• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Bose QuietComfort 35 II Review (Noise Cancelling Headphone)

Robbo99999

Master Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
6,874
Likes
6,672
Location
UK
Minimal eq to get him where he wanted to be (v similar to his k371 profile) and sounded good. What more is there to his opinion?
K371 is easier to EQ than this headphone, it's less jagged in the treble, that's one of the reasons why I don't like the measurements of this headphone - enough said from my point of view, I've made all my points in my previous posts.
 

MayaTlab

Addicted to Fun and Learning
Joined
Aug 15, 2020
Messages
924
Likes
1,512
Nice write up, but "exact FR curve at your eardrum" We are not fully there yet, not willing to diminish how great and creative those engineers are.

As written : below 1khz or so. The feedback mechanism present in most ANC headphones won't work above that.
Below that frequency variation across listeners is moderate and the most signifiant variance happens at the lowest frequencies because of breach of seal - what the feedback mechanism in some ANC headphones seems quite capable to address. It's only above 1khz or so that variation across listeners starts to become unmanageable because of anatomical variations / position :
1449446210135.png

https://www.aes.org/e-lib/browse.cfm?elib=16877

The next frontier for headphones design is indeed to pass the 1khz barrier and deliver an exact, constant and custom / tailored FR curve at the listener's eardrum above that frequency. As you said we're not there yet but I was only commenting on what happens below 1khz (and more specifically asking if Amir had an idea whether or not the QC35II was able to maintain a constant FR curve regardless of seal even when ANC is switched off).
 
Last edited:

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,941
Location
Michigan
Thanks for the review Amir. I have been trying some travel headphones and I am interested in what performs well.
 

Jimbob54

Grand Contributor
Forum Donor
Joined
Oct 25, 2019
Messages
11,066
Likes
14,697
K371 is easier to EQ than this headphone, it's less jagged in the treble, that's one of the reasons why I don't like the measurements of this headphone - enough said from my point of view, I've made all my points in my previous posts.

You wouldn't want to sleep on either ;-)

BckogGzG0kJSfxB_M5aExnG3ehu6tSpJ4_-VbRIzBaAPyJe0KoCy49ZM099aQiQZF36SeYNVXms4-unPzvzNKEU6-tXb5EbgrbnzDquqbww6K7wRIgEAQCVSq_FWagrHK7dQuwnR


T4L_5XztuQOYe-vQmzppcx_Y7n8a5ZCeej7C64nZnbs_dPi8MjLsFietUmdQugR0lQ2UsXrIcHSTBEoE6XZR_zkYx_3tv-W6UE-GK55ZAlvbH6UC8URHzQTjcpdqP04qLJeDpi2y
 

Attachments

  • 1613928953118.png
    1613928953118.png
    81.1 KB · Views: 89

mt196

Active Member
Joined
Apr 7, 2019
Messages
229
Likes
156
Location
Rome, Italy
I understand that. Which curve would you use as a reference in place of Harman 2018?
I won't use a reference curve, I will just measure it then anybody can judge for their likings. Putting votes/panthers just based on that curve is completely stupid and that is the biggest problem of headphones reviews here. It's way easier to judge dac or amp with measurements than something so subjective as headphones
 

Robbo99999

Master Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
6,874
Likes
6,672
Location
UK
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,386
Location
Seattle Area
Putting votes/panthers just based on that curve is completely stupid and that is the biggest problem of headphones reviews here.
The panther rating is my sum total *judgement* of a headphone. It is based on measurement, listening tests, and equalization of the same. For you to think this is stupid means you don't believe in any of these factors. That, would be stupid. :)
 

infinitesymphony

Major Contributor
Joined
Nov 21, 2018
Messages
1,072
Likes
1,806
K371 is easier to EQ than this headphone, it's less jagged in the treble, that's one of the reasons why I don't like the measurements of this headphone - enough said from my point of view, I've made all my points in my previous posts.
You wouldn't want to sleep on either ;-)

BckogGzG0kJSfxB_M5aExnG3ehu6tSpJ4_-VbRIzBaAPyJe0KoCy49ZM099aQiQZF36SeYNVXms4-unPzvzNKEU6-tXb5EbgrbnzDquqbww6K7wRIgEAQCVSq_FWagrHK7dQuwnR


T4L_5XztuQOYe-vQmzppcx_Y7n8a5ZCeej7C64nZnbs_dPi8MjLsFietUmdQugR0lQ2UsXrIcHSTBEoE6XZR_zkYx_3tv-W6UE-GK55ZAlvbH6UC8URHzQTjcpdqP04qLJeDpi2y
K371 is going to have more bass tilt and be more forgiving of some tracks with that sensitivity range cut (or simply sound dull), QC35 II might have a slight comb filtering effect in the treble that may or may not work for certain music. But overall, these headphones aren't so far off from each other.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,386
Location
Seattle Area
I just say: use your ears, graphs are useful but as a plus, moreover if we consider that everyone's hearing and perception are very different.
If that is true, that we are all "very different" when it comes to perception, then you sharing yours is of zero value to us. It is a random data point according to you so can't be used for any purpose. So best not post it over and over again.

On my side, I have compared countless headphone measurements to listening tests and importantly, corrections of what is shown in measurements. I have built up quite a bit of confidence in value of measurements in predicting a) preference and b) what would make the sound of a headphone more correct and preferred. If you don't agree with this, then don't read or comment on these reviews. They are not for people who believe in randomness of everyone's perception.
 

Dogen

Senior Member
Forum Donor
Joined
Aug 31, 2018
Messages
360
Likes
612
Location
Durham, NC USA
Bose knows how to achieve their design goals. We may not agree with the goals, but they’ve got a lot of skill at their disposal. Here, they do a great job.
 

Not Daijoubu

Member
Joined
Feb 2, 2021
Messages
24
Likes
22
So these or Sennheiser HD650? :p

Surprisingly good measurements, though I personally feel our current repertoire of measurable aspects do not tell the whole picture of whether a headphone is "good" or "bad."

I'd be interested in a study in which two different drivers using the same enclosure/headband and FR (with EQ) can be differentiated in a blind test. Something like Koss Porta Pro vs. Yuin G1A on PP headband, Sennheiser HD559 vs HD560S, HD58X vs HD650, DCA Ether 2 vs Voce etc. Is there not a way to quantitatively measure things like "microdetail" and "dynamics" that audiophiles claim to exist? In such instances, I wonder if a listener can tell differences with ABX after familiarizing themselves with the headphones.

Recently I watched this Schiit interview in which Jason remarks he and others were able to differentiate the Magni 3+ and Heresy under blind testing conditions. Interestingly, he also says it was much harder to differentiate the Heresy from other integrated designs. As for how much is believable is debatable, but I am inclined to take his word since he states it was under blind, not just sighted conditions, and they have been testing their designs far longer than audiophile would probably demo the devices. Although this does not make any difference to me, who cannot hear variations, it does make me wonder if there is justification to go for higher model headphones as opposed to EQ-ing lower end ones with good measurements.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,386
Location
Seattle Area
As for how much is believable is debatable, but I am inclined to take his word since he states it was under blind, not just sighted conditions, and they have been testing their designs far longer than audiophile would probably demo the devices.
A test being blind is insufficient to draw any conclusions. The test needs to be repeated a number of times to reduce the chances that the person just got lucky. If I flipped a coin and predicted tails and that is what became, would you believe that I can always predict a coin toss?

I have run blind ABX tests where I got a number right only to get lost from there on. Here is an example:

foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/07/19 07:09:23

File A: C:\Users\Amir\Music\Arny's Generational Loss\sb20x_original.wav
File B: C:\Users\Amir\Music\Arny's Generational Loss\sb20x_pass1f.wav

07:09:23 : Test started.
07:09:40 : 01/01 50.0%
07:09:54 : 01/02 75.0%
07:10:07 : 01/03 87.5%
07:10:26 : 02/04 68.8%
07:10:36 : 03/05 50.0%
07:10:51 : 04/06 34.4%
07:11:03 : 05/07 22.7%

07:11:14 : 05/08 36.3%
07:11:29 : 05/09 50.0%
07:11:44 : 06/10 37.7%
07:12:01 : 07/11 27.4%
07:12:16 : 08/12 19.4%
07:12:37 : 09/13 13.3%
07:13:23 : 09/14 21.2%
07:13:34 : 10/15 15.1%
07:13:47 : 11/16 10.5%
07:14:00 : 12/17 7.2%
07:14:13 : 13/18 4.8%
07:14:23 : 13/19 8.4%
07:14:37 : 14/20 5.8%
07:15:09 : 14/21 9.5%
07:15:22 : 15/22 6.7%
07:15:39 : 16/23 4.7%
07:15:59 : 16/24 7.6%
07:16:12 : 16/25 11.5%
07:16:34 : 17/26 8.4%
07:16:48 : 18/27 6.1%
07:17:09 : 18/28 9.2%
07:17:53 : 18/29 13.2%
07:18:11 : 19/30 10.0%
07:18:34 : 19/31 14.1%
07:19:13 : 19/32 18.9%
07:19:18 : Test finished.

----------
Total: 19/32 (18.9%)

As you see, I got 19 out of 32 right at the end which is not a valid result. But the bolded section appears to make a case that I was guessing correctly four times in a row.

You have to get results like this to be definitive:

foo_abx 1.3.4 report
foobar2000 v1.3.2
2017/11/26 10:50:22

File A: C:\Users\Amir\Documents\Test Music\Headfi RR Samples\07094.wav
File B: C:\Users\Amir\Documents\Test Music\Headfi RR Samples\27776.wav

10:50:22 : Test started.
10:51:03 : 01/01 50.0%
10:51:17 : 02/02 25.0%
10:51:25 : 03/03 12.5%
10:51:36 : 04/04 6.3%
10:51:42 : 05/05 3.1%
10:51:52 : 06/06 1.6%
10:52:00 : 07/07 0.8%
10:52:10 : 08/08 0.4%
10:52:17 : 09/09 0.2%
10:52:25 : 10/10 0.1%
10:52:36 : 11/11 0.0%
10:52:41 : Test finished.

----------
Total: 11/11 (0.0%)

Now probability of guessing is pretty much the same as zero.
 

majingotan

Major Contributor
Forum Donor
Joined
Feb 13, 2018
Messages
1,509
Likes
1,781
Location
Laguna, Philippines
Is there not a way to quantitatively measure things like "microdetail" and "dynamics" that audiophiles claim to exist?

I answered this question on a different post: https://www.audiosciencereview.com/...y-mdr-7506-review-headphone.19099/post-626877
THD and CSD/impulse/step plots are the answer for detail. High amounts of distortion at 1KHz - 4KHz can give a pleasing timbre and “texture” plus faux details as seen with HEDDPHONE.

BTW, a worse 2nd and 3rd harmonic plots on 1 KHz to 4 Khz can also indicate a presence of "microdetail" and tuning the FR to a small dip (instead of a rise to from 1 KHz to 4 KHz) can significantly increase "dynamics" as the bass and treble are now slightly more forward than the mids so when you crank the volume up to keep the mids volume high, you hear a larger presence of dynamics in the bass and treble region
 

Phorize

Major Contributor
Forum Donor
Joined
Apr 26, 2019
Messages
1,533
Likes
2,060
Location
U.K
Well the audiophile in me so wanted to hate these, but good engineering triumphs over prejudice it seems, shame on me.
 

Not Daijoubu

Member
Joined
Feb 2, 2021
Messages
24
Likes
22
A test being blind is insufficient to draw any conclusions. The test needs to be repeated a number of times to reduce the chances that the person just got lucky.

Obviously it needs to be repeated over and over so that the results are undoubtedly not just due to random variation. Supposedly, a good sample size is how Harman reached their conclusions on the target curve after all. I'm not a complete neophyte when it comes to statistics and proper experimentation.

And regarding you not being able to find differences with ABX, it does not discount some people may be able to. Personally I can't either, but I'm not gonna try to dispute the results of an ABX done multiple times with multiple participants. From Jason's words in the interview, it seems they did take blind testing seriously, though I obviously cannot say whether they for sure did it properly without actually knowing exactly how much they tested and how they tested.

Really, I'm just curious if a hypothetical blind test of different headphones, hypothetical since it would be expensive, impractical, and too niche to implement an actual one that is authoritative, would produce interesting results in terms of preference or sound differentiation.
 
Last edited:
Top Bottom