Sennheiser HD650 Review (Headphone)

[David Ellis]

I thought this might be helpful. I took some measurements of the Sennheiser 650 headphones my son gave me for Christmas. I took these measurements using a UMIK 1 with a 90 degree calibration file. I placed the microphone at a few locations around my ear canal and placed the headphones over my ear and the microphone. This is the first time I have done this and it seemed a reasonable approach. The above graphs are from my right ear and the left ear was about the same - incredibly flat overall, with a mild hump around 5-6khz then a dip and some mess in the graph after a dip at 8khz. There are likely plausible explanations for the ripples, but I am darn happy with how they sound - MUCH better than my Sennheiser 598 headphones. Strangely, the 650 headphones sound bass heavy, but the graph doesn't necessarily convey this. I also realize my graphs look considerably different than the initial graphs posted by Amir. Perhaps this resulted from differences in our measurement methods. Or perhaps my new headphones actually measure different than Amir's "older" headphones.

 

[staticV3]

View attachment 417524
I thought this might be helpful. I took some measurements of the Sennheiser 650 headphones my son gave me for Christmas. I took these measurements using a UMIK 1 with a 90 degree calibration file. I placed the microphone at a few locations around my ear canal and placed the headphones over my ear and the microphone. This is the first time I have done this and it seemed a reasonable approach. The above graphs are from my right ear and the left ear was about the same - incredibly flat overall, with a mild hump around 5-6khz then a dip and some mess in the graph after a dip at 8khz. There are likely plausible explanations for the ripples, but I am darn happy with how they sound - MUCH better than my Sennheiser 598 headphones. Strangely, the 650 headphones sound bass heavy, but the graph doesn't necessarily convey this. I also realize my graphs look considerably different than the initial graphs posted by Amir. Perhaps this resulted from differences in our measurement methods. Or perhaps my new headphones actually measure different than Amir's "older" headphones.

When sharing graphs, make sure that you click on Limits, then Fit to data first

Though I'm afraid that this method of measuring headphone frequency response really isn't good for much of anything..

 

[isostasy]

View attachment 417524
I thought this might be helpful. I took some measurements of the Sennheiser 650 headphones my son gave me for Christmas. I took these measurements using a UMIK 1 with a 90 degree calibration file. I placed the microphone at a few locations around my ear canal and placed the headphones over my ear and the microphone. This is the first time I have done this and it seemed a reasonable approach. The above graphs are from my right ear and the left ear was about the same - incredibly flat overall, with a mild hump around 5-6khz then a dip and some mess in the graph after a dip at 8khz. There are likely plausible explanations for the ripples, but I am darn happy with how they sound - MUCH better than my Sennheiser 598 headphones. Strangely, the 650 headphones sound bass heavy, but the graph doesn't necessarily convey this. I also realize my graphs look considerably different than the initial graphs posted by Amir. Perhaps this resulted from differences in our measurement methods. Or perhaps my new headphones actually measure different than Amir's "older" headphones.

This isn't a bad approach, you're sort of measuring the response before it reaches your ear. The main downside is it's impossible to do without breaking the seal, this is probably why you don't see the bass emphasis you'd expect. As the microphone is going in at a right angle you're also filling up a lot of the space with the microphone body.

Amir's rig has a calibrated microphone in a tube with some surrounding chambers (3 I think) of differing volumes to emulate an ear canal. Outside of this is a flexible "ear". So when he measures headphones the frequency response is modified by the "ear" then further by the impedance of this "ear canal" before it reaches the microphone which is like the "ear drum". Although you've put the headphones on your head, what you're essentially doing is measuring the response of the driver before it reaches your ear.

Therefore constructing a flat-plate coupler would be much more convenient to use and give more repeatable results. At bare minimum all you need is a roughly head-width cardboard box with a hole for the microphone to fit through one end and come out the other side. Then you end up with a measurement like this:

The EQ I made by cross referencing with GRAS measurements and my own listening. You can see the bass emphasis centered around 100-150Hz which is probably why you hear them as bass heavy. The small correction around 2kHz isn't really necessary, it's only really noticeable as a little more clarity to distorted guitar parts in very dense songs like metal.

N.B. the suggestion to "fit to data" is reasonable but you will end up seeing data below 20Hz if you do this. I suggest limits 20Hz-20kHz then manually set the y-axis so the measurement isn't compressed to look flatter than it really is.

 

[patoulol]

I just came across the great site of @solderdude

I loved as much as I hated my HD650 that I ended up selling for two reasons:

-keeps the ears too warm
-veiled in the high end of the spectrum

It dated from 2011, which probably explains the veil!

 

[Mr Swing King]

I never quite understood the ‘veiled’ comment with regards to the classic Sennheisers (HD580/600/650/6XX). Sure the 650 has a smidgen of warmth, but it still sounds extremely natural.
The only way I can wrap myself around this conundrum is if the user comes from brighter headphones - cans with more ‘sparkle’ up top. Then again most folks I know who feel the Sennies are veiled sounding usually prefer Beyerdynamics, AKGs and Grados. If your brain is used to those types of sound signatures, an HD650 is going to sound dark or even muddy.

 

[patoulol]

I never quite understood the ‘veiled’ comment with regards to the classic Sennheisers (HD580/600/650/6XX). Sure the 650 has a smidgen of warmth, but it still sounds extremely natural.
The only way I can wrap myself around this conundrum is if the user comes from brighter headphones - cans with more ‘sparkle’ up top. Then again most folks I know who feel the Sennies are veiled sounding usually prefer Beyerdynamics, AKGs and Grados. If your brain is used to those types of sound signatures, an HD650 is going to sound dark or even muddy.

The veiled HD650 (a 2014 silverscreen by the way) clearly has 3 to 4 dB less output compared to the quite new HD650 (2017).
Even the older black screen HD650 (fitted with new pads) has about 2dB more presence & treble than the veiled HD650. Also the lows roll-off somewhat sooner and steeper.
This HD650 thus has audibly less output and sounds ‘darker’ than the HD650’s I heard and measured before.

HD650

back to Sennheiser back to measurements home NO SMOOTHING is applied to the shown plots. Most measurement sites have some smoothing applied which ‘irons flat’ sharp peaks and ‘wig…

diyaudioheaven.wordpress.com

 

[mike7877]

Esoteric info about the drivers in the HD650

- their maximum movement is about 2.6mm peak to peak!

For anyone who was curious, now you know! I'm approximately 97% confident that it's 2.6mm +- 0.1mm (and probably a little higher like 2.63 but don't quote me)

How'd I find this out? Well, first I used my eyes... wait, no... that was last -

Tanchjim Space is worth talking about, but not here very much, past it's very small and has a maximum output voltage of 4.0V RMS. I was powering my 650s with it, when, at 26Hz, 4.015V (or whatever 0dBfs is...) my HD650 bottomed!



It wasn't a fierce bottoming, like a car's shocks after 5 feet of air over the train crossing, no. It was like, soft. A marshmallow.

I further refined it, and the highest frequency 4.0V can make the HD650s bottom out at is 26Hz and 3.984V
As you go lower, it just gets easier. This rule, very very VERY oddly, continues down to the 3Hz I could measure to, getting slightly more aggressive over the span, but not much (considering that's basically two octaves...)



Also, from what I've been able to see, I can't say for sure for sure that it's definitely, 100%, the driver which is the cause

Due to visibility restraints *unless you fully disassemble the headphones (which I wasn't willing to do because I don't want things rattling prematurely if I can stop it), we don't know if the driver's limited mechanical movement is from being mounted within a finite distance between two other object - there's only the ability to see the middle 50% of the driver through the hold opening for it in the very centre!

 

[mike7877]

I never quite understood the ‘veiled’ comment with regards to the classic Sennheisers (HD580/600/650/6XX). Sure the 650 has a smidgen of warmth, but it still sounds extremely natural.
The only way I can wrap myself around this conundrum is if the user comes from brighter headphones - cans with more ‘sparkle’ up top. Then again most folks I know who feel the Sennies are veiled sounding usually prefer Beyerdynamics, AKGs and Grados. If your brain is used to those types of sound signatures, an HD650 is going to sound dark or even muddy.

They don't sound veiled. I actually tone their top end down a bit in the treble -

Highs (peaking filters)
3330Hz -2.20dB Q = 0.976 These two adjustments make the top end sound SO much more natural!!!
5270Hz -1.50dB Q = 3.9 Get rid f the extra sparkle that just become fatiguing after 15 minutes (to me, anyway - I'm used to a flat system)

Mids (peaking filter)
1143Hz -0.6dB Q = 2.3 ;;;;; this one fixes some voices coming across too nasaly

Bass (peaking filters{except 42/84})
190Hz -1.6dB Q = 0.85 ::::: this removes the slightly excess upper bass
84Hz + 3.4dB Q = 0.8 this is a low shelf filter
42Hz + 3.8dB Q = 0.8 this is a low shelf filter
20Hz + 1.7dB Q = 0.75 this is a normal peaking filter
The low shelf filters are there to extend the useful bass down to 20Hz. It's basically flat.


AFTER
A very polite way to adjust the overall presence of top end by EQ is with a high shelf filter with Q = 0.5 set at 2300Hz. Normally not touched, for bad recordings or when listening really loud and treble becomes too much to enjoy. Q = 0.5 doesn't introduce new ringing, so guilt free!

And the bass, of course bass! Bass the base - when you're listening quietly, you need more basse! When someone mixed the song too loudly, you need more bass! There are too many scenarios to list where you need more bass, so obviously there's something for bass.
Frequency: Between 75 and 85Hz, older rock can be up to 98 in some circumstances. Q = 0.85 works well here. I'd say, usually, you wouldn't need to add or remove more than 2-3dB except in rare cases. I haven't removed bass yet.

 

[mike7877]

View attachment 417524
I thought this might be helpful. I took some measurements of the Sennheiser 650 headphones my son gave me for Christmas. I took these measurements using a UMIK 1 with a 90 degree calibration file. I placed the microphone at a few locations around my ear canal and placed the headphones over my ear and the microphone. This is the first time I have done this and it seemed a reasonable approach. The above graphs are from my right ear and the left ear was about the same - incredibly flat overall, with a mild hump around 5-6khz then a dip and some mess in the graph after a dip at 8khz. There are likely plausible explanations for the ripples, but I am darn happy with how they sound - MUCH better than my Sennheiser 598 headphones. Strangely, the 650 headphones sound bass heavy, but the graph doesn't necessarily convey this. I also realize my graphs look considerably different than the initial graphs posted by Amir. Perhaps this resulted from differences in our measurement methods. Or perhaps my new headphones actually measure different than Amir's "older" headphones.

Just saw your post... WOW was I SPOT ON lmao!
Did I or did I not say I'm used to flat speakers lol

 

[David Ellis]

The main downside is it's impossible to do without breaking the seal, this is probably why you don't see the bass emphasis you'd expect.

Breaking the seal.... not audible for me. Hmmm,

Can you gents hear the impact of breaking the seal ?

First, I appreciate you gents taking the time to read my measurement and offer reflections.

I am not certain that breaking the seal with the microphone reduces the level of measurable bass. I am not sure of this, but I intentionally broke the seal with a couple fat sharpie pens (simulating microphone placement) and listened to Al Hirt, Night life, on the Sugar Lips album. There is a steady string bass throughout the song so this is very easy to track. The audible impact of the bass doesn't not seem to change with respect to seal integrity. Maybe my ears just can't perceive this.

Are you gentlemen able to perceive an audible difference in bass impact with the seal of your headphones broken mildly - with a sharpie pen / microphone poking out from the bottom of the ear / seal ?

Also, I am highly impressed with the sound quality of these headphones. They are wonderfully clean, clear, smooth and wonderfully balanced overall - albiet slightly heavy in the bass region. I can also add the bass region sounds like a straight response line - not humped. I can easily hear the full harmonic spectrum from String Bass and Cello.

I have listened to many "bumpy" bass speakers where it's very difficult to hear the difference between a real bass instrument and a keyboard because the fundamental frequency is elevated and the harmonics are depressed. The 650s do NOT suffer this problem make this distinction among bass instruments VERY easy.