*NEW* AKG K702 has been measured by Oratory!

[Nango]

Hi, like the title says, Oratory measured my K702 I sent him, I've got two pairs and I sent him my old pair from year 2015 (made in China, not Austria). Yes, so Oratory published the result today:
https://www.dropbox.com/s/033kxyf8a0o3iv6/AKG K702.pdf?dl=0
https://www.reddit.com/r/oratory1990/wiki/index/list_of_presets
View attachment 127401

I've tried his EQ & I do like it, previously I was using my own EQ from a measurement by Crinacle.....he uses a GRAS measuring device too, so both Crinacle & Oratory measurements are compatible with the Headphone Harman Curve that we know & love (some hate though!)

Oratory measured my headphone both with the old very worn pads from year 2015 vs new pads which he put on for me, and here's the result of worn vs fresh pads:
View attachment 127402
Oratory advised me to ignore the increased bass roll off of the worn pads, as he said he "only did the quick measurement type, which doesn't give as accurate measurement results below ~50 Hz", so he said to assume for the worn pad bass frequency response it should just track the same shape as the fresh pads in the sub 50Hz, rather than paying any attention in the graph that the bass decrease accelerates faster than the fresh pads...so you should assume that those two lines won't cross over each other in the bass. Yes, so that's not bad, that's quite easy to correct the frequency response for old pads.

Oratory also did some distortion measurements of my K702 at different input voltages, with the maximum level being the equivalent of 104dB, and here's the results, (I have asked if he'd send me through graphs showing measurements at individual levels as the following graph is a bit congested to read, so I might update the graph if he sends me through what I've asked for):
View attachment 127403
So we can see that distortion in the bass is not bad, not an issue. There are a couple of distortion peaks at 1250Hz & 2000Hz, which seem to vary significantly depending on playback volume.


Oratory also measured Channel Matching of my headphone:
View attachment 127404
We can see that channel matching is not perfect here! Oratory did say that our hearing sensitivity in the bass area is quite low so the increasingly large difference in the bass as you go down into the subbass is not as bad as it looks. When I get my headphone back I'll be doing Channel Matching EQ on them so I'll be able to flip channel matching on & off whilst listening to Oratory's EQ, thereby I'll see if I can recognise any benefit of channel matching.

Group Delay:
View attachment 127407
Excess Group Delay:
Not sure on significance of difference to above graph.
View attachment 127440
EDIT: Oratory did say he'd have a go at finetuning the EQ some more, so it's possible that the EQ will be updated in the near future, so I'll edit this post if anything changes. I may also add my own personal tweak(s) I've added to Oratory's EQ. I'll also probably be editing this post to further to put the distortion measurements into better perspective re the dB values rather than the input voltage, and that graph may be updated with seperate graphs done for each level so it's more readable. I'll also try to see how those distortion measurements equate to Amir's standards that he sets, for instance the graph where he draws the +40dB threshold line on his 94dB distortion graph, or this may come to light in the discussions in this thread. I'll also update this post as more is learned through people's comments & analysis in this thread, and I may include a K702 EQ I've done from Crinacle's measurement too, so this may become a K702 EQ thread as well as initially displaying Oratory's measurements which he did just this week.

Finally, I'd like to say thanks very much to Oratory for measuring my headphones & the great work/contributions he makes to the headphone community....nice one Oratory, you're much appreciated!


@samwell7 , @Ravix , @bobbooo , @jhaider , @thewas , @Nango , @Seta Seta Pop , @Jimbob54 , @Bob-23 , tagging you all as you either have the K702 or expressed interested in Oratory's measurement & EQ of them.

Great to see you happy with the outcome!!

 

[Bob-23]

The filters provided by Oratory1990 are for Parametric EQ (PEQ), these are applied in a quite different way from the Graphic EQ (i.e. fixed frequency bands) that you have on your sound card software. You can apply PEQ to your whole system audio with EqualizerAPO on Windows. Adding the PEACE front end UI to eAPO makes things easier.

You can read about equalization and the differences between PEQ and GEQ here: Equalization (audio) - Wikipedia

Equalizer APO download | SourceForge.net
Peace Equalizer, interface Equalizer APO download | SourceForge.net

PEACE setup with Oratory1990's PEQ for AKG371...

View attachment 127439

This results in and EQ graphs like this...

View attachment 127441

Thanks, Robbo! Lookin' forward to trying out the settings for our beloved soundstage monster! And, of course, thanks to Oratory, à la recherche du son perdu.

 

[Robbo99999]

A question for you folks. I'm in the process of trying to put distortion measurements into perspective (and how that relates to distortion in the K702 headphone), and I wanted to work out what is the peak level that I listen at? I think I've figured out a way to work this out by taking into account the following:

• my DAC outputs 2V max
• My JDS Labs Atom headphone amp is unity gain at max volume on Low Gain
• my K702 headphone as measured by Oratory has a sensitivity of 102.2dB per Volt

With the above taken into account I decided to run my JDS Labs Atom headphone amp at max volume on Low Gain (so unity gain), then I dialed down the negative preamp in EqualiserAPO until I reached my loudest subjective listening volume that I use (I was listening to my favourite track for this test). When I did this I found that I had EqualiserAPO on a -24.5dB preamp:

On top of that I also run my windows volume at 79% which equates to another -3.5dBFS, so given the above then that's a total of -28dBFS in relation to the DAC's max voltage of 2V.

So then I wanted to work out what -28dBFS was in voltage when applied to this 2V max of my DAC, I did that with the aid of this website:
https://www.h-ear.audio/2019/06/01/headphone-power-calculator/
First I used the above website to work out what the max theoretical loudness that my DAC could drive my headphone, so 2V applied to my headphone, and as you can see that's 108.2dB as a max theoretical level:

However, earlier I worked out through my listening testing that I'm running at -28dBFS below that level shown above, so 108.2 - 28 = 80.2dB. So this means my max peak listening level at 1kHz is 80.2dB. This equates to the following which is only 0.08V:

So, 80dB and 0.08V (or 80mv) is my max peak listening level which is the second lowest measured level in Oratory's distortion graph that he did on my K702 (the blue line):

I think that puts into perspective the levels I listen at and the associated distortion levels, unless I've gone wrong in my calculations/experimentations here. What do you guys think, have I missed anything here?

EDIT: ah, bass EQ boosts level there by 5dB, so my bass is experiencing whatever distortion is pictured at 85dB which is a tad below the dark green line in Oratory's distortion graph above. Conversely, due to the EQ the 2kHz distortion peak is lowered by another 4dB, which would theoretically take distortion down at that point to whatever level it would be at when run at 76dB, but Oratory didn't measure at such a low level.

EDIT #2: I did the same experiment & calculation with most of my other headphones to work out the max peak level I listen at with each headphone: NAD HP50 = 84.5dB, HifiMan HE4XX = 86.1dB. These two headphones were calculated using the manufacturer claimed headphone sensitivity and not actual measured. So we see a different max peak level that I listen at with each headphone from 80dB for the K702 up to 86dB for the HE4XX....this is either due to inaccuracies of manufacturer reported sensitivity, or inaccuracies in my listening test to determine the feel of the max volume I listen at, or it could be that for some reason I have different tolerances for how loud I like each headphone.....however it does seem that worst case scenario I listen to music at loudest 86dB for a max peak.

 

[solderdude]

That means you do not need any amp at all.
Even the lousiest phone/dongle can easily deliver 0.5V giving you a comfortable 6dB headroom.
80-85 dB peak is well below the level where the recordings are monitored/mixed (around 80dB average) at explaining why you like Harman bass boost.

When you really want to know how loud you listen you would need to measure actual peak voltages.
When you have a PC/laptop that can record you have all the tools needed.

As the K702 has the same voltage efficiency as the HD600/HD650 you can use my assessment 1:1

The part of the recording shown below was playing very softly, think soft level when listening in the evening. The level where you can easily listen for a whole evening while having the urge to turn up the volume a bit.

This resulted in the following (rounded off) numbers:

RMS levels = 23mV = 0.0017mW = 70dB (average)
Peak levels = 245mVPP = 87mVRMS = 0.023mW = 82dB

 

[JIW]

A question for you folks. I'm in the process of trying to put distortion measurements into perspective (and how that relates to distortion in the K702 headphone), and I wanted to work out what is the peak level that I listen at? I think I've figured out a way to work this out by taking into account the following:

• my DAC outputs 2V max
• My JDS Labs Atom headphone amp is unity gain at max volume on Low Gain
• my K702 headphone as measured by Oratory has a sensitivity of 102.2dB per Volt

With the above taken into account I decided to run my JDS Labs Atom headphone amp at max volume on Low Gain (so unity gain), then I dialed down the negative preamp in EqualiserAPO until I reached my loudest subjective listening volume that I use (I was listening to my favourite track for this test). When I did this I found that I had EqualiserAPO on a -24.5dB preamp:
View attachment 127682
On top of that I also run my windows volume at 79% which equates to another -3.5dBFS, so given the above then that's a total of -28dBFS in relation to the DAC's max voltage of 2V.

So then I wanted to work out what -28dBFS was in voltage when applied to this 2V max of my DAC, I did that with the aid of this website:
https://www.h-ear.audio/2019/06/01/headphone-power-calculator/
First I used the above website to work out what the max theoretical loudness that my DAC could drive my headphone, so 2V applied to my headphone, and as you can see that's 108.2dB as a max theoretical level:
View attachment 127683

However, earlier I worked out through my listening testing that I'm running at -28dBFS below that level shown above, so 108.2 - 28 = 80.2dB. So this means my max peak listening level at 1kHz is 80.2dB. This equates to the following which is only 0.08V:
View attachment 127684

So, 80dB and 0.08V (or 80mv) is my max peak listening level which is the second lowest measured level in Oratory's distortion graph that he did on my K702 (the blue line):
View attachment 127685

I think that puts into perspective the levels I listen at and the associated distortion levels, unless I've gone wrong in my calculations/experimentations here. What do you guys think, have I missed anything here?

EDIT: ah, bass EQ boosts level there by 5dB, so my bass is experiencing whatever distortion is pictured at 85dB which is a tad below the dark green line in Oratory's distortion graph above. Conversely, due to the EQ the 2kHz distortion peak is lowered by another 4dB, which would theoretically take distortion down at that point to whatever level it would be at when run at 76dB, but Oratory didn't measure at such a low level.

EDIT #2: I did the same experiment & calculation with most of my other headphones to work out the max peak level I listen at with each headphone: NAD HP50 = 84.5dB, HifiMan HE4XX = 86.1dB. These two headphones were calculated using the manufacturer claimed headphone sensitivity and not actual measured. So we see a different max peak level that I listen at with each headphone from 80dB for the K702 up to 86dB for the HE4XX....this is either due to inaccuracies of manufacturer reported sensitivity, or inaccuracies in my listening test to determine the feel of the max volume I listen at, or it could be that for some reason I have different tolerances for how loud I like each headphone.....however it does seem that worst case scenario I listen to music at loudest 86dB for a max peak.

The peak of a sine is 3 dB higher than the RMS so your peak SPLs are 3 dB higher than you calculate. For instance, the peak SPL of the K702 for a 0 dBFS sine wave at full output from the DAC and unity gain from the amplifier is 111.2 dB.

 

[Robbo99999]

That means you do not need any amp at all.
Even the lousiest phone/dongle can easily deliver 0.5V giving you a comfortable 6dB headroom.
80-85 dB peak is well below the level where the recordings are monitored/mixed (around 80dB average) at explaining why you like Harman bass boost.

When you really want to know how loud you listen you would need to measure actual peak voltages.
When you have a PC/laptop that can record you have all the tools needed.

As the K702 has the same voltage efficiency as the HD600/HD650 you can use my assessment 1:1

View attachment 127727
The part of the recording shown below was playing very softly, think soft level when listening in the evening. The level where you can easily listen for a whole evening while having the urge to turn up the volume a bit.

This resulted in the following (rounded off) numbers:

RMS levels = 23mV = 0.0017mW = 70dB (average)
Peak levels = 245mVPP = 87mVRMS = 0.023mW = 82dB

Yes, you're explaining that you can actually measure the peak voltage to know 100%, and I was showing what my theoretical max peak would be, so there's a difference, but I take it I've made some reasonable assumptions and calculated it correctly for a theoretical max peak?

 

[Robbo99999]

The peak of a sine is 3 dB higher than the RMS so your peak SPLs are 3 dB higher than you calculate. For instance, the peak SPL of the K702 for a 0 dBFS sine wave at full output from the DAC and unity gain from the amplifier is 111.2 dB.

Hmm, ok, I've not heard of that before. That doesn't really tally with Oratory's distortion graph though, where he shows mVrms and the associated frequency response vs dB....as in the mVrms and the shown dB marry up to the calculations, with nothing to show re +3dB you mention, so I'm confused by that.

 

[JIW]

Hmm, ok, I've not heard of that before. That doesn't really tally with Oratory's distortion graph though, where he shows mVrms and the associated frequency response vs dB....as in the mVrms and the shown dB marry up to the calculations, with nothing to show re +3dB you mention, so I'm confused by that.

No, it does not contradict anything Oratory provided to you. The measured SPL corresponds to the RMS voltage but the peak SPL corresponds to the peak voltage, which is 3 dB higher for a sine, i.e. greater by a factor of square root of 2 (https://en.wikipedia.org/wiki/Root_mean_square#In_common_waveforms).

 

[solderdude]

To simplify.
A pure sinewave of 1Vrms has a peak voltage voltage of 1.41V and a peak-to-peak voltage of 2.82V.
So even when you think you reach 1V you will need an amp that can provide a peak voltage of 1.41V (+3dB)

 

[Robbo99999]

No, it does not contradict anything Oratory provided to you. The measured SPL corresponds to the RMS voltage but the peak SPL corresponds to the peak voltage, which is 3 dB higher for a sine, i.e. greater by a factor of square root of 2 (https://en.wikipedia.org/wiki/Root_mean_square#In_common_waveforms).

To simplify.
A pure sinewave of 1Vrms has a peak voltage voltage of 1.41V and a peak-to-peak voltage of 2.82V.
So even when you think you reach 1V you will need an amp that can provide a peak voltage of 1.41V (+3dB)

Ok, although I think the format that Oratory's graph is in, showing mVrms & dB, then that correlates with my calculations.....ie my calculations and the mVrms of the graph are both not taking into account your +3dB. That's my understanding, so when comparing my calculations with the relevant line on the distortion graph then I take it I've done that correctly rather than adding an extra 3dB onto sensitivity values of the headphone in the online calculator.

 

[solderdude]

distortion from headphones is mostly 2nd harm so low enough to be masked.
When you hear 'distortion' in music you can be pretty certain it's in the recording.

 

[Robbo99999]

distortion from headphones is mostly 2nd harm so low enough to be masked.
When you hear 'distortion' in music you can be pretty certain it's in the recording.

(don't bother answering the question then....if in doubt and to make no admissions, state a somewhat related fact)

 

[solderdude]

It is related. The highest distortion peaks at your listening levels are 54dB down. As those are 2nd harmonic and there is something like masking then at 60dB SPL frequencies in that area mask distortions at 10dB because the distortion products are below the 50dB difference.
So when you think you hear distortions at the levels you listen at you can be certain they are in the recording.

I just mentioned it as re-assurance for you so you don't have to worry about the fidelity of your headphone when it concerns HD.

 

[JIW]

Ok, although I think the format that Oratory's graph is in, showing mVrms & dB, then that correlates with my calculations.....ie my calculations and the mVrms of the graph are both not taking into account your +3dB. That's my understanding, so when comparing my calculations with the relevant line on the distortion graph then I take it I've done that correctly rather than adding an extra 3dB onto sensitivity values of the headphone in the online calculator.

The SPL at peak voltage corresponds to that measured for a 3 dB greater RMS voltage so one level higher in the figure. The 3 dB have to be added to the voltage not the sensitivity.

 

[Robbo99999]

It is related. The highest distortion peaks at your listening levels are 54dB down. As those are 2nd harmonic and there is something like masking then at 60dB SPL frequencies in that area mask distortions at 10dB because the distortion products are below the 50dB difference.
So when you think you hear distortions at the levels you listen at you can be certain they are in the recording.

I just mentioned it as re-assurance for you so you don't have to worry about the fidelity of your headphone when it concerns HD.

That's fine, but I just know there's an element of point scoring with your answers, especially when it comes to replies to me.....I don't like it.

The SPL at peak voltage corresponds to that measured for a 3 dB greater RMS voltage so one level higher in the figure. The 3 dB have to be added to the voltage not the sensitivity.
View attachment 127744

It seems like the convention with distortion measurements and frequency measurements is to measure associated with RMS and not peak, as seen in Oratory's distortion graph, and even Amir's measurements fit to that convention, so my calculations that I showed initially are relevant and accurate to that. I'm not sure why we're nitpicking here, I was merely trying to work out and show the approximate dB that I would listen to at 0dBFS, in order to put the distortion graphs into perspective regarding my worst case, and it seems like you're nitpicking on a seperate point, because my understanding is that convention is to display this information in RMS, for instance here note the y-axis in RMS (from one of Amir's reviews, not my K702 of course):

and here:

So I think you're creating an unnecessary diversion from the experimentation & calculations that I did regarding putting into perspective the levels that I listen at and their associated distortion on my K702 headphone. So whilst you might have an additional point I see it as nitpicking and not particularly relevant, so I'd rather not let it clutter up this thread any further.

 

[Fernando]

This is my measurement of the Akg K702 made with my unscientific system with the Audyssey mic and a little deplorabe invent made with expanded polyethylene (white cork)

Sure there are bugs, I'm starting with REW, to name one of the reasons ...

 

[Robbo99999]

This is my measurement of the Akg K702 made with my unscientific system with the Audyssey mic and a little deplorabe invent made with expanded polyethylene (white cork)

View attachment 127803

Sure there are bugs, I'm starting with REW, to name one of the reasons ...

Ha, I guess it's an interesting experiment and fun project.....but you got Oratory's measurement now, and Crinacle's is out there too, both on GRAS devices, so probably better to use those no? Where do you want to end up with your measurements?

EDIT: channel matching looks good, which wasn't the case with mine!

 

[Fernando]

Ha, I guess it's an interesting experiment and fun project.....but you got Oratory's measurement now, and Crinacle's is out there too, both on GRAS devices, so probably better to use those no? Where do you want to end up with your measurements?

EDIT: channel matching looks good, which wasn't the case with mine!

Channel matching really has a bit of a catch, but it's pretty good, less than 1db of difference.
My intention is to fine tune the equalizations, since I have been able to verify that not all headphones of the same model measure the same.
The culprit is an LCD2 that I have and it doesn't work well with any EQ I've found, they sound horrible with them.
Now I have to learn to import the Harman curve and test equalizations, although it works quite well by eye.
I also need to visualize the curve as it appears in the reviews, as sloping upwards, but I have not been able to do it.

 

[JIW]

That's fine, but I just know there's an element of point scoring with your answers, especially when it comes to replies to me.....I don't like it.

It seems like the convention with distortion measurements and frequency measurements is to measure associated with RMS and not peak, as seen in Oratory's distortion graph, and even Amir's measurements fit to that convention, so my calculations that I showed initially are relevant and accurate to that. I'm not sure why we're nitpicking here, I was merely trying to work out and show the approximate dB that I would listen to at 0dBFS, in order to put the distortion graphs into perspective regarding my worst case, and it seems like you're nitpicking on a seperate point, because my understanding is that convention is to display this information in RMS, for instance here note the y-axis in RMS (from one of Amir's reviews, not my K702 of course):
View attachment 127795
and here:
View attachment 127796

So I think you're creating an unnecessary diversion from the experimentation & calculations that I did regarding putting into perspective the levels that I listen at and their associated distortion on my K702 headphone. So whilst you might have an additional point I see it as nitpicking and not particularly relevant, so I'd rather not let it clutter up this thread any further.

You were talking about 'max peak level' which to me reads as the instantaneous SPL corresponding to the highest digital sample value, which is 3 dB higher than the RMS SPL corresponding to a full scale sine wave, the peak of which is the highest digital sample value. I have not said anything about how you should assess the distortion. This was about the peak level alone.

 

[Ron Texas]

Fairly typical needing bass boost and various random peaks in the high end.