Audeze LCD-X Over Ear Open Back Headphone Review

[mitchco]

@Robbo99999 The intent is not to equalise or listen over speakers... The intent is to listen to the binaural recorded music over headphones and listen to the tonal differences between the sonic signature of the headphone relative to the track itself or comparatively to the HP50. Of course, there are HRTF differences, but the comparison is a relative one. And comparing the recorded headphones sonic signature to just the stereo track reveals the signature of the headphone that most people can hear over their own headphones and own HRTF... Just a fun experiment that yielded some interesting sonic results that are consistent with the known sonic signature of the headphone. Have you downloaded the binaural recordings and given a listen over headphones? That was the intent.

 

[BARDUKE]

Man! The morning after I pack up my Mobius and start the return on the Sundaras, I find this review! All of the other reviews online were A+++, but it seems like the general consensus is a NO for the LCD-X. I really wanted some quality open back planars since the Mobius impressed me at first, then my HD660S were so much clearer, I wanted to see what all the hype about open back planars was. Now I might have to reconsider.

 

[KeithPhantom]

Man! The morning after I pack up my Mobius and start the return on the Sundaras, I find this review! All of the other reviews online were A+++, but it seems like the general consensus is a NO for the LCD-X. I really wanted some quality open back planars since the Mobius impressed me at first, then my HD660S were so much clearer, I wanted to see what all the hype about open back planars was. Now I might have to reconsider.

Actually, regular dynamics impress me more than planars at this point. The way you can easily tune them and how they generally measure pretty good (many non-audiophile dynamic headphones have THD and all their components under 1%) is what makes me ask "how good you can do with a really good to excellent dynamic headphone?" Also, planars have some issues trying to stop moving after an impulse (usually, it isn't noticeable), and you can see this in the CSD many planar headphones produce. They have a lot more spurious movement/resonance/ringing than dynamics such as the HD 600. They aren't bad though.

 

[MZKM]

No, I can only generate an impulse response.

Can you show group delay & spectral decay? Showing this in comparison to the Sennheiser will show if there really is any difference in how "fast" it is to respond as well as how quickly it decays.

 

[Robbo99999]

@Robbo99999 The intent is not to equalise or listen over speakers... The intent is to listen to the binaural recorded music over headphones and listen to the tonal differences between the sonic signature of the headphone relative to the track itself or comparatively to the HP50. Of course, there are HRTF differences, but the comparison is a relative one. And comparing the recorded headphones sonic signature to just the stereo track reveals the signature of the headphone that most people can hear over their own headphones and own HRTF... Just a fun experiment that yielded some interesting sonic results that are consistent with the known sonic signature of the headphone. Have you downloaded the binaural recordings and given a listen over headphones? That was the intent.

I'll have a listen to them tomorrow, just got back from work, no time for contemplative listening! I think there might be some value in listening to your binaural recordings if I had some in-ear headphones that would sit in my ears at roughly the same depth as your microphones you used for your binaural recordings, but if you playback the binaural recordings on regular over ear headphones then you're adding another layer of "HRTF" on top of your own (mitchco's) HRTF that you recorded during your binaural recordings (not to mention your HRTF is also different to other peoples), so I imagine that to be quite a "messed up" situation so I doubt what can be derived from any comparisons you've recorded HP50 vs LCD-4z. But I will have a listen to them tomorrow.

(I found it interesting what you were saying in your original post about convolution & how it can be implemented in quite a flexible way regarding "Q-values" as long as you're using the right software, what software do you use?).

EDIT: to try and get my terminology more accurate I should probably replace all instances of "HRTF" above that I mentioned with "HpTF" (headphone transfer function), as there's a difference and we're talking about the latter.

 

[Theriverlethe]

Can you show group delay & spectral decay? Showing this in comparison to the Sennheiser will show if there really is any difference in how "fast" it is to respond as well as how quickly it decays.

It seems pretty unlikely that any headphone would have perceptible differences in group delay or decay time. Room decay time is usually measured in hundreds of milliseconds to upwards of a second. I've never seen spectral decay plots of a headphone showing more than ten milliseconds.

 

[MZKM]

It seems pretty unlikely that any headphone would have perceptible differences in group delay or decay time. Room decay time is usually measured in hundreds of milliseconds to upwards of a second. I've never seen spectral decay plots of a headphone showing more than ten milliseconds.

Which is why I don’t believe these sound audibly “faster” than regular headphones; but, since those measurements show how a driver behaves over time, that’ll satisfy those wanting to know for sure if it is actually “faster” or if it’s like Amir said and it sounds like that simply due to the lacking of bass.

 

[KeithPhantom]

Which is why I don’t believe these sound audibly “faster” than regular headphones; but, since those measurements show how a driver behaves over time, that’ll satisfy those wanting to know for sure if it is actually “faster” or if it’s like Amir said and it sounds like that simply due to the lacking of bass.

Even having the times, they are usually too short to even make a difference. I think what audiophiles call "speed" is FR-dependant. An experiment with EQ should be enough to give us some kind of clue.

 

[Theriverlethe]

Even having the times, they are usually too short to even make a difference. I think what audiophiles call "speed" is FR-dependant. An experiment with EQ should be enough to give us some kind of clue.

I suspect the lack of distortion could make the bass sound “tighter,” even if it’s EQ’ed to match the Harman curve.

 

[KeithPhantom]

I suspect the lack of distortion could make the bass sound “tighter,” even if it’s EQ’ed to match the Harman curve.

That's something I would like to test.

 

[BYRTT]

.....(I found it interesting what you were saying in your original post about convolution & how it can be implemented in quite a flexible way regarding "Q-values" as long as you're using the right software, what software do you use?).....

Not im mitchco but below two free program is flexible and can shift any minimum phase filters to linear phase and vice versa plus they offer pure phase filter manipulation too, but they havent any automation as so many seems to like in REW's EQ menu and one needs to know what one is doing as with most other crafts, for inspiration i traced your raw HD650 curve plus target in one the programs below and set your 10x filters to toggle into animations.

Rephase..

VituixCAD, tip to create a convolution file create any filter correction be it minimum phase linear pase pure phase or a mix, then menu "File/Export/Filter response of driver", now open another new program session (for both VituixCAD and Rephase they can be open in multiples) and load the exported filter curve to one transducer and in menu "View/Impulse response" dialog window adjust the output settings..

 

[Robbo99999]

Not im mitchco but below two free program is flexible and can shift any minimum phase filters to linear phase and vice versa plus they offer pure phase filter manipulation too, but they havent any automation as so many seems to like in REW and one needs to know what one is doing as with most other crafts, for inspiration i traced your raw HD650 curve plus target in one the programs below and set your 10x filters to toggle into animations.

Rephase..
View attachment 88948


VituixCAD, tip to create a convolution file create any filter correction be it minimum phase linear pase pure phase or a mix, then menu "File/Export/Filter response of driver", now open another new program session (for both VituixCAD and Rephase they can be open in multiples) and load the exported filter curve to one transducer and in menu "View/Impulse response" dialog window adjust the output settings..
View attachment 88949

The fact that they don't have any automated filter creation doesn't bother me, as when I do headphone EQ's in REW I just do them all manually, so I've learned over time to judge quite well what kind of Q I need for a filter, etc. It might be worth my while to have a bit of a play with those two pieces of software, often the best way to learn. Am I right in thinking I could use Rephase to correct phase in a headphone, whereas that can't be done using normal parametric filters? Isn't there supposed to be some sound quality benefits of that? I'd have to do some research on it. I mean I'm perfectly happy using REW & parametric filters to correct headphones and speakers, but if there's some sound quality benefits of using Rephase & getting involved with phase then that's something I'd learn how to use, but is there a benefit?

 

[BYRTT]

The fact that they don't have any automated filter creation doesn't bother me, as when I do headphone EQ's in REW I just do them all manually, so I've learned over time to judge quite well what kind of Q I need for a filter, etc. It might be worth my while to have a bit of a play with those two pieces of software, often the best way to learn. Am I right in thinking I could use Rephase to correct phase in a headphone, whereas that can't be done using normal parametric filters? Isn't there supposed to be some sound quality benefits of that? I'd have to do some research on it. I mean I'm perfectly happy using REW & parametric filters to correct headphones and speakers, but if there's some sound quality benefits of using Rephase & getting involved with phase then that's something I'd learn how to use, but is there a benefit?

One can correct any phase for heaphones into Rephase, trouble is get a measurement chain calibrated to really tell the truth, that is stuff as USB interfaces can give some trouble, chain needs use dual channel loopback to adjust for clock and never use some I/O that dangle on its own seperate clock generator, I/O needs sync to same clock. Now say that measurement chain bench phase correct then yes to phase filter correction should give sound quality benefits for single transducer head phones provided there are anything to come for, but for example LCD-X is of Planar Magnetic and could be a 2-wayer that crossover at the 5kHz blib seen in impedance charts and if that is case then its impossible correct the phase mismatch at crossover point downstream of the acoustic sum, for that case one need to get upstream of the two summing band passes and first correct slopes if theres any imperfection then correct for delay if theres any imperfect acoustic center mismatch. There is equalizer software out there that can dial on pure phase domain but they not realtime in it takes time to correct for time mismatch and therefore convolution method is often used as solution to manipulate in time domain and its not so hard to output multiple different convolution corrrections to judge sound (A/B) using those different filter string settings.

A sidenote is REW actual can output convolution files its just than it need files with pure impulse response to create them where a txt-file response curve cant be output as wav-file, as example if you highlight the measurement with your EQ filter string and in menu "File/Export/Export filters impulse response as wav" get one the convolution file of the EQ string but it cant do phase manipulations except one import some phase manipulation from Rephase or other program as wav-file and do the (A*B) command on "ALL SPL" tab in "Controls/Trace arithmetic" and then export that sum of (A*B) as wav-file.

 

[Sanlitun]

I have the original 2013 driver LCD-X and love them.

Interesting test results.

 

[Robbo99999]

One can correct any phase for heaphones into Rephase, trouble is get a measurement chain calibrated to really tell the truth, that is stuff as USB interfaces can give some trouble, chain needs use dual channel loopback to adjust for clock and never use some I/O that dangle on its own seperate clock generator, I/O needs sync to same clock. Now say that measurement chain bench phase correct then yes to phase filter correction should give sound quality benefits for single transducer head phones provided there are anything to come for, but for example LCD-X is of Planar Magnetic and could be a 2-wayer that crossover at the 5kHz blib seen in impedance charts and if that is case then its impossible correct the phase mismatch at crossover point downstream of the acoustic sum, for that case one need to get upstream of the two summing band passes and first correct slopes if theres any imperfection then correct for delay if theres any imperfect acoustic center mismatch. There is equalizer software out there that can dial on pure phase domain but they not realtime in it takes time to correct for time mismatch and therefore convolution method is often used as solution to manipulate in time domain and its not so hard to output multiple different convolution corrrections to judge sound (A/B) using those different filter string settings.

A sidenote is REW actual can output convolution files its just than it need files with pure impulse response to create them where a txt-file response curve cant be output as wav-file, as example if you highlight the measurement with your EQ filter string and in menu "File/Export/Export filters impulse response as wav" get one the convolution file of the EQ string but it cant do phase manipulations except one import some phase manipulation from Rephase or other program as wav-file and do the (A*B) command on "ALL SPL" tab in "Controls/Trace arithmetic" and then export that sum of (A*B) as wav-file.

Ah, well I understand some of that, but I think it's because I don't fully understand Phase. So with the measurements that Amir is providing, is it possible to use that information to correct phase in headphones, or do we need different measurements, or is his measurement equipment not able to provide the right information? I think you mentioned that Phase can't be inferred or corrected from a text file frequency response (which is what we see from Amir & Oratory right now?)? I probably don't know enough about Phase to be asking you questions on this!

EDIT: Actually, thinking about it (in a nutshell), could I for example take my parametric filters from say my HD600 EQ and use those programs to correct for phase given that I know the parametric filters (which I created manually in REW) and have a text file frequency response of the stock headphone from Oratory via the Jaakopasanen website, or I need other measurements? If I can then I'll definitely have a play with the software and see if I can work it out.

EDIT #2: And how is the final Phase Corrected EQ implemented in day to day usage of the headphone, can it be implemented by Equaliser APO or you need other software? Is it implemented as a convolution file or other type/name?

 

[BYRTT]

Ah, well I understand some of that, but I think it's because I don't fully understand Phase. So with the measurements that Amir is providing, is it possible to use that information to correct phase in headphones, or do we need different measurements, or is his measurement equipment not able to provide the right information? I think you mentioned that Phase can't be inferred or corrected from a text file frequency response (which is what we see from Amir & Oratory right now?)? I probably don't know enough about Phase to be asking you questions on this!

About understand how minimum phase works think one can get a good feel open up Rephase and on "Minimum-Phase Filters" tab try set a HP filter as 1st order then 2nd order 4th order etc and see how per order out of band is a 90º degree rotation per order and in band or say at the dialed in frequency knee its a 45º rotation, now try set 2nd order BW HP @20Hz and 2nd order BW LP 20kHz and see how phase is behaved for a perfect audio domain passband, that exercise should be understandable but its still hard to grasp the right phase curve when there is havoc in amplitude domain but a good rule is whenever one see wiggles in magnitude domain there should also be wiggles in phase domain and if not we have interference that could be reflections or misalligned acoustic centers in allpass summing systems (multi-wayer). Also as exercise its possible export whatever passbands in Rephase as 64bit IEEE wav-file and read them in REW to study how phase/time domain looks into a waterfall or a wavelet, waterfall present only the time domain after time zero where wavelet can present before and after and can include group delay curve, if one create and use target curves or whatever reference curves into Rephase to compare to live measurements then remember always use the same sample rates as measurement chain.

Files Amir shared for LCD-X was without the phase column (same as linear phase) so we cant extract if there is mismatch in time domain for LCD-X cans, his measurement gear is first class so cant imagine it cant provide the right information but imagine it takes too much work in software for Amir to export that information to shared files and in that regard think its okay thinking of how much time he use on publish graphs and text into his acoustic reviews.

Think the thing about txt-file was that REW cant output or export a wav-file from a txt-file, but VituixCAD can do the trick as seen in the text above the animation for VituixCAD.

 

[Robbo99999]

About understand how minimum phase works think one can get a good feel open up Rephase and on "Minimum-Phase Filters" tab try set a HP filter as 1st order then 2nd order 4th order etc and see how per order out of band is a 90º degree rotation per order and in band or say at the dialed in frequency knee its a 45º rotation, now try set 2nd order BW HP @20Hz and 2nd order BW LP 20kHz and see how phase is behaved for a perfect audio domain passband, that exercise should be understandable but its still hard to grasp the right phase curve when there is havoc in amplitude domain but a good rule is whenever one see wiggles in magnitude domain there should also be wiggles in phase domain and if not we have interference that could be reflections or misalligned acoustic centers in allpass summing systems (multi-wayer). Also as exercise its possible export whatever passbands in Rephase as 64bit IEEE wav-file and read them in REW to study how phase/time domain looks into a waterfall or a wavelet, waterfall present only the time domain after time zero where wavelet can present before and after and can include group delay curve, if one create and use target curves or whatever reference curves into Rephase to compare to live measurements then remember always use the same sample rates as measurement chain.

Files Amir shared for LCD-X was without the phase column (same as linear phase) so we cant extract if there is mismatch in time domain for LCD-X cans, his measurement gear is first class so cant imagine it cant provide the right information but imagine it takes too much work in software for Amir to export that information to shared files and in that regard think its okay thinking of how much time he use on publish graphs and text into his acoustic reviews.

Think the thing about txt-file was that REW cant output or export a wav-file from a txt-file, but VituixCAD can do the trick as seen in the text above the animation for VituixCAD.

Ah, so we need the Phase Column to be included in the measurement before we can even start thinking about correcting for phase. So in this case standard parametric filters are more applicable until we can see measurements from Amir (or others) that include the Phase Column.

 

[Chocomel]

Headphones are usually minimum phase systems so correcting the FR will also correct the phase response. That's assuming you use minimum phase EQ of course. That's also why i would advice against using linear phase EQ for headphones.

For some examples of the excess group delay of some headphones.

https://imgur.com/a/FGvIldH

As you can see the headphones are all basically minimum phase throughout the audible range, be it open or closed, planar or Dynamic. There's two features you'll probably have noticed to, in the lowest frequency there's some deviation. This is caused by noise and with higher level Measurements (or lower noise) you would see a reduction in that. The other is some wiggling in the top octave. I'm not entirely sure what causes that but considering this Measurement rig (a hms ii.3 ) isn't accurate there anyway and headphone Measurements there being troublesome in general i wouldn't read anything into personally

There's some cases where headphones will deviate from the minimum phase behavior, for example the monoprice m1060 around 4-5k, but even then it's for just a part of the frequency range. For those cases however it can be nice to also have a look into decay graphs, CSD still sucks then though and personally prefer something like burst decay.

Excess group delay of a m1060 https://cdn.discordapp.com/attachments/392997158283902980/620264684666355750/unknown-16.png

And it's burst decay
https://cdn.discordapp.com/attachments/392997158283902980/768768254000234496/unknown-123.png

 

[50°F=Hellfire]

BTW the only Audeze headphones that I found to have an acceptable sound (among those I heard) are the LCD-XC (not speaking of other issues like weight...).

I've had XC's for a while, and they seem equal in almost all things, with slightly more pronounced treble (which doesn't offend me, but I know that most people's d*cks fall of at the mere thought of frequencies above 4hz having the audacity to exist) and much larger soundstage than my Ether CX's (which Amir loves). They also feel more comfortable due to the thicker pads, but that's mostly a "me" thing, because my right ear sticks out rather far and ends up touching the drivers of most headphones. And the weight is actually a plus for me, because I'm autistic and these feel like the headphone equivalent of a weighted blanket.

 

[zachary80]

Audeze LCD headphones non-EQed preference scores:

81 - LCD-2 Closed (crinacle)​

76 - LCD-2 Closed (oratory1990)​

74 - LCD-1 (oratory1990)​

71 - LCD-XC (crinacle)​

70 - LCD-2 (oratory1990)​

70 - LCD-2 Fazor (oratory1990)​

65 - LCD-1 (crinacle)​

52 - LCD-4 (oratory1990)​

51 - LCD-2 Classic (oratory1990)​

50 - LCD-X (oratory1990)​

48 - LCD-2 Classic (crinacle)​

40 - LCD-X (crinacle)​

31 - LCD-3 (crinacle) I have these... and rather like them.​

Scores: https://github.com/jaakkopasanen/AutoEq/blob/master/results/RANKING.md

Martin

​

This is interesting to me because, despite liking to EQ my headphones to Harman target and owning multiple from the list, it is almost completely unaligned with my subjective experiences. At which point I'm either wrong, or frequency response isn't the whole story, and honestly it would be a lot easier if we could just all buy PSB M4U 8 and be done with it.