Recently, having received several new headphones (new to me actually) for testing and evaluation and possibly to keep permanently, I was surprised by the wide range of variation in sound between the different headphones.  I found it easy enough to adjust to strong bass -vs- weak bass as long as the bass was not boomy or bloated to a significant degree.  I also found it easy to adjust to differences in treble strength as long as there were no major annoyances such as strong sibilants and harsh tones.

Once I eliminated the non-contenders for whatever reasons, I was left with my own personal stable of headphones: Sennheiser HD-800, PX-100-II, PX-200-II; Beyerdynamic DT-48E, DT-1350, DTX-300p; and the B&W P5 (my second set, acquired today).  For me, the most intriguing of the new acquisitions was the Beyer DT-1350, which some people refer to as the replacement for the T50P model.

Having done an extensive search for reviews of the T50P and finding very little, other than general agreement that the T50P’s bass is very light, my own experience with the DT-1350 tells me that Beyer fixed that little problem.  On the other hand, the sound of the DT-1350 from the midrange on up appears to be a repeat of the T50P’s sound, which prompted my investigation here.

Using Tyll Hertsen’s (formerly of headphone and amp seller ‘Headroom’) frequency response charts, and my own experience in correlating dips in the upper mids to lower highs with perceived detail in that area (or lack thereof), I’ve made the following list of headphones and the amount of drop in decibels from the midrange around 400 hz to the area from about 4000 to 7000 hz where each headphone has its greatest drop (or “suckout” as it is sometimes referred to).  For reference, the Sennheiser HD-800 has a drop of about 8 db, the Audeze LCD-2 has a 17 db drop, and the new HiFiMan HE-500 has a drop of 12 db.

I’ve made two tables, the first of which is sorted my manufacturer and model.  The second table is sorted by the drop in number of decibels.  In the second table I’ve drawn lines between “groups” to represent what I think of as “bright”, “normal”, “recessed”, and “depressed”.

These tables were formatted as plain text using the Courier fixed font, in case the columns don’t line up when viewed with a proportionally-spaced font.
Mids to Mid-Highs Drop by Manufacturer
————————————–
Audeze LCD-2            -17.0 db
Audio-Technica M50      -19.0 db
B&W P5                  -12.0 db
Beyerdynamic DT-1350    -25.0 db
Beyerdynamic DT-48E     -22.5 db
Beyerdynamic T1         -19.0 db
Beyerdynamic T50P       -25.0 db
Beyerdynamic T5P        -20.0 db
Denon D-2000            -09.0 db
HiFiMan HE-500          -12.0 db
HiFiMan HE-6            -07.5 db
Monster Beats Pro       -19.0 db
Sennheiser HD-25-1      -16.0 db
Sennheiser HD-800       -08.0 db
Sennheiser PX-100-II    -17.0 db
Mids to Mid-Highs Drop by Decibels
———————————-
HiFiMan HE-6            -07.5 db
Sennheiser HD-800       -08.0 db
Denon D-2000            -09.0 db
——————————–
B&W P5                  -12.0 db
HiFiMan HE-500          -12.0 db
——————————–
Sennheiser HD-25-1      -16.0 db
Audeze LCD-2            -17.0 db
Sennheiser PX-100-II    -17.0 db
Audio-Technica M50      -19.0 db
Beyerdynamic T1         -19.0 db
Monster Beats Pro       -19.0 db
Beyerdynamic T5P        -20.0 db
——————————–
Beyerdynamic DT-48E     -22.5 db
Beyerdynamic DT-1350    -25.0 db
Beyerdynamic T50P       -25.0 db

In addition to the above tables, I also did another comparison – of two headphones only – the Beyer DT-48E and DT-1350, in an effort to quantify the total and average amount of deviation from the anchor point (400 hz), and to try to get a feel for the smoothness of those deviations (i.e. the amount of change from measure point to measure point, as opposed to the amount of difference from each measure point to the anchor point).

In the table below, you can see that the total deviations (in absolute values) of the DT-48E add up to 94, and the DT-1350’s to 149.  But those are the deviations from the anchor point of 400 hz.  When you see that the measure-to-measure changes for the DT-48E add up to 88.5, and only 60 for the DT-1350, it appears that the DT-1350’s overall response from 400 to 11000 hz is smoother, albeit consistently lower in frequency from the anchor point.

When you consider that all of the better headphones exhibit the dips and choppiness in mid-upper frequency response that is shown in the measurement graphs, then obviously the data in each of those graphs could be “normalized” to subtract out the differences from an idealized response as graphed.  So, while Tyll Hertsen’s measurement graphs are extremely important and useful to at least partially validate what you hear in a particular headphone, there are several questions I think could be answered if someone would perform additional statistics on the data in those graphs.

Q: Is 400 hz a valid anchor or pivot point for measuring response deviations, or should that be set differently for each individual headphone? 

For example, if I moved the anchor point of the DT-1350 to 2000 hz, the total deviations would be 85 instead of 149.  2000 hz is probably too high, but it illustrates the point.

Q: If the data were normalized by subtracting out the deviations of an idealized response, would that make the results clearer or more accurate? 

At this point I don’t know, but I did perform a short test using the Sennheiser HD-800’s graph as the “ideal” response curve.  In this case, the DT-1350’s response matched the HD-800’s more closely than the DT-48’s did, which suggests that the deviation results noted above may not be as good a test as the smoothness and normalized-response tests.

Q: Would each headphone’s bass response have a mitigating effect on the results of tests performed on the mid-to-high frequencies, especially if the bass response were stronger than the midrange or anchor point? 

I would think so, but I don’t know at this point how that would work.

Frequency     DT-48E                    DT-1350
———     ———————     ———————

              Deviation    Change       Deviation    Change
              From         From         From         From
              400 hz       Previous     400 hz       Previous
              ——–     ——–     ——–     ——–
  600 hz      -02.5 db     02.5 db      -01.0 db     01.0 db

  800 hz      +02.5 db     05.0 db      -03.0 db     02.0 db

 1000 hz      +05.5 db     03.0 db      -05.0 db     02.0 db

 2000 hz      +06.5 db     01.0 db      -08.0 db     03.0 db

 3000 hz      -02.5 db     09.0 db      -16.0 db     08.0 db

 4000 hz      -07.0 db     04.5 db      -15.0 db     01.0 db

 5000 hz      -13.5 db     06.5 db      -15.0 db     00.0 db

 6000 hz      -22.5 db     09.0 db      -18.0 db     03.0 db

 7000 hz      -07.5 db     15.0 db      -23.0 db     05.0 db

 8000 hz      +07.5 db     15.0 db      -13.0 db     10.0 db

 9000 hz      +11.5 db     04.0 db      -03.0 db     10.0 db

10000 hz      -02.5 db     14.0 db      -11.0 db     08.0 db

11000 hz      -02.5 db     00.0 db      -18.0 db     07.0 db
————————————————————

Total of
Deviations
(Abs. Value)   94.0 db                  149.0 db

Total of
Changes From
Prev. Value    88.5 db                   60.0 db

An Investigation of Headphone Brightness and Detail Relative to the Midrange
4 (80%) 1 vote