MEASUREMENTS: Sennheiser HD650 (~2010-2012) and HD800 (~2010). A comparison of two "Reference Class" headphones, and a look/listen to the "Sennheiser veil / darkness". [Inexpensive 2TB SSD and inflation in consumer electronics...]

As you can see in the picture above, this is a bit of what the "box opening" looks like when you buy either one of these "Reference Class" Sennheiser headphones.

Much has already been said about the headphones being measured today. The Sennheiser HD650 was originally released back in 2003. Of interest historically, there was a fire in the Ireland factory back in 2004 so there probably are not many older pre-fire HD650's floating out there. Tyll Hertsens' article about the Sennheiser HD580/600/650 is a great summary and SolderDude has a ton of stuff on the HD650 on DIY-Audio-Heaven. SBAF has this "compendium" on the HD650 which I think covers everything... and more! Oh yeah, even NwAvGuy got one of these back in 2011.

These days, there's the Massdrop x Sennheiser HD6XX which would be the latest version of the HD650 discussed here, while the Sennheiser HD660S looks similar but has redesigned drivers.

Likewise, you already no doubt would have read about the Sennheiser HD800 over the years. These came out in 2009, sporting the large 56mm "ring radiator" driver (vs. 42mm transducer of the HD650). The HD800 has since been revised with the HD800S in 2016 (here's Tyll's take with mods discussions). Multiple reviews on Head-Fi. The HD800S sounds very similar to the HD800 to me.

For this post, let's look at measurements comparing the two and I'll put in my 2¢ on the sonic differences as I hear them.

The Sennheiser HD650 (built ~2010-2012)

My friend linnrd bought this headphone back in 2013 used and was able to lend them to me for a listen. Peeking behind the metal grille in the image above, we see that this is not the oldest version based on the color of the driver surrounding material which is silverish/white rather than the pre-2010 version that would have been all black (based on a comment here). So based on this, unless there is a more precise way to find manufacturing date using the serial number (1409XX), this specific headphone was "Made In Ireland" at some point likely between 2010-2012. (I've sent an E-mail to Sennheiser with the serial number, will update if they send me more info.)

Knowing the age of these headphones appears to be important because over time, reportedly there have been changes to the tonality of these headphones.

The 2-pin cable connectors to the ear pieces are easy to remove and replace. I'm somewhat envious of how inexpensive balanced cables are for these HD650 headphones as opposed to the significantly higher price for HD800's (even generics)!

They feel quite comfortable and light (260gm without cable) when worn but that's variable depending on one's head shape and size of course. Clamping force is stronger than the HD800. I measured these with the stock pads on, and while many years old I don't think the headphones have been used too much and the pads look to be in good shape, are clean, have maintained "springyness", and with no noticeable wear/tear of the velour material.

Here are the core measurement results:

These are "flat" frequency response headphones from 100Hz to 3kHz. There is however a broad but not terribly deep dip from 3kHz to 10kHz which spans the upper mid, "presence", and "brilliance" frequencies down to about -10dB maximally around 6kHz. Bass roll-off below 100Hz isn't bad with about -4dB at 50Hz (based on a model Harman 2018 target as discussed before) and dropped to around -10dB by 20Hz so sub-bass is weak compared to something like the closed-back AKG K371.

These are high impedance headphones ranging from 300Ω from 1kHz to 6kHz and up to 500Ω peak at a 90Hz resonance. They are quite sensitive, achieving almost exactly 100dB SPL/0.25V which would be 112dB/Vrms. The official Sennheiser specs seem to list a "sound pressure level" of "103dB" but I'm not sure if this should be interpreted as 103dB/V. In any event, the result I'm getting here seems consistent with Benchmark's voltage sensitivity result also of 112dB SPL/Vrms.

The waterfall plot above looks good with a smooth decay. Harmonic distortion is not a problem at less than 0.1% by 100Hz and remains <0.2% to 3kHz while playing the ~95dB SPL signal. Note that the (H)EARS measurement device has a strong resonance at ~4.2kHz thus artificially elevating distortion results so I cannot accurately comment on distortion level higher up.

As an open headset, don't expect much noise isolation or low leakage:

Sure, there's a little isolation from outside noise but only <10dB into 6kHz. You'll still be able to hear what's happening around you. For comparison, a closed headphone like the AKG K371 will already show >20dB of isolation from outside noise by 6kHz.

With the HD650 playing at 100dB SPL, average leakage at 1" is 83dB SPL (100Hz - 10kHz) as measured just behind the driver where the maximal leakage appears to be coming from. This translates to about 61.5dB SPL at 1 foot; high compared to the AKG K371 with 33.4dB SPL at 1'.

Finally, let's have a peek at the time-domain performance with an audio bandlimited 100Hz square wave (bandwidth up to 20kHz):

That looks good. Relatively smooth shape and the peaks are achieved by 0.75ms.

The Sennheiser HD800 (2010)

BTW: Notice the 3 bumps over the ear cup for tactile confirmation of the left side (Braille for "L"). The HD650 also has 3 bumps over the top of the left ear piece.

That's my Sennheiser HD800 bought in 2010, possibly one of the first to show up for sale in Canada. It's an early model with S/N: 012XX, "Made In Germany". Back in the day Sennheiser sent a copy of their individualized frequency response graphs for each HD800. Unfortunately, I've lost this over the years but I do recall it looked very flat with their "loudness diffuse field" target. If you're curious, it looked like this. I think it's classy to have individualized graphs to hand to the customer for a top-tier product.

I have since replaced the old earpads with some Accessory House Hybrid Sheepskin / Protein Leather (~US$40) which I find quite comfortable and likely will be more durable than the soft microfiber stock pads that literally gradually disintegrated over time! Headphone weight without cables is 390gm. The only downside is that the sheepskin can get a little warm after awhile.

Core measurements:

As you might have read elsewhere, the main criticism of the HD800 has to do with the aggressive high frequencies. We see that the frequency response does dip in the upper-mids and then rebounds strongly into the "presence" and "brilliance" regions. Most measurements of these headphones I've seen show a strong peak into around 6kHz. However, I can confirm that with mine, the treble is audibly strongest around 7kHz. I wonder if this is to some extent a result of the replacement ear pads. Indeed, different ear pads will significantly change frequency response of your headphones (some examples of the variation on Dekoni's site showing their different pads with the HD800).

Bass response does roll off but it's not bad. I'm getting less than -3dB at 50Hz and about -6dB by 20Hz compared to 1kHz. The bass sounds a bit fuller on the HD800 than the HD650 but not to a huge degree; perhaps this can be correlated to the larger transducer. 

As per specs, these are high impedance headphones and we can see that the impedance varies between 350-800Ω. Using a 1kHz tone at 0.25V, I'm measuring 97dB SPL which means a sensitivity of 109dB/V which is higher than the company's specifications of 102dB/V but less than the measured HD650 sensitivity above by 3dB. I see that Reference Audio Analyzer has the sensitivity measured at 103.8dB/V. Although the impedance is higher, the sensitivity is quite good. While it would be best to drive these headphones with a low-distortion amplifier like the Drop + THX AAA 789 I have, lesser amplifiers should be able to drive them to adequate volume.

The waterfall plot decays quite smoothly across 10ms without significant irregularities. Harmonic distortion is not a problem and stays below 1% essentially across the whole audible spectrum at 95dB SPL output level (<0.2% at 1kHz).

Again, we obviously cannot expect good isolation and sound leakage is also rather high:

Yeah, barely any isolation from the 90dB SPL external noise, <5dB reduction up to 7kHz.

Extrapolated with the headphone playing 100dB SPL, we're looking at around 67dB at 1' for leakage! These guys are like little speakers broadcasting to those sitting nearby - forget privacy. ;-)

Here is the 20kHz bandlimited 100Hz square wave as measured from these transducers, output level set at 95dB SPL at 1kHz:

Notice that compared to the HD650, the HD800 is a bit "rougher" on those peaks. We can even see an irregularity during the steepest part of the transitions. If we look at the rise time duration between the square wave edge and how quickly the waves peak, the HD800 achieves this by around 0.75ms, similar to the HD650.

Subjective comparison & Summary of HD650 vs. HD800

BTW: Notice the size of the HD800 ear piece and how it's wider, covering the screws that hold the miniDSP (H)EARS silicone piece in place.

As you might imagine, these are different sounding/feeling headphones. As different sounding as they look!

While both are open drivers, the HD800 feels more comfortable with its larger ear cup, less tight clamping force and overall a more "spacious" feel with a great sense of "air" around the ears (at least for me!). I think psychologically this plays a big part in how listeners also experience the sound.

As for the sound itself, the HD800's drivers are placed a little further out from the ears and driver mounts are angled medially from back to front which also helps to convey space and externalizes the sound a bit, the sound just seems to be coming at the listener slightly from in front (this is good). Realize though that this is still far from sounding like real speakers in a real room! Unless I'm listening to a well recorded binaural recording or using DSP to model such a sound, stereo "imaging" is still very much "inside the head" and more a lateral side-to-side experience rather than 3D with true depth or height.

After all these years of use, I'm very familiar with the sound of my HD800's. I spent a few evenings listening to the HD650 for a subjective comparison using the reference Drop + THX AAA 789 headphone amplifier on the computer Workstation serving up music through Roon and Amazon Music HD.

Let's talk mainly about the HD650 since the sound is less familiar to me.

Recently, I've been playing some NieR: Automata (Square Enix video game, 2017) and listened to the NieR: Automata Piano Collections. Some sweet Japanese video game score compositions on solo piano which came through nicely with good articulation, intimate harmonic details, smooth decays, and beautiful conveyance of emotion on the HD650. I love "Weight of the World". While good, compared to the HD800, it's more constrained and "smaller" sounding, like listening to the piano more intimately in a large room in a house instead of the spaciousness of the HD800 and a piano in a concert hall.

If you like John Williams' film music, Anne-Sophie Mutter's Across The Stars (2019, DR11, this is a "true" 24/96 hi-res recording) features a number of tracks from the Star Wars movies as well as a few non-sci-fi flicks like Schindler's List, Memoires of a Geisha, and Munich. Mutter's solo violin sounded round and organic (I was almost tempted to use the cliché "analog", but this sounds better than an analog medium like LPs ;-) through the HD650. Instrument separation between her and the rest of the orchestra on the title track was rendered with ease. Very well done "hi-res" rendition of the "Schindler's List Theme" on this album. While this album is far from bass heavy, the better bass response of the HD800 is noticeable but we're not talking about "night and day" differences. Without the ability to switch quickly with the headphones side by side, I don't think I could tell the difference.

Comparatively, this HD650 is definitely "darker" in tonality. I've heard the term "Sennheiser veil" applied to these headphones and Tyll back in the days of InnerFidelity said this as well.

To start, I think it's useful to think about what the subjective experience of a "veil" is. For me, a "veiled" sound would be one which is "muffled" which implies both "dull" tonality and a loss of resolution which one likely would be able to correlate with higher distortions and poor time-domain performance. A "veil" would result in loss of highs as well as a reduction in how well we might make out the words in a song or the various background instruments in an orchestral piece. Furthermore, a "veiled" sound to me would cause a loss in soundstage definition, "unclean" right-left separation, poorly defined virtual 3D soundstage when listening to binaural recordings. This broad concept of "veil" is not what I heard with the HD650.

Rather, to me the HD650 sounds very detailed, demonstrated excellent channel separation, good conveyance of the 3D binaural effect (have a listen to the classic binaural barbershop), and instruments in an orchestra were not difficult to pick out (I was also listening to the Mo-Fi remaster of Walter Susskind's Holst: The Planets the other day). I believe the problem is simply that the frequency response with the relatively broad dip particularly from 5-10KHz is just a bit too much, making the HD650 (at least of this vintage) sound a bit "boring", "dull" especially with acoustic music.

A live album like Def Leppard's recent Hits Vegas [Live at Planet Hollywood] was disappointing and sounded lifeless with the HD650 compared to the HD800 (still far from the best live recording I've heard, mind you).

No need to fret though because EQ will improve this (discussed below), and I think there are genres that can benefit from this "dark" stock sound. Have a listen to some of the synthetic soundtracks like the recent Benjamin Wallfisch's Mortal Kombat (2021, DR10 avg). The first track "Techno Syndrome 2021" I think sounded better with the HD650s, less grating and harsh compared to the HD800. Likewise, modern mastered pop like ohhh.... Justin Bieber's Justice... Sam Smith's Love Goes? ;-) So I think if you're into pop, studio rock, and recordings which sound "cooler" like old '80s stuff (eg. Don Fagen's The Nightfly or Midnight Oil's 1987 first pressing of Diesel and Dust), this kind of HD650 frequency response could be pleasing. 

Interestingly, the ASR measurement of the HD650 did not show the same amount of treble sag that I'm seeing (and hearing!) here. However, the RTINGS measurement is showing a similar frequency response profile. I believe there is a school of thought suggesting that newer HD650s are less "dark" and you can see measurements on DIY-Audio-Heaven (eg. 2014 vs. 2017 model) showing significant differences. Seems like a bit of a crap shoot if this is true and if one is buying used, it would be important to run a check on the serial number. Another possibility is the ear pads themselves causing the tonality to change with age. Here's a nice example from Crinacle. I'm not sure how much deterioration or age there needs to be to cause these changes though.

Another look at the Sennheiser headphones and ear pads... Stock velour pads of the HD650 and Accessory House Hybrid Sheepskin/Protein Leather for the HD800.

Interesting what SolderDude said in his DIY-Audio-Heaven post on the HD800: "To me, without doubt, the HD650 is technically the better headphone." Indeed, even with a "dark" sounding HD650, it has a flatter measuring overall frequency response, demonstrates smoother waterfall decay, has less overshoot and "roughness" when looking at the square wave reproduction (larger size and mass of the HD800 driver probably contributing), and lower overall harmonic distortion. Although I'm measuring my HD800 with the aftermarket replacement earpads, I have no reason to believe these characteristics would change even if I had used new official Sennheiser pads.

Having said that... Without a doubt, I prefer the overall feel and sound of the HD800.

As humans, the frequency response I think trumps everything else and the "dark" HD650 sound is not to my liking. Furthermore, the way the music and sound "tickle" the ears, the voluminous impression of "space", "transparent" atmosphere, subjective "speed" and the taste of somewhat-outside-of-the-head "soundstage" delivered by a pair of headphones playing standard stereo music make the HD800 a superior experience. Beyond the frequency response, many of these characteristics probably cannot be measured in the traditional sense, and likely would be idiosyncratic for each listener anyways. Perhaps the best I can do is suggest that you listen to the tracks "Bubbles" and "Letter" from Yosi Horikawa's binaural EP Wandering through the Sennheiser HD800(S). You might experience the same impressive qualities I described above. (There are many very good headphones these days, the tracks may sound very impressive already!)

There's a tactile intimacy about the headphone experience. These kinds of experiences which cannot be easily quantified reminds me that as an audiophile, while I maintain the "More Objective" perspective, this isn't "All Objective". Having specs, graphs, and other data in hand is important and while these facts may explain qualities like the output of DACs thoroughly, once we move into transducers and acoustics, there will be other factors which we must integrate in order to reach the balance for what's best for ourselves.

Regarding the HD800, while there's an upper-mid dip, the restoration of the "presence" and "brilliance" content along with a bit more bass simply makes the sound more pleasurable with a subjective "sparkle" and "excitement". Psychoacoustically, I think we can make a case that the 3-4kHz dip reduces harshness - echoes of the "Gundry/BBC Dip" or Linkwitz's discussion here. Yes, the 6-7kHz peak (6kHz in most measurements of the HD800) could be a bit annoying with bright recordings. Here's a graph from Sennheiser showing the frequency response difference between the HD800 and HD800S:

Clearly we can see Sennheiser's intent to dampen the frequency response at the 6-7kHz region when they released the HD800S.

To end, let's just talk briefly about adjusting that "dark" sound of this Sennheiser HD650. Based on what I have heard/seen here, and if I owned these headphones, I would try:

1. Getting new ear pads, genuine replacement pads are not expensive. There is some evidence that as pads age, the high frequencies weaken and potentially results in that "dark" or "veiled" sound as suggested here. SolderDude speaks about it here also. 

2. Apply EQ: A simple +6dB, 7kHz peak with a broad Q2 will restore most of the treble dip for the Sennheiser HD650 measured here. I've tried this myself and I like the result. Make sure you apply some headroom to avoid clipping. 

3. Research and look for mods that could brighten up the headphones. Check this out to start.

As for my HD800, I'm quite happy with the stock sound for the most part. A couple of EQ settings with a peak of +4dB, 3.5kHz, Q1.5 and a dip of -3dB, 7.2kHz, Q2 will help fill in the upper-mid dip and reduce the aggressive upper treble.

Both the HD650 and HD800 can benefit from a low-shelf to boost the bass: +3dB, 80Hz, Q1 will do the job, IMO. No point forcing bass further IMO, pushing distortion higher. I don't think most people will complain of the amount of bass from these headphones.

Here's what it looks like in Roon for linnrd's HD650 and my HD800:

HD650 EQ

HD800 EQ

As usual, experiment for yourself with EQ settings. Although my measurements reference a Harman-like headphone target curve as a useful empirically-derived "standard", Harman is not Gospel. ;-) Also, unless you're specifically measuring your own headphones, make sure to look around at other measurements out there to consider broad tendencies rather than being glued to a specific set of EQ parameters thinking that what works for others would apply to your own headphones! For example, the typical 6kHz peak with most HD800 measurements is different than for mine.

Identifying the general tonality will help suggest broad EQ settings that might be "good enough" for a particular headphone model. As for the parametric EQ settings themselves, as the Q gets larger, this results in a narrower frequency effect. Before blindly applying settings, make sure that narrow EQ settings (like Q>3) are not just idiosyncratic findings for a specific unit. Be vigilant to "do no harm".

Thanks again linnrd for lending me this HD650 to test out!

--------------------

I think the news of Sennheiser wanting out of the consumer business reported back in February 2021 is unfortunate given the company's long history and their reputation among audiophiles. Of course, things change and such is the business world with major competition these days in consumer audio and shifts in trends. 

I will refrain from looking around at audiophile news this week and commenting on what other people say. Audioholics' editorial of the week is interesting. I still contend MQA is every bit a scam as any faith-based AudioQuest cable, silly "jitter" insects and such. ;-)

I added an upgrade to my humble gaming PC recently. My old AMD FX-8320 CPU + ASUS M5A97 2.0 motherboard + 3GB nVidia GTX 1060 + 24GB DDR3 RAM still is able to play modern games at 1080P quite well. The old AMD CPU could be a bottleneck these days but I'm too lazy to rebuild the machine. Here's the hardware upgrade:

It's great how inexpensive SSD drives have become in the last number of years with the newer TLC and QLC (triple-level and quad-level cell) NANDs. The multilayer "3D" drives along with multi-level cell NANDs have allowed for higher densities so they can start to rival hard drive capacities. However, these multi-level cells are slower to access and have lower endurance due to limited program/erase cycles compared to SLC NAND, but still way faster than a spinning hard drive.

These days I think an inexpensive TLC model like the Silicon Power 2TB SSD, currently <US$180 and look for sales (I have the A58, but should be similar to A55) with (pseudo-)SLC caching which makes up for lack of DRAM cache hits the sweet spot of budget price for non-critical applications. Similar 2TB inexpensive drives around this price point can be had from Crucial and Western Digital. Pay a little more for the Samsung 870 EVO with great reputation for more critical applications.

Previously, I did not have enough storage for Mortal Kombat 11 on my gaming PC - not a problem now.

CrystalDiskMark x64 8.0.1 speed test:

Ran the CrystalDiskMark right after I used Macrium Reflect to clone the OS over.

Notice the nice increase in speed comparing a circa 2012 OCZ Vertex 2 SSD drive based on the SandForce SF-1200 controller and the new Silicon Power based on a more modern Phison PS3111-S11 (or related model) controller even on a modest machine. I know some folks have issues with the S11 controller - will see how this works out for me. The SATA interface has been bumped up from SATA II to SATA III. Obviously, back in 2012 when I bought the OCZ, the idea of an affordable 2TB SSD would have been foreseeable at some point but the exact timeframe would not have been clear. Here we are 9 years later, >16x the storage space, better performance across the board, for around the same price.

Speaking of price, this brings us to one last thing. With COVID supply chain disruptions and the explosion in the broad commodity complex, clearly the governments of this world are OK with inflation and forcing their definition of inflation "targets" through deficit spending. This screws folks on fixed-income which is a discussion for another time and place. Things are getting more expensive in consumer electronics where margins are thin and manufacturing expenses are going to have to be reflected in the final price to the consumer. Already, many items cannot be obtained. Anyone tried to buy a latest generation nVidia RTX 30X0 graphics card lately? Or Radeon 6000-Series at non-scalper prices?

Component prices (such as Phison that makes controllers for many SSDs) have been signaling price increases for months. I don't know how this is all going to play out eventually, but given this trend, for now at least, if you find a good deal on something on sale, probably not a bad time to grab stuff you might need/want for the next year.

I hope you're doing well, dear audiophiles and music lovers.

Enjoy the music as we enter May. Catcha later...