Fostex TH-X00 Measurements and Subjective Impressions

I've spent nearly 2 years trying to tune the TH900, by modding its ear pads, driver damping, and cups damping. I managed to get its bass much better, slightly improved the 3-4 kHz dip and the treble, but not by any means I could get the 500 Hz dip fixed, not even by re-tuning the Onken loading of the TH900 drivers (BTW hats off Fostex, their factory tuning was optimal). Since the measurements without the cup and ear pads (wonderfully flat FR) don't show the 500 Hz dip, I assume it's a cancellation mode because both the cups and ear pads. The search continues, but here is the current state. The green line is the modded TH900, the red line the same when some Sorbothane damping was applied. The plots are 1/6 octave smoothed.

With the modified ear pads and cups damping (which are near optimal I'd say), using the original Fostex dampers (the grey foam ring around the drivers) results in 4-6 dBs more elevated bass from 100 Hz and down, going in a straight line to 20 Hz. The 1kHz peak is lower. It sounds like a bass boost, but most of the time it's not needed.



While it sounds much better (even in the bass) than the stock TH900, it still doesn't measure nowhere near as good as the TH-X00. I may end up buying one.

 

Have you tried making the baffle leak on the ear side? I've had success reducing the 'port-related' dips by using a leaky baffle.

 

This sounds interesting. Leaking what way? It is already leaking by construction, through the baffle and the ring damper. Did you make another hole in the baffle? Where?

 

For instance, the T50RP mk3 has a 'double baffle' with ports from the ear-chamber to the outside, not back into the cups. From pictures it is not clear whether the TH900 baffle leaks to the outside this way.

 

Just joining the club. Landed me a pair this morning. First pair of closed cans.

 

The TH900 also has double baffle, but there is little room between them to make an opening on the edge. Multiple small holes may be an option, but it's hard to calculate, with the already super complicated loading. Perhaps the earpads are loose fit on a purpose... But then the X00 has tighter fit, yet no dip, and no port visible either. Anyway, I will make a leaking version of the pads and let's see.

 

I'd look into making a custom mounting ring that would allow you to sandwich a ring of felt underneath to allow for controlled leaking before permanently modifying the baffle or pads.

 

Nice idea, but the construction of the pads does not allow that. One needs to modify the pads or use new ones - but that is easy, and replacement pads are relatively cheap. I am also considering making a completely alternative ear pad. Anyway, that probably belongs to another topic in mods.

 

You're right, as a closing thought, Stratocaster has experimented with these 'leaky baffle' schemes for longer than I, he could give you some pointers.

 

To conclude the case of the modded TH900 here - I got it better without leaking baffle. Very close results using 3 different mods, all seem converging here and I suspect it won't get much better. This one has modified Fostex earpads, with a custom carbon foam insert, 5mm thick wool felt damper, and original cup damping. Also, similar results were obtained using Stax 009 and 007 earpads with custom foam filling, original Fostex damper and original cup damping. However, this is the best sounding of them and doesn't require an expensive new pad, just about 50 bucks worth of materials and tools, and a lot of tinkering. There are a few more mods with different cups damping, but they have a slight ringing around 3.5 kHz on the waterfall, so I have ruled them out.

The X00 still measures better, but this is not that bad either, especially considering from where it all started.

FR: https://goo.gl/photos/hVtCdhvx6itqs4uXA


Distortions: https://goo.gl/photos/LneNTTJkryyNZxPg7


Waterfall (rollercoaster): https://goo.gl/photos/n9d5BbYxfCFe9nrj8


Spectrogram: https://goo.gl/photos/dDtNY6pgivh2fBLVA


The bass overshoot is very benign and mids are not masked in listening tests. Distortions are quite low, and most of them is 2nd harmonic. Still a bit bumpy FR, and as expected, doesn't sound as smooth as my Stax 007 Mk1 and 009, but very good for a closed "fun" headphone, driveable from almost any portable device.

For bassbanging fun, I can put back the original Fostex dampers, and end up with this FR.
https://goo.gl/photos/4s5JbXYSbCLHiCfEA


Despite the apparently small, ~2 dB difference in bass, the green line sounds much more bass-heavy with music and does mask the midrange. With the original Fostex pads it's about 6 dB's difference below 100 Hz.

 

That looks quite listenable. If you want to remove the midrange bump you will have to perforate the bottom of the donuts a-la jergpad, but that would leave you with a bloated bass (go here to see how it affected my Philips Uptown, similar results with the T50RP with the same pads).
Can you provide pad-less measurements to see if the higher frequencies become smoother?

 

Actually I think there is no midrange bump, but dips at 500 Hz, 2 kHz, 4.5 kHz, 9 kHz. Pad-less measurements would require a different microphone position and would be more near-field than with pad. I tried that, but putting a felt disk around the mike had more effect on smoothing high frequencies. Some of the raggedness, especially deep and steep dips in the treble comes from measurement setup limitations. I wonder if in-ear measurements (with right compensation) would be better for measuring the highs more accurately.

Good you mentioned the effect of perforations on the pads. The Bose QC15 has that, and since I will fly soon and will have to use it, I need to check if its bloated bass can be somewhat addressed with a different pad.

 

1- I would really encourage you to try in-ear measurements. In the end I feel that every coupler is a compromise and will mask or amplify relative differences (like you noted with the ~2db difference in bass). Obviously getting great repeatability will be much more difficult than with a normal coupler. You really need both perspectives.

2-You can always use IEMs under the QC15 for maximum isolation and best sound if you want.

 

QFT, I now use both solid-plate and ear-mounted setups to paint a better picture of what is there.

 

Thanks! Any pointers towards in-ear setups? Is the Roland CS-10 good?
I have strange ear canals so probably won't use them on my own head, but I could perhaps bribe my wife ("darling, could I use your not-so-dummy-at-all head?" .

 

Oh, I just use a Panasonic WM-61A that I put in a silicone ear tip. Dirt cheap and pretty good results. The whole setup cost me less than 20€ (for 10 mics and a preamp).

 

I got mine from BlueMonkeyFlyer @ Head-Fi, he may still have some to sell.

 

Just received these headphones, which I bought from @JoshMorr

Very first impression - Wow, that is a lot of bass! I wasn't expecting such bass monsters. It's very good bass though.

Overall very pleased so far. I've never had a pair of closed headphones that really did it for me. Oppo PM-3 seemed very neutral and well behaved but were not enjoyable to listen to. In contrast, these may not be very neutral, but they are very enjoyable to listen to. Certainly will relegate my DT770 to dust collecting status.

 

Wrong thread!