lastly, did the Oppo engineers mention anything about their design? I am rather bewildered as to why they chose the transducer to be confined and the opening like a typical dynamic driver. I was under the impression that the main thing that helps the sound of the planar magnetics to be like a 2 channel system is the planar wavefront as well as the greater flow of air.

post #599 of 915

Well, one thing that is true is that the sound of headphones (relative to speakers) depends on the head, ears, torso, etc. of the listener. i.e. the way real life and speakers sound is influenced by anatomical factors that vary from person to person, and some of these are bypassed or are different (never mind if the headphones fit onto the head differently) when using headphones. So to produce a natural / perceptually flat sound on headphones, you need a different tuning for each person, technically. Thus, having a comparison to other headphones can be useful, and there is some justification of being suspicious of "hey it sounds flat like calibrated speakers in room."

 

That's definitely not what most people are thinking or talking about, though.

post #600 of 915
I did rough checks using Boom for Mac, but when we did the tone sweeps, my sound engineer friend brought over his signal generator and pro grade digital equalizer. We checked each other at each step and were probably accurate to about 1dB.

I had two beta versions at once that Oppo said were the same sound, just different headbands. One had leatherette and the other had velour. They came out to pretty much the same response... within 1dB or so which is just variation between copies. The subjectivists in the test group swore the velour softened the sound, but we didn't find any difference at all sound wise. (I wouldn't imagine there would be with the drivers pointing straight down your ear canal.)

I'm afraid I'm not up on headphone design theory. There was a point early on where one of the beta testers suggested a modification to the damping and it was incorporated into the next beta, but I don't know what that was all about. The designer we were speaking with kept speaking way over my head. I was lost in some of those conversations. He was an expert in planar designs, having designed speakers with the same concept. I got the impression that he was working on his own theories based on trial and error and experimentation, not generally accepted thought on the subject.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Skeptico Saloon: An Objectivist Joint