Hifiman Ananda Review (headphone)

[phoenixsong]

I think these measurements are important for us to have so we have comparison points for quality and Amir is doing us a great service by doing his work.

However, I think attitudes on here can be on the extreme end of dogmatic to the point of sounding absurd at times. There are countless people commenting that they get equivalent or better sound quality listening to cheap AKG headphones. I don't see how this is possible. It reminds me of Tyll saying he liked the ATH-m40x over the Z1R - after that statement, it was more or less widely perceived as fact. In that case, I've heard both and I actually own the Sony's. There's no comparison between those headphones.

I see a similar thing happening here. People patting themselves on the back saying their $100 AKG headphones sound better than Focal Clears and simultaneously sighing a relief that they just crossed off the Ananda's from their wishlist and didn't waste their money. It's funny to see honestly. I have tried the Anandas, I don't own them. I own the Aryas, and they are relatively close but not quite. There is a presentation to the music that Amir hints to that is truly a special experience - one that you don't get with HD 650's, for example. Whether or not this quality is worth the asking price is up to consumers to decide. For me, the Arya is one of the most unique and exciting presentations of music I've ever heard and are worth it to me.

I think looking at graphs for most things reviewed on this site is pretty cut and dry.. amps, dacs, etc. But at least for headphones, I would really advise for people to listen to them in person before claiming to know what they sound like. These reviews are great references for headphones' hardware technicalities and also are great for hearing Amir's subjective takes on the qualities of particular headphones. With that said, I truly believe it is the latter which matters more for these products.

Also will just add: I can see Amir likes bass/subbass. Oh how I would really love to see some thoughts on the Sony Z1R's...

The Z1R did not sound clean to my ears. Well, we can only wait for the measurements as proof or otherwise (although it will be useless to those who do not trust them )
However, I do agree that people have different subjective preferences and are fully entitled to pursue that which they choose. Also, with many factors in play, headphone measurements may be slightly off as highlighted repetitively. That being said, I do not agree with those who claim objective measurements are mistaken simply because they do not match the overall subjective pleasure perceived upon personal listening

(From a purely subjective standpoint, what makes your subjective listening impressions more entitled than those who enjoy the sound from their AKG headphones more?)

 

[Nango]

Ananda ranks well behind Sundara on this ranking: https://github.com/jaakkopasanen/AutoEq/blob/master/results/RANKING.md

 

[Nango]

The Z1R did not sound clean to my ears. Well, we can only wait for the measurements as proof or otherwise (although it will be useless to those who do not trust them )
However, I do agree that people have different subjective preferences and are fully entitled to pursue that which they choose.

If I may add, not only people have different subjective preferences, one person has different subjective preferences depending from so many variables, day or night, after or before work .... and so on. I do question if it is wiseful to trust to only what you hear when testing this and that.

 

[phoenixsong]

If I may add, not only people have different subjective preferences, one person has different subjective preferences depending from so many variables, day or night, after or before work .... and so on. I do question if it is wiseful to trust to only what you hear when testing this and that.

Indeed. I have personally experienced that too, as well as the effects of conditioning

 

[GWolfman]

Thanks!

 

[Nango]

Indeed. I have personally experienced that too, as well as the effects of conditioning

I add again, unless you are people like @solderdude or Amir, given their level of experience and background I think they are more like machines when judging and talking about headphones, they are not subjective or biased (they can't!!!) any more, but purely objective.

 

[Robbo99999]

Wow, seems best to buy the HE4XX or HE400i if you're gonna buy a HifiMan headphone! Bit of an overall ugly result for this expensive headphone, seemingly unfixable distortion type issues north of 1kHz as seen in the fine grass jaggedness of the frequency response and the Group Delay issues. Doesn't even have low distortion in the bass which is normally the plus point of planar headphones.

 

[Nango]

Other than hard to drive, I dont see how anyone can dislike he6 or se. But hey your money, you chose where to spend it. -

This sentence we have heard so often. And I disagree. But I admit it is difficult to explain. It is not about telling you (not you personally, of course) where to spend on or not. But under certain circumstances you got to admit, it was then a mistake, which you normally won't do. You are free to do mistakes, of course.

 

[solderdude]

regarding seal issues.

I too have a flatplate and had seal issues with several similar sized Hifiman when measuring.
Some only had it on one side which puzzeled me at first.
Most pads are perfectly 'flush' around the ear (smaller area) but these Hifiman, because of their size, are shaped to facial contours.
The pads, also because of their large diameter, take quite some force to flatten enough when seal isn't great on my flat rig.
What I do in these cases is wrap velcro tape around the headphone and rig and tighten it so I get good bass response.
With some hifiman I had to apply quite a lot of force.
This force is not needed on my head b.t.w. it follows the contours well.
The thing is when you have a good seal (Amir probably had on his head) the 'Harman boost' is not present and thus sounds not so impressive in the sub lows (as seen on my plots).

Some are bothered by the treble peak which is masked by the large dip around 9-10kHz. It is good to see Amir's EQ adresses this as opposed to Oratory and even worse Jaakko who even increases the treble 2dB.
I lower the 8kHz peak about 5dB using a passive filter with the same BW as per my plots and have had responses from owners that were really happy to use it without EQ but just using the filter.

John may actually have a point that the treble peak becomes lower when positioning it in a less intuitive way. This is true for many headphones by the way.

Eyeballing Amir's EQ I would think this makes a LOT more sense than the EQ expert's attempts.

 

[solderdude]

Wow, seems best to buy the HE4XX or HE400i if you're gonna buy a HifiMan headphone! Bit of an overall ugly result for this expensive headphone, seemingly unfixable distortion type issues north of 1kHz as seen in the fine grass jaggedness of the frequency response and the Group Delay issues. Doesn't even have low distortion in the bass which is normally the plus point of planar headphones.

Before coming to such a conclusion I would recommend to own both for at least a few weeks, preferably side by side.
When one is on a budget I would agree. Using ears and as is the Ananda was a clear winner to me. IMO only needs the treble peak addressed and some Harman-esque subbass boost.

Distortion at 90dB is VERY low (would be a 'normal' all day long listening level where most folks would use it)
It reaches the very bottom of my test rig and is even lower than both 400i I measured.

 

[Robbo99999]

Before coming to such a conclusion I would recommend to own both for at least a few weeks, preferably side by side.
When one is on a budget I would agree.

Distortion at 90dB is VERY low (would be a 'normal' all day long listening level where most folks would use it)
It reaches the very bottom of my test rig and is even lower than both 400i I measured.

Fair enough, that's another data point.

Ah, one question I've been meaning to ask, do you know if your distortion measurements on a flat plate are directly comparable to distortion measurements on an artificial ear like the GRAS units? In other words, do they produce the same distortion results or does the inclusion of the pinna, etc influence the measured distortion results? The pinna does influence the frequency response, so you might think that distortion % could be affected in the areas that are boosted (or otherwise).

By the way, I like your website, particularly I've used it a lot in the past to look at distortion measurements in headphones (as well as your physical ear cup dimension measurements).

 

[Chocomel]

[QUOTE="amirm, post: 739041, member:
As I have explained, the proof of such measurements is in controlled listening with EQ on and off. In this case, my listening tests agree with weak sub-bass performance. If you want any more certainty than this, I don't know how to provide it.[/QUOTE]

In-ear Measurements like for example rtings does would be a great way to confirm it.

 

[paolomo]

Thanks @amirm for the great work and @cursive for providing the sample! I am happy to report that my humble EQ settings are pretty close to Amir's, this makes me a happy panther

 

[solderdude]

Amir measures at levels others never do. Most headphones will be in trouble but at 114dB SPL. In the lowest bass less than 10% is not really audible at these listening levels.
When one is reaching 114dB at around 5kHz (even peaks in music) the bass would already be closer to 130dB or so. One wouldn't want to have this headphone on your head at these levels when upper mids reach those levels.
Bass levels... sure, certainly with a lot of bass boost that would be enjoyable (for the duration of a song) with 114dB bass boost.
It looks weird and poor quality at 114dB but in reality is looking at levels most will never ever reach (those that value their hearing).

Amir provides a measurement point others don't with the 114dB. Personally I look at the plot of 114dB and then disregard anything that happens above 100Hz and look at the 104 and 94dB traces to get an idea of distortion in the mids (where hearing is most sensitive).

Ah, one question I've been meaning to ask, do you know if your distortion measurements on a flat plate are directly comparable to distortion measurements on an artificial ear like the GRAS units? In other words, do they produce the same distortion results or does the inclusion of the pinna, etc influence the measured distortion results? The pinna does influence the frequency response, so you might think that distortion % could be affected in the areas that are boosted (or otherwise).

By the way, I like your website, particularly I've used it a lot in the past to look at distortion measurements in headphones (as well as your physical ear cup dimension measurements).

A pinna probably does alter distortion at mid-high frequencies somewhat Can't say how much a few dB probably at most.
I use a $ 0.60 microphone capsule with a fixed gain home made mic pre-amp and nowhere near AP quality distortion levels ADC (EMU 0402) .
Above 95dB the measurement chain distortion becomes a thing and at just above 100dB becomes too high so alas cannot measure higher.

 

[Chocomel]

Great review, as always!



But which type of measurement better represents the audibility of the distortion?

Rtings will have better correlation since it's at least somewhat weighted. THD(and also IMD) have quite poor correlation with sound quality, now rtings uses somewhat of a custom weighting so that makes a comparison difficult without more research but i would feel comfortable saying it's better than nothing. There's other metrics for distortion that have been the result of research and show way better correlation with sound quality for example GedLee, especially at lower levels of distortion. I wrote i small bit about it here https://www.reddit.com/r/headphones/comments/ltwyz9
Links to the papers
http://www.gedlee.com/Papers/Distortion_AES_I.pdf
http://www.gedlee.com/Papers/Distortion_AES_II.pdf

 

[Robbo99999]

A pinna probably does alter distortion at mid-high frequencies somewhat Can't say how much a few dB probably at most.
I use a $ 0.60 microphone capsule with a fixed gain home made mic pre-amp and nowhere near AP quality distortion levels ADC (EMU 0402) .
Above 95dB the measurement chain distortion becomes a thing and at just above 100dB becomes too high so alas cannot measure higher.

That would be something to bare in mind, but you do have another data point there which is useful in comparison.

I too don't hold much water to the 114dB distortion measurements, but just like you I see the benefit of that measurement in the bass area when taking bass EQ into account when listening at normal to loud levels.

 

[charleski]

While I do agree the graph may be heavily smoothed, you can move your cursor to any point on the rtings one and see the actual readout. So not comparing the graph, but the actual data it seems to be radically different as well. For instance the Rtings graph doesn't go over 1ms delay until under 64hz, which we can see Amirs easily surpasses 2 and 3ms at various points throughout. Perhaps I'm reading it wrong, or misunderstanding, but it seems on Amirs group delay it can have positive and negative delay at the same frequency? Or is that just due to the horizontal scale we're looking at? It seems on rtings at any given point it's either positive or negative delay in ms.

The data readout shows the values that are plotted on the graph, and includes whatever smoothing they’re doing. Unfortunately they don’t specify the smoothing they use. The purpose of measuring group delay here is to detect areas where the headphone response isn’t minimum phase, as EQ can only really fix minimum phase errors. So you’re looking for sharp upward spikes that indicate group delay in excess of that present in a minimum phase system. Any sort of smoothing will simply flatten these and render the plot useless for this purpose. At higher frequencies amir’s plot is showing a thicket of rapidly-changing delay values that cluster together on the scale shown.

There’s a discussion of this here: https://www.roomeqwizard.com/help/help_en-GB/html/minimumphase.html
This covers speakers, which experience far more group delay than headphones.

 

[amirm]

One reason to NOT mess with distortion plots is to see what distortion mechanisms are in play. This can point out to issues that need resolving. It is like mains hum. If you a-weight this, a lot of it disappears yet you may very well be hit with a ground loop in some other conditions.

 

[Chocomel]

The data readout shows the values that are plotted on the graph, and includes whatever smoothing they’re doing. Unfortunately they don’t specify the smoothing they use. The purpose of measuring group delay here is to detect areas where the headphone response isn’t minimum phase, as EQ can only really fix minimum phase errors. So you’re looking for sharp upward spikes that indicate group delay in excess of that present in a minimum phase system. Any sort of smoothing will simply flatten these and render the plot useless for this purpose. At higher frequencies amir’s plot is showing a thicket of rapidly-changing delay values that cluster together on the scale shown.

There’s a discussion of this here: https://www.roomeqwizard.com/help/help_en-GB/html/minimumphase.html
This covers speakers, which experience far more group delay than headphones.

Would be easier if amir showed excess group delay if we wanna check for minimum phase. Incidentally here is such a graph for the ananda

https://imgur.com/4p0tfXv

. Measurement done with a hms ii.3 (not by me)

 

[amirm]

Would be easier if amir showed excess group delay if we wanna check for minimum phase.

Unfortunately Audio Precision software doesn't have that capability.