• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Measurement of applied EQ question?

Jeromeof

Addicted to Fun and Learning
Forum Donor
Joined
Jul 25, 2021
Messages
616
Likes
965
Location
Ireland
So, I am 95% in the objective camp and 5% in the subjective camp as far as Measurements are concerned but I have 2 thoughts / concerns and apologies if they have been discussed previously.

My main concern is around what I might call " measurements loopback" (i.e. measuring and comparing the EQ/DSP changes that we might apply multiple times to make on bit of audio gear sound like another) and this question is sort of inspired by the Truthear Review, basically early on ( probably 10+ years ago now) , I was a heavy adopter of Android mainly because of the Viper4Android and the promise I could turn my relativity cheap IEM's into something sounding like an expensive pair using low level (system wide at the time) DSP which would first normalise my IEM (assuming I had a reasonable consistent model - i.e. little manufacturer variance) into a much more expensive sounding device - so I could turn them into whatever I fancied that day (e.g. want a bass boost - let them sound "exactly" like a Beats headphone today). This was Nerdy fun and useful at the time. But we all grow up :) and it is only in recent months have got the time to get back to this "hobby", so my interest in the true capabilities of EQ/ DSP and measurements have re-emerged, so....

Anyway the real "concern" from my technical brain is now that we can properly measure the audio output of headphones and we all happy that a relatively cheap pair (I am still waiting for my Thruthear's to arrive) can "measure" sound wise like a crazy expensive pair of IEM's. Now with IEM's (especially as there is no 'room' reflections etc) the DSP itself can be scientifically verified cheaply (we don't need room's like Amir or Erin)? I.e. measure a $1000 IEM and compare it with the Thruthear with the EQ set to match the $1000 pair - or alternatively normalise both pairs to the Harman standard and compare?? Mostly, I am guessing that people will still prefer the expensive pair as some things will still be subjective ("the quality of the material of the cables") but I am still thinking that for the extra money it provides, something from the $1000 IEM will cause it to be "better" than the $50 can provide? So what is that? Can that be measured?? With Speakers, I always felt there were too many variances (around the room and the analog / noise introduced into the flow) but headphones and DSP can now verified easily I believe.

The other thought / concern is similar - but based on the excellent video by @Amir
it sort of concerns me about smoothing - so at lower frequencies we can distinguish between different frequencies easier - but then using the same smoothing options like 1/3 maybe 1/6 seems a bit too broad at these frequencies, like we smoothing over these variations at these lower / mid frequencies that might make a difference - this would in my opinion sort of be similar to imagine using a photo "smoothing" algorithm like the "beauty filters" that can be applied to say instagram photos -( they remove details to provide perceived perfection) - so we might be smoothing out the "detail" to make the sound match the generally perceived average "beauty" for that frequency - I am not necessarily saying its wrong just we might be de-emphasising detail in a very specific frequency for a particular speaker that to make that speaker sound more "average" - I know (as I have also watched) the various videos about the Harmon Curve and the methodology of blind testing speakers - but my worry is that is mostly "over" averaging to end up with most people's example of bland beauty can be destructive in detail.

Anyway, being Irish - I would invite anyone willing to a few pints in a nice pub in Dublin to discuss these concerns over some of the "black stuff".
 
Last edited:

fpitas

Master Contributor
Forum Donor
Joined
Jul 7, 2022
Messages
9,885
Likes
14,191
Location
Northern Virginia, USA

fpitas

Master Contributor
Forum Donor
Joined
Jul 7, 2022
Messages
9,885
Likes
14,191
Location
Northern Virginia, USA
To the OP's point about speakers (since I've never even seen an IEM)...I suppose you are making the speaker "average". The alternative is that it has a sound of its own. That's fine for musical instruments, but speakers are supposed to be neutral.
 
OP
Jeromeof

Jeromeof

Addicted to Fun and Learning
Forum Donor
Joined
Jul 25, 2021
Messages
616
Likes
965
Location
Ireland
Well the question(s) were more less that we ca now change the sound of IEMs especially ( but also speakers ) using DSP to normalise them to a standard curve and potentially make them sound like any other measured IEM. If that is the case what is the subjective thing that makes the difference between a $50 IEM and a $1000 IEM !!

With speakers the question is sort of asking what is the difference ( after proper EQ applied ) between say a good quality $400 speaker and an $1000 speaker and I was wondering is that difference related to loosing some measured “detail” when we smooth curves as part of the process of measuring and comparing speakers
 

fpitas

Master Contributor
Forum Donor
Joined
Jul 7, 2022
Messages
9,885
Likes
14,191
Location
Northern Virginia, USA
Well the question(s) were more less that we ca now change the sound of IEMs especially ( but also speakers ) using DSP to normalise them to a standard curve and potentially make them sound like any other measured IEM. If that is the case what is the subjective thing that makes the difference between a $50 IEM and a $1000 IEM !!

With speakers the question is sort of asking what is the difference ( after proper EQ applied ) between say a good quality $400 speaker and an $1000 speaker and I was wondering is that difference related to loosing some measured “detail” when we smooth curves as part of the process of measuring and comparing speakers
As I said, I have no idea about IEMs. I'm sure the IEM crowd will be here soon.

But as to detail lost or gained by EQing a speaker. My experience has been the flatter I get mine, the more revealing they are. I guess on one particular track, a peak might help bring out an instrument or voice etc. Usually that's at the cost of sounding bad on other tracks.
 
OP
Jeromeof

Jeromeof

Addicted to Fun and Learning
Forum Donor
Joined
Jul 25, 2021
Messages
616
Likes
965
Location
Ireland
As I said, I have no idea about IEMs. I'm sure the IEM crowd will be here soon.

But as to detail lost or gained by EQing a speaker. My experience has been the flatter I get mine, the more revealing they are. I guess on one particular track, a peak might help bring out an instrument or voice etc. Usually that's at the cost of sounding bad on other tracks.
Yes, I have a similar experience with a "flat" response - and I agree that possible for a given track a certain instrument might be emphases and that is sort of my 'concern' - if say a famous perfectly tuned instrument that might have been used in various classic recordings (think a Stratocaster or Fender guitar), that instrument might use common very narrow frequencies on particular cords. If those details were smoothed out too much by EQ - this might explain why it is very common to see people wax lyrical about particular 'high' end speakers while the same classic recordings might not sound great on other speakers with exactly the same 'flat' response, but would have lacked a specific "spike" in a particular frequency.
 

ZolaIII

Major Contributor
Joined
Jul 28, 2019
Messages
4,069
Likes
2,409
If driver or combination of them has no large offender's (spikes and deeps in ± dB horizontal line and it's not much out of the horizontal axis) and if it has a low distortion (as low as possible [realistically and better on hedaphone one's]) you will have more DSP headroom and be able to reach higher LPS level keeping distortion low. Including bass level adjustments and other preferences and excitement's (like female overtones for example). So rule is adres only biggest spikes end do as much as possible leveling inverse [self level adjustment for highs and lows if it's overlay bright or boomy and bringing mids up front] and adjust the overlay output minus at the end for the applied biggest gain in parametric EQ filter.
While this is cheapest achievable with IEM's (starting with good one's for as low as 50~70 $/€ to my better knowledge and even TWS minuse BT lacks in highs and if you can hear them still heavy depended on codec and it's implementation for 100~130 $/€) and even on very loud levels still they are not best way on the long run for a bunch of reasons (to long discussions even for a pub). On the platform (Android) this day's for DSP capabilities I would mention Neutron as a cheap but buggy or just to upset your nostalgic about old V4A which whose progressive to a point of option of doing a correction on separate (PC) system and exporting it as convolution kernel to it.

No magic, no turning frog into the royalty, realistic expectations and physics.

As you are to far away and I got upset (about large black, strong finished one's) I will have to go down and pore my self a (couple) apple rakija's and... green.
 
Top Bottom