Unlike linear frequency response errors, hearing non-linear distortion requires skill. Proper testing usually involves listening training and filtering out those who do not have critical listening skills.
The listeners were all employees of Harman International who were trained listeners with normal audiometric hearing. All listeners had successfully achieved level 7 or higher in the Harman How to Listen Training software.
And such distortion may very well be level dependent.
After equalization, all five headphones were measured at four different calibrated levels; 82, 88, 94, & 100 dB SPL.
The relative playback levels of the headphones were adjusted to the same level based on ITU-R 1770.2 level. All of the listening tests were conducted at an average playback level of approximately 80 dB (Bweighted, slow) for both listening test sessions.
An important feature of this test is that the playback level is lower than the level which was used in the recordings. The 90 dB recording level would have been uncomfortable and possibly unsafe for the listening subjects. Absolute ecological validity is thus not achievable in this case. If anything, this should result in an increase in the sensitivity of the subjects to audible distortion.
@GaryH, Harman listener training has nothing whatsoever with detecting non-linear distortions. It is completely about hearing frequency response errors.
— @Sean Olive 's blog.The current Harman listener training software “How to Listen” includes training tasks on spectral distortion as well as spatial, dynamic and various types of nonlinear distortions
As to listening levels, 100 dBSPL is too low for bass response. It is trivial for me to create audible distortion at some listening level in a headphone. "90" dBSPL *flat* response is indeed very loud. Not because of bass but higher frequencies. Real music has energy piled on at very low frequencies so it is not nearly that loud even at 114 dBSPL.
Reference level for recreating live venues calls for 120 dBSPL (actual measurements of multiple concerts at listening position).
As to listening levels, 100 dBSPL is too low for bass response. It is trivial for me to create audible distortion at some listening level in a headphone. "90" dBSPL *flat* response is indeed very loud. Not because of bass but higher frequencies. Real music has energy piled on at very low frequencies so it is not nearly that loud even at 114 dBSPL.
The TI-1a was created for the CA-1a, not SR-1a. It does include barrels for baffle compensation so can be used with the SR-1a. I have the same one. The TI-1b is the same internals but in a polycarbonate or plastic housing.Sorry, but no one noticed that Amir tried the ca1a with the old interface created for the sr1a? the correct one is TI 1b
Yes:Sure?
TI is a successor to RI (the resistor interface) and we won't be making those any more.
TI-1a is in a metal box (available now), TI-1b is in a molded box (available in 4 weeks). There is just a limited run of metal box TI's and we won't be making those any more.
I think that for a short period of time, they will overlap until all TI-1a metal boxes are sold out.
The difference between the TI-1a and TI-1b is the housing. The TI-1a is housed in a metal box, very similar to our original resistor Ribbon/Amp Interface. The TI-1b is electrically identical to the TI-1a (same circuit and components) but housed in a moulded box that echos CA-1a shell material.
They are. And in the majority of reviews of their competitors, in terms of price point anyway, they are often trounced for deviating so far from the Harman Curve, but get glowing reports on listening tests, once EQ is applied (like here).The Dan Clark Stealth and Expanse are well reviewed here.
well, the treble curve has a strange trend, as if the SR1's compensation filters were used
Anyone who spends more than $500 on headphones is special. Probably rich too, but absolutely special.Like their other here previously reviewed model looks and performs like amateur DIY...
Guess their target group must be very special.