Hifiman Ananda

[Lbstyling]

Hello,

Aim: I want the best sounding mobile headphone setup possible. This could be high efficiency phones using my mobile to drive, or lower efficiency, but with a small light dongle (such as the E1DA)

I am currently using Superlux HD861 headphones with the official velour pads. System is now simply with my phone (xaoimi K20pro) since I found the android 'wavelet' app (100%recommended) and the preloaded Oratory's EQ profile for those headphones.

The sound is frankly shockingly good already, and a large part of me thinks I just cant see me beating them significantly as they are just so good with EQ! (£35 Inc pad is just silly for SQ like this.) But I would like to try and see just how good it can get.

I would like to find the best performing headphone that will run directly from a mobile phone or portable Bluetooth DAC/amp that I can for under £700.

The exact EQ profile will be adjusted to the Harman profile (with a couple of Db's of added bass as per my preference)

The strongest candidate I have found so far is the Hifiman Ananda. In many respects, I would have preferred the Sundara, partly as it costs half the amount of the Ananda. In terms of FR the current Sundara version is a better match to the Harman curve but it seams it would require an amp to get loud whereas the Ananda does not.

Rtings rate the Ananda as having a wider soundstage. The question is, is this entirely a function of the EQ profile, or is there more going on?

Sundara:

Ananda:

Arya

Given the now widespread availability of EQ, I feel the vast amount of reviews of headphones now don't offer the useful information they once did.

I know Solderdude suggests the data Rtings offer above is based on some testing (using the sennheiser HD800 as a benchmark) but is there any correlating evidence? Even Amir appears to describe soundstage as somewhat separate from FR (albeit impacted by it). Sean Olive appears to feel it is largely if not entirely a function of FR.
I'm sure many of us have tried EQing headphones to match the HD800s. It doesn't work. This leaves me following the Rtings data.

My other concern is decay:

Much of this is 45db down, but at 900hz it looks closer to only 20db down at what? 13us?

Any thoughts welcome!!

 

[maverickronin]

Rtings rate the Ananda as having a wider soundstage. The question is, is this entirely a function of the EQ profile, or is there more going on?

Quoting my self from another thread.

I've pieced together a few rules of thumb from DIY experiments over the years.

The basic shape of the soundstage is often determined by FR and can usually be modified to a degree by EQ. IMO a DF EQ usually give the most natural soundstage with more forward projection. Reducing DF's 2.7kHz peak and/or shifting its frequency can adjust the depth and width to a degree.

Overall "openness" is another factor. This is kind of akin to cabinet resonances in speakers. An open design can just let the backwave and any other excess energy radiate away. Closed designs need to damp and absorb it.

IMO, the biggest factor in soundstage is the positioning of the driver with respect to the pinna. When the driver is is larger, farther away, or at an angle with respect to your pinna it adds more of your personal pinna transfer function to the final sound reaching your eardrum. This makes it sound just a bit more natural and will increase the size of the soundstage.

Putting these all together we can examine a couple designs.

The HD800 has a larger than average driver held at an angle and distance for the pinna in a very open design. This gives a greatly enhanced sense of space. The downside of holding the driver in that position is the less than ideal FR with a peak corresponding to the wavelength of the diameter of its cups.

In contrast the HD650 has a smaller driver pressed directly against the pinna. This makes for a much more controlled "room" for the driver but greatly reduces the sense of space.

Electrostatics are also known for their large soundstages.

Compared to dynamics they have much larger drivers. They are also very open. Their baffles are sealed (or ported) on the ear side, but the drivers themselves very acoustically transparent. Some planar magnetics can nearly match a 'stat in driver size but the drivers themselves are less acoustically transparent. This can be easily demonstrated by just holding your hand an inch or two out from the back of headphone and hearing how much the sound changes. No current production 'stats I'm away of though have angled drivers. There is the old Stax Sigma though.

Overall, this makes a very open sound with a great sense of space. The non-angled drivers drivers keep it from having much in the way of forward projection though.

As to why this kind of thing is copied more? I'm not entirely sure.

Part of it is probably difficulty in getting a good FR in a physical design like the HD800. The HD8XX is now their third try and the prototype still needs just as much EQ as the original.

Part of it is probably cost. The other path to this is large drivers and those are expensive if you want them to sound good.

'Stat drivers aren't inherently expensive, but the amps they require are, and that is a hard market to break into.

Large planar drivers are expensive too. Lots of strong magnets trying to to tear apart the stators, lower yields on larger diaphragms, probably more issues tensioning them, and then the issues in making them acoustically transparent.

If you make a dynamic driver too big it stops being a tweeter and turns into a midrange. There are plenty of 40 and 50mm headphone drivers which are already twice as large as average tweeters on speakers, but which are expected to got up to 20khz on their own none the less. That's probably why Sennheiser went to the trouble of making the HD800's 56mm driver a ring radiator.

...

Anyway, these are just some random thoughts before I've even finished my first cup tea this morning...

Short answer is that EQ can help some but, just as much of the soundstage is inherent to the headphone's physical construction. You have to experiment with EQ on your own too. Aiming for a generic target curve will at best give you generic, middling results.

Rtings measurements should be seen as an indication of how much the headphone and driver interact with the pinna of the wearer. When you wear the headphone it will interact with your own pinna, add more of you personal pinna transfer function to the final FR reaching your ear drum, and increase the sense of space.

 

[Lbstyling]

Quoting my self from another thread.



Short answer is that EQ can help some but, just as much of the soundstage is inherent to the headphone's physical construction. You have to experiment with EQ on your own too. Aiming for a generic target curve will at best give you generic, middling results.

Rtings measurements should be seen as an indication of how much the headphone and driver interact with the pinna of the wearer. When you wear the headphone it will interact with your own pinna, add more of you personal pinna transfer function to the final FR reaching your ear drum, and increase the sense of space.

Thanks for this.
Are you suggesting I can widen the soundstage by EQing up/down a specific range that matches the shape of my ears?

 

[solderdude]

If that were the case one could make any headphone sound like a HD800 just by EQ.
You can make them tonally very close but not in other aspects.
Driver size, ear distance and above all angle, as well as the info in the recordings used are important for the 'imaging'.
Even when you were to put HD650 drivers (smaller in diameter) in the HD800 housing they still would not have the same effect due to the surface area and driver construction (ring radiator)

I have never been able to EQ a HD650 to HD800 tonality AND make the stereo imaging closer. The different tonality (more treble) does help a bit with pinpointing sounds and 'detail' but not in sharper defined instrument location.

 

[maverickronin]

Thanks for this.
Are you suggesting I can widen the soundstage by EQing up/down a specific range that matches the shape of my ears?

You can usually improve a specific headphone's soundstage with basic PEQ, but it won't transform the spatial presentation into that of another completely different headphone.

In the end the differences in soundstage between two headphones do just come down to FR, but the headphone with the better soundstage interacts more with you pinna and gives you a more personalized FR. To full emulate that you would need to use convolution with personalized measurements of each headphone on your own head.

As examples Impulicifer and and the Smyth Realiser emulate speakers over headphones in that manner. Thy also account for inter aural timing and level differences which are beyond the scope of a passive headphone.

...

Also, are you just looking for a wide soundstage or do you want something with a little forward projection that sounds closer to speakers?

 

[Lbstyling]

You can usually improve a specific headphone's soundstage with basic PEQ, but it won't transform the spatial presentation into that of another completely different headphone.

In the end the differences in soundstage between two headphones do just come down to FR, but the headphone with the better soundstage interacts more with you pinna and gives you a more personalized FR. To full emulate that you would need to use convolution with personalized measurements of each headphone on your own head.

As examples Impulicifer and and the Smyth Realiser emulate speakers over headphones in that manner. Thy also account for inter aural timing and level differences which are beyond the scope of a passive headphone.

...

Also, are you just looking for a wide soundstage or do you want something with a little forward projection that sounds closer to speakers?

As close to real as possible always. Large image is always nice for me.

In terms of speakers, I like the deep image with instruments placed Infront and behind of each other that I get projected from my horns, but I can imagine this is hard to do on headphones.

I will look into the speaker emulation software you list. I assume special mics may be needed. I have a galaxy cm140 at the moment, but that's it.

 

[maverickronin]

As close to real as possible always. Large image is always nice for me.

In terms of speakers, I like the deep image with instruments placed Infront and behind of each other that I get projected from my horns, but I can imagine this is hard to do on headphones.

If you want depth, and don't end up using something like Impulcifier of the Realizer then the HD800 is the way to go. With crossfeed, or some other kind of HRTF simulation software you may be able to get something approaching a very near field speaker setup.

With fully personalized impulse response convolution, anything that's open and suitably eq-able should give good result.

I will look into the speaker emulation software you list. I assume special mics may be needed. I have a galaxy cm140 at the moment, but that's it.

AFIK the Realiser is still unobtanium with a years long wait time. Impulcifier is free, but you need your own in ear mics to take the measurements.

For both of them, you need reference speakers/room, but it sound like you already have that.

...

I used to use TB Isone (now discontinued?) which was much better than basic crossfeed but could only really be used with some playback scenarios since it was a VST plugin. I got sick of every source sounding different so now I do all the DSP outboard in the RME ADI-2 DAC which has crossfeed, PEQ, and a bunch of other nifty features.

 

[Cahudson42]

Just in case you might not have considered it, the N90Q is still apparently available from Harmon for the next 10 hours at $299, rather than the $1499 list (usually Amazon is around $400 -$500).
https://www.harmanaudio.com/akg/N90+Q.html

As an 'all in one' it will run directly off your phone. It also has an apparently good soundstage via it's crossfeed implementation.

Tyll did both a video and detailed written review for InnerFidelity, which is still available - just Google for it.


Video:

At $299, if it doesn't suit, you most likely can resell for close..

 

[MayaTlab]

Rtings rate the Ananda as having a wider soundstage. The question is, is this entirely a function of the EQ profile, or is there more going on?

Frequency response above a few kHz is heavily influenced by your own anatomy. Above around 5000hz it's likely to become quite difficult to predict what the peaks and dips will be like on your own head. Too bad as FR, particularly in that range, is one component of how humans locate sound cues.
Rtings' PRTF measurements presume that their findings on their pinna will translate to similar findings on other test pinnae and humans, but I'm not sure that it's been effectively tested. And remember that it's only an evaluation of the FR differential between pinna and no pinna, if the basal FR curve is out of whack to start with and doesn't correspond to what your own anatomy mandates results will be sh...t anyway.
"Soundstage" is a word that has so far not found much common definition and I personally struggle to understand what people mean by what.
For what it's worth I never found that the HD800 gave me a particularly good rendition of reverbs or a sense of space, or a capacity to effortlessly peer into the track (whatever that means, honestly I'm not sure) and running sweeps at higher frequencies made it quite apparent why (too many oddly placed painful peaks or dips on my head that I found difficult to effectively EQ by ear, EQing trebles past 5000hz or so is a b*tch and measurements won't help much as it varies too much between individuals). I found it easier to reach a satisfying point with a pair of HD650. Your own impressions will necessarily differ from mines.