MiniDSP EARS: First Measurements

[MRC01]

We have started a MiniDSP EARS on its way round-robin for people to try it out then mail to the next person. Details and sign-ups here: https://audiosciencereview.com/forum/index.php?threads/wanted-minidsp-ears.26597/post-959400

I've shared my initial measurements of vintage Sennheiser HD-580 and Audeze LCD-2F headphones. The measurements are generally in the right ballpark, but some look a bit "off", not sure whether it's the device or how I'm using it, your advice is welcome:

Wanted: MiniDSP Ears

Hey folks, I sold my UMIK-1 to a member here and it all went smoothly, so I figured I'd see if anyone wants to get rid of a MiniDSP Ears. It seems the kind of thing you might buy and use to measure & tune your personal headphones, then no longer have a need for it. At least that is how I want to...

audiosciencereview.com

I also measured some IEMs I had lying around: VSonic VSD1S, Vsonic GR07 and Etymotic ER6. I'll post those soon. Teaser: the HRTF correction curves for these seem to be more challenging than over-the-ear headphones.

 

[MRC01]

Next are the IEM measurements. Frequency responses are in order: raw (corrected for mic only), HEQ correction (for over-the-ear headphones, same as used for the LCD-2F and HD-580), and IDF (which MiniDSP uses for in-ear monitors). The manual for the MiniDSP EARS says they used the raw response curve of the Etymotic ER4 as the standard for their IEM curve. I measured each twice, completely removing them from the EARS head and reinstalling them in between.

Here's the Etymotic ER6. I bought these for about $100 15-20 years ago.
Raw

HEQ

IDF

The channel matching is decent, almost within the margin for error after I removed & reinstalled them on the head.
Distortion is quite low. Again, the 120 Hz spike is my computer fan.

Here's the VSonic GR07. I bought these for about $100 about 6 years ago.
Raw

HEQ

IDF

These had the best channel matching and consistency. Their response is quite smooth, but it rises a bit later than Etymotics, so the IDF curve which is based on the Etymotic house curve punishes them.
GR07 Distortion:

The GR07 have nice low distortion too. Almost as low as the Audeze LCD-2F! With the great channel matching, smooth response and low distortion, these are a deal for $100.

Now for the VSonic VSD1S. I bought these for less than $50 about 3-4 years ago.
Raw

HEQ

IDF

It looks like the R side never sealed. But I tried it several times, not just twice, and it always had this lack of bass response. So it is either defective, or this is just extremely poor channel matching. They're cheap, maybe you get what you pay for?

Distortion, I measured twice because the R channel had some wonky response:
Left looks quite good, low distortion:

Right also has low distortion, but has some wonky behavior around 700 - 2500 Hz (still below 1% / -40 dB). This came up in every measurement.

Conclusion: of these IEMs, the GR07 measured the best. And subjectively, they sound the best. Frequency response for headphones is always subjective, what HRTF curve should one use? Every person is different both physically and subjectively. With IEMs the proper curve is different, and this challenge seems even harder.

 

[MRC01]

If anyone wants to borrow this MiniDSP EARS unit and do your own testing, post in the other thread and join the round-robin.

Wanted: MiniDSP Ears

Hey folks, I sold my UMIK-1 to a member here and it all went smoothly, so I figured I'd see if anyone wants to get rid of a MiniDSP Ears. It seems the kind of thing you might buy and use to measure & tune your personal headphones, then no longer have a need for it. At least that is how I want to...

audiosciencereview.com

 

[MRC01]

This MiniDSP EARS is a neat device that's easy to use. I'm surprised that with all the curious gearheads here nobody wants to give it a test drive on their own headphones.
If you do, all it will cost is postage. Reply and let me know before Thanksgiving. Details here.

 

[Robbo99999]

F*cking A! I was just toying with buying one. What's the postage to the UK??? My main motivation is to use it for balancing left & right channels throughout the frequency range for the headphones I own. I may also try using at as a means to establish my own preferred target curve.

Also, what's been your best usages for this device so far? I don't think it can replace GRAS & the Harman Curve, but I'm planning to use it for L/R matching and perhaps establishing my own target curve based on headphones that have been EQ'd to the Harman Curve & then tweaked by me, which I would then measure on this device as a means of me setting an EARs target curve for myself for EQ'ing other headphones.

EDIT: just ordered one, doesn't cost more than a mid range headphone and could be useful for the future.

 

[Keith_W]

I found this topic via a search. Perhaps we should send one of these units to Amir for comparison to his B&K setup. If we measure a few headphones on both setups, we will be able to compare the two against the reference to see how far off it is.

 

[SIY]

I did a short article on this fixture. Bottom line is that it's a fun toy and can be useful for some specific relative measurements. But it is not comparable to a true headphone test fixture, and the frequency response is dominated by the geometry of the straight cylindrical "ear canal" to the microphones.

 

[Robbo99999]

I did a short article on this fixture. Bottom line is that it's a fun toy and can be useful for some specific relative measurements. But it is not comparable to a true headphone test fixture, and the frequency response is dominated by the geometry of the straight cylindrical "ear canal" to the microphones.

I've got one, I don't see it as a toy. It's 100% valid for creating channel balancing EQ's through throughout the frequency range. It can also be useful for knowing what levels you're listening to, and for matching SPL levels if comparing headphones. Distortion measurements can be useful and interesting when comparing between headphones/units. It can also be used to determine unit to unit variation. Has some limited use if you create a miniDSP to GRAS conversion curve, but it's not always that accurate as it will be a different conversion curve for each model of headphone (not universally applicable between different headphone models) and you always have the element of unit to unit variation that muddies the conversion if you're comparing measurements you find on the internet (Oratory/Resolve/Crinacle/Amir), but you can get round that by measuring the same unit on a GRAS and then measuring it on a miniDSP EARS, but you're still left with a difference of how the measurer did the measurements (their measurement protocol) which is an inaccuracy of unknown quantity. I've found the miniDSP EARS very useful, & most certainly for the bit I put in italics at the start of the post.

EDIT: ah, forgot to mention, it's also extremely useful for choosing aftermarket pads, as you can see the effect on the frequency response vs your stock pads, and therefore you can derive compensatory EQ's so that your chosen aftermarket pads don't negatively effect the sound vs stock.

EDIT#2: I've also found it exceedingly easy to get reproducible/repeatable measurements on the miniDSP EARS, so the precision from measurement to measurement when picking up & replacing the same headphone on the fixture is very high & reliable.....bottom line I've found it very easy to get reliable measurements from it with my various headphones.

 

[asrUser]

I've got one, I don't see it as a toy. It's 100% valid for creating channel balancing EQ's through throughout the frequency range. It can also be useful for knowing what levels you're listening to, and for matching SPL levels if comparing headphones. Distortion measurements can be useful and interesting when comparing between headphones/units. It can also be used to determine unit to unit variation. Has some limited use if you create a miniDSP to GRAS conversion curve, but it's not always that accurate as it will be a different conversion curve for each model of headphone (not universally applicable between different headphone models) and you always have the element of unit to unit variation that muddies the conversion if you're comparing measurements you find on the internet (Oratory/Resolve/Crinacle/Amir), but you can get round that by measuring the same unit on a GRAS and then measuring it on a miniDSP EARS, but you're still left with a difference of how the measurer did the measurements (their measurement protocol) which is an inaccuracy of unknown quantity. I've found the miniDSP EARS very useful, & most certainly for the bit I put in italics at the start of the post.

Your post is really helpful. Could you please elaborate how to do the miniDSP to GRAS conversion curve? Which guides could you recommend to get the EARS kicking?

 

[Robbo99999]

Your post is really helpful. Could you please elaborate how to do the miniDSP to GRAS conversion curve? Which guides could you recommend to get the EARS kicking?

Oratory measured one of my K702 on his GRAS rig, I then measured it on my miniDSP EARS rig. I then used REW to EQ the miniDSP measurement to the Oratory measurement - that set of EQ filters is what turns a K702 miniDSP measurement into a GRAS measurement. It will be a different conversion curve for each headphone model. I did a similar thing converting HD560s measurements, and I decided to use that conversion curve on some of my other headphones as it just seems like a more "generic" headphone design with the HD560s vs the K702, so I think would be more applicable to other models, but really it's only most useful in terms of conversion curves if you have the same unit measured on GRAS and then measured on miniDSP, and then it's applicable just for that model of headphone. The most bonafide uses for miniDSP EARS will be for making comparisons within a headphone model - the affect of different pads and the channel matching - creating EQ's to match the channels, creating EQ's to allow use of aftermarket pads without changing the sound. The conversion curve stuff is not it's prime usage in my opinion, unless you have a unit measured on GRAS & then on miniDSP....and as I said only valid for that model of headphone.

REW is a great tool for EQ'ing, you can look up some Youtube vids to get you started, and it'll probably be based around room correction for speakers those vids......but with some creativity & imagination once you know how all the parts of REW fit together then you can use it to help you EQ any headphone measurement to any target curve you want.

 

[asrUser]

creating EQ's to match the channels

Well, up to now I just applied Eq settings to both channels at once. I didn't see an application of matching the channels yet and don't know how to do this. Is this advanced use?

When I read your text I think that even other earpads could mess the conversion curve up for 1 specific headphone model, couldn't it? So I think it's not worth the time and I'd be better off to just measure with the EARS as it is. But before that calibrating it according to the manual.

 

[Robbo99999]

Well, up to now I just applied Eq settings to both channels at once. I didn't see an application of matching the channels yet and don't know how to do this. Is this advanced use?

When I read your text I think that even other earpads could mess the conversion curve up for 1 specific headphone model, couldn't it? So I think it's not worth the time and I'd be better off to just measure with the EARS as it is. But before that calibrating it according to the manual.

You don't need to calibrate the EARS, it already comes with calibration files for that unit of EARS rig, which calibrate the mics - so you just have to use the cal files supplied by miniDSP.

Depending what music player or EQ system you use to implement your EQ to the music, then a few of the players or EqualiserAPO (being system-wide) allows you to apply EQ per channel - so that's how you EQ each channel to match - with a different EQ that matches each channel to the average of the two channels (that's one way of doing it).

Yes, if you choose earpads that differ to a large extent in their design vs stock, then that could throw off the conversion curve and make it no longer as valid. If you choose a pad which is similar then I think the conversion curve will still hold true and you can use the measured differences to EQ out the differences - thereby making sure your aftermarket pad is not affecting the sound of the headphone from stock.

 

[Keith_W]

@amirm would it be possible for this unit to be sent to him for review. I would be really interested to see if we can generate a transfer function to convert this unit's measurements to something industry standard.

 

[solderdude]

According to Oratory this cannot be done.
Only for 1 specific headphone this is possible.
The ear simulator is very different from the various standards that exist.
The thought is nice though, similar to saying... I buy a HD201 apply some EQ and turn it into an HE1.
Would be fun if you could buy a cheap soundcard and turn it into an AP analyser as well.

The EARS can be used for some things like getting an idea of the tonality (with the proper compensation/target) or to see what changes occur when modifying headphones or check for production tolerances perhaps so it does have some worth.

 

[Keith_W]

The thought is nice though, similar to saying... I buy a HD201 apply some EQ and turn it into an HE1.
Would be fun if you could buy a cheap soundcard and turn it into an AP analyser as well.

I get what you are saying. But then Bob Carver did prove that if you know the transfer function of a speaker-room system you can replicate it to the extent that professional audio reviewers at the time could not tell the difference in a double blind test. And apparently, BACCH can do the same to the extent that you are unsure if you are listening to the system or the headphone.

 

[solderdude]

Yes, but the thing is that in hearing there are much larger tolerances allowed than in measurements.
There is a difference between adding some extra output resistance and making a bandwidth of an amp smaller to fit that of another amp (what Bob basically did) and not changing distortion profiles and levels (which is not needed to make amps sound similar anyway) as the hearing is not that discriminative.
In other words the hearing is not even remotely close to being a measurement device and easy to fool as well.

You simply can not turn the cheap EARS into a GRASS measurement system but both can produce a squigly that, by changing the correction curve, can be similar on one headphone but will differ on another headphone (so not be a GRASS for a lot of other headphones). But... probably close enough to be able to create a decent 'starting point' to create EQ to. This was the idea behind the miniDSP b.t.w.

 

[thewas]

But then Bob Carver did prove that if you know the transfer function of a speaker-room system you can replicate it to the extent that professional audio reviewers at the time could not tell the difference in a double blind test.

Didn't Carver only prove that for the amplifier part which is more trivial as its "single dimensional" while loudspeakers/headphones/room/ear interaction is a 3D problem?

 

[SIY]

Didn't Carver only prove that for the amplifier part which is more trivial as its "single dimensional" while loudspeakers/headphones/room/ear interaction is a 3D problem?

He didn't even do that. This was a marketing campaign for a line of amps he was making then. He roped magazines into some badly controlled demos, and the legend has taken root.

 

[Keith_W]

Didn't Carver only prove that for the amplifier part which is more trivial as its "single dimensional" while loudspeakers/headphones/room/ear interaction is a 3D problem?

My apologies, you are correct. He showed that his amplifier was indistinguishable from any other amplifier as long as the transfer function of the other amplifier was known.

 

[Robbo99999]

@amirm would it be possible for this unit to be sent to him for review. I would be really interested to see if we can generate a transfer function to convert this unit's measurements to something industry standard.

Solderdude already answered you, but I'll word the same thing in a different way: you can do this but it would be a different conversion curve for each different model of headphone, so it's not like you can create a generic miniDSP EARS to GRAS conversion curve for all headphones, but you could for example do one for a K702, and then do another for the HD560s - both of the conversion curves would be different - so that's the limitation.