EQ'ing to Harman curve doesn't give me pleasing results - why not?

[ozzy9832001]

I also would like to point out that you will configure the FR to one thing and it will be completely different once you sit down and have a chair. This is from my left ear.

Blue line represents the FR with no chair or body. As we typically would measure it.
Yellow line represents mic pointed at ear height right in line with my ear and speaker with me physically sitting in my chair (as I would sit)
Purple represents mic point upward 90 degrees me sitting in chair.
Pink line represents me turning my head 45 degrees me sitting in chair.

The best response is me in the chair with ear aligned with speaker, as it should be. My point was to illustrate that the response can and will change even when just moving your head. And it's best to have a mic stand so you can actually measure with you sitting down. The human body is actually a really good absorber and depending on what your chair is made of it will either absorb, reflect or both.

This is where I stopped. It's about as good as I can get with a 2.0 system. After awhile you change one thing and it messes something else up. The only way to get a better response would be even more treatment...and blocking the windows just doesn't seem like a viable alternative.

 

[Thomas_A]

The head and body is always there and should not be corrected for. The direct sound should have a flat frequency response at LP.

 

[dasdoing]

yea, real sound events have no correction for our body,
the seat is problematic though. we should correct for it's influence, but without the body it will meassure differently. unless it's fully absorbive (unlikly this exists) we can't meassure with it in place cause all it will cause is crazy comb filtering

 

[theswede]

I’ve used Dirac and rew for 6 years always changing something and consequently having done many hundred measurements with both Dirac and rew.
My experience is if you use the same measurement points and target curve as Dirac the after measurement results are as close to each other as you can get and the subjective tonal balance is the same. There are 2 notable differences with Dirac imho and in my 7.2 setup. First is the timing or spacial coherence between all the speakers. Second is overall resolution which is higher with Dirac and not always preferred or should I say source dependent.
Not sure if this is due to Dirac measuring the distances between the speakers better or if it’s their acclaimed phase correction.
That said if I measure the speakers near field after Dirac the phase response is very good or almost perfect.
All in all I use both and switch depending on what’s playing.
Example is YouTube and Netflix typically end up with rew filters while UHD Blu-ray hd audio gets the Dirac filters.

 

[tmtomh]

This has been an erudite and lengthy discussion; apology but I have only scanned it.-- my aged vision makes it difficult for me to read lengthy texts, "subjectively" my eyes bleed.

However I especially appreciate @alex-z's providing the Toole/Olive subject preference curves:

My own preference corresponds most closely to the "trained-listeners" preference per the graphs; actually on the high end I might be a little closer to the "all listeners" curve but that would be on account of my ears' decline in HF sensitivity. I've had no "training" other than my own 50+ years of listening.

For EQ I use a crude & easy method, that is, MathAudio's Room EQ plug-in in my Foobar2000 player. I use its "Neutral" setting, slightly modified, rather than its "Bright", (i.e. flat) setting, as measured at my listening position. Works very well.

Interesting. I have to say, I'm not quite sure what to make of this. If I used the Trained Listeners target curve on my speakers in my room - that is, +3-4dB in the lower bass and down what looks to be about 7dB at 20kHz - my speakers would sound bass-shy, treble-shy, and/or midrange-heavy. I don't like smiley-face EQ and I'm not a bass-head, and maybe I'm doing something wrong with my room measurements, but I've used a "Harman" curve that's similar to the All Listeners curve from about 100Hz-2kHz, but then about halfway between All Listeners and Untrained Listeners below 100Hz, and halfway between All Listeners and Trained Listeners above 2kHz, and it doesn't sound bass-heavy or otherwise off to me.

 

[ozzy9832001]

Interesting. I have to say, I'm not quite sure what to make of this. If I used the Trained Listeners target curve on my speakers in my room - that is, +3-4dB in the lower bass and down what looks to be about 7dB at 20kHz - my speakers would sound bass-shy, treble-shy, and/or midrange-heavy. I don't like smiley-face EQ and I'm not a bass-head, and maybe I'm doing something wrong with my room measurements, but I've used a "Harman" curve that's similar to the All Listeners curve from about 100Hz-2kHz, but then about halfway between All Listeners and Untrained Listeners below 100Hz, and halfway between All Listeners and Trained Listeners above 2kHz, and it doesn't sound bass-heavy or otherwise off to me.

I would say the all listeners is about what I EQ too as well. I listen to a lot of metal and I think the bass is crazy good to not be heard. Maybe a combo. I don't mind a bit more energy between the 1k-3.5k ish mark. Makes the guitar solos pop a bit more.

 

[neRok]

If one speaker is appearing louder or the phantom image has shifted then you probably have to lower or raise the gain on one or the other. That has to do with the way the reflections are playing off the two sidewalls. Probably best to lower gain.

I don't think it's a phantom center problem, because vocals sound central. It might be a hearing problem because my left ear canal feels a bit "itchy" this morning (allergies), and the music sounds a bit more to the right, lol (it also sounds a bit unrefined without much EQ atm). But I will have a go at tweaking the relative levels and see what happens.

 

[DualTriode]

Ever try the "microphone in the box" method of measuring/estimating the speaker bass response? None of the in the room issues come into play.

Thanks DT

 

[Rednaxela]

Cool. Never heard of it.

Measuring Loudspeaker Low-Frequency Response

With the availability of low-cost, PC-based acoustic data acquisition systems, experienced hobbyists can make highly accurate loudspeaker frequency-response measurements without an anechoic chamber. The power of these systems comes at a price. These systems window the time-domain measurement to...

audioxpress.com

 

[Gorgonzola]

Interesting. I have to say, I'm not quite sure what to make of this. If I used the Trained Listeners target curve on my speakers in my room - that is, +3-4dB in the lower bass and down what looks to be about 7dB at 20kHz - my speakers would sound bass-shy, treble-shy, and/or midrange-heavy. I don't like smiley-face EQ and I'm not a bass-head, and maybe I'm doing something wrong with my room measurements, but I've used a "Harman" curve that's similar to the All Listeners curve from about 100Hz-2kHz, but then about halfway between All Listeners and Untrained Listeners below 100Hz, and halfway between All Listeners and Trained Listeners above 2kHz, and it doesn't sound bass-heavy or otherwise off to me.

That's the thing about subject preferences: we're all a bit different. So if you're measured response is different from, say, 'Trained Listeners' or whatever, but that's what you like, who's to say your wrong?

It's not necessarily your room, speakers, or measurements if you like a little bass boost, but it might be the variety of music you tend to listen to.

 

[sigbergaudio]

seeing the graph on post 58 for the perhaps 10th time I now realize that this target is actually the

View attachment 295632

This is a pretty important point. You're not supposed to EQ your system to look like this, it's what you would expect to see in a normal room with decent speakers.

Below Schroeder (2-400hz or so), you can dampen peaks and adjust the bass to your preference, leave the rest alone.

 

[neRok]

I didn't get around to doing any proper testing today, but I did quickly whip up an EQ a while ago. I took yesterdays MMM in REW EQ and loaded the estimated-in-room-response in as the target. Then I fiddled around with its EQ methods, kind of doing it in chunks (eg I did one 600 to 1200Hz so it would make a single PEQ to pull down the 850Hz peak). I then stitched them together, but then thought it looked like everything below 500Hz could do with shelving down instead, so I reduced the -dB PEQ's a bit and did the shelf. I didn't do it super scientifically, just winged it (as a trial run / proof of concept).

Then I exported that estimate, did a MMM with that EQ applied, and the good news is that the EQ translated very well to the MMM. I think this is a step in the right direction, as it sounds nice enough and doesn't have the previous "void" when I toggle the EQ on and off.

Looking at the resulting measurement, I'm going to investigate if anything can be done phase wise at 250Hz. 700Hz needs a tweak. The bottom end rolloff needs sorting (haven't touched it beyond the 48Hz PEQ and shelf). But it's showing promise. I wonder if it was good how I EQ'd and shelved it so that the peaks below 300Hz are on one slope, and the valleys are on another slope? I'm having a listen now and will again tomorrow, and will report back then.

Regarding target curves - I'm not even at the point of judging "bass boost" and "high end slopes". My original intention by using a target was to find where in all the ups and downs the speakers "baseline" was. Actually using the EIRR seems much better for that purpose, so that was a good prompt from that poster. Someone also commented they EQ flat/smooth and then apply their shelving preference on top, and that seems a decent idea, and is kind of what I've tried to do with the trial EQ in this post (I've tried to tame the problem areas as much as reasonably possible).

Anyway, time will tell, because it wouldn't be the first time I've clicked on an EQ and thought "yer that's alright", but then a few hours later thought "yer nah, somethings not right". I am noticing a bit more punch in the bottom end which I have been missing since switching to these speakers though, and I think it might be thanks to that 90Hz phase tweak.

 

[Rednaxela]

Then I exported that estimate, did a MMM with that EQ applied, and the good news is that the EQ translated very well to the MMM.

Good call. It is also my experience that a post-EQ MMM tracks REW’s prediction rather closely. It’s great to have this confirmation and it is one of the things like about MMM. This is quite a bit harder to achieve with a multi-point measurement.

 

[ozzy9832001]

Your frequency response looks pretty good in the last measurement; I'd be happy with that. If the bass is lacking then you can EQ to taste or use the speakers built in tone controls. You may want to start looking into decay times and things of that nature.

If you wouldn't mind, can you post the entire .mdat file in a zip so we can take a look at the other aspects?

 

[dasdoing]

You're not supposed to EQ your system to look like this, it's what you would expect to see in a normal room with decent speakers.

it works very well as a starting point for me in my reflection treated room. the normal Harman on the other hand sounds "broken" to me ears and in my room

 

[Deleted member 48726]

I think you make very good points. As I read the Dirac manual, it provides three options: "very focused" with 9 measurements, "focused" with 13 measurements, and "wide" (or something like that) with 17 measurements. I chose the "focused"/13 measurement option simply because it was the middle option and I'd never used Dirac before. I've never actually checked into whether I could do even fewer than 9 measurements and then just skip to the "create filter" step. That might be interesting.

I also was under the impression that Dirac requires, or at least defaults to, a minimum of 9 measurements because it uses differences among the various measurements to create a quasi-anechoic model of the speaker's response vs the room effects. But of course I could be wrong about that.

As for what sounds better with EQ, I've always found it pretty obvious - typically tighter bass with fewer obvious peaks/resonances, and smoother mid/treble response with no objectionable peaking and hopefully no harshness.

However, I have managed to get generally pleasant, linear-sounding results with fairly crude EQ methods, for example a 10-band graphic EQ adjusted based on single-point readings at each octave with an SPL meter.

All that said, while I was waiting to receive my MiniDSP SHD unit after I got my latest pair of speakers, I did some quick and dirty EQ using the graphic EQ in my music playback software. The EQ immediately produced a better-sounding (to me, at least) result than no EQ. But after installing the SHD and creating some Dirac profiles, I disabled Dirac and re-enabled that graphic EQ setting in my music player app. And the sound was noticeably duller and less balanced than with the more sophisticated EQ in Dirac.

So based on what I've heard over the years, I lean towards EQ, especially high-quality digital EQ like with the SHD and Dirac, since the THD and noise penalty of such processing is negligible.

Just a personal side note of anecdotal (questionable) quality; I've come to realise that a NEUTRAL response often sounds "DULL" compared to, lets say, a peaky one. An eye-opener in this regard, for me was when I started out with REW and a UMIK-1 and did a ****-ton of measurements in my learning process. One can easily get used to a certain frequency response, whether that be neutral or +12 dB bass boost etc. So I figured that I may as well get used to as neutral sound as possible to, you know, get as close to reproduction 1:1 as possible..

 

[dasdoing]

One can easily get used to a certain frequency response

totally....but I think that once used to something near neutral it is hard to get back. you kind of realise it sounds natural, while when you were used to "boomboxes" it "just sounded good because it has nice bass" for example.
it's like with todays pictures with those stupid filters. people who grew up with it don't see nothing wrong, but if you grew up with more realistic pictures it just seams off

 

[Deleted member 48726]

totally....but I think that once used to something near neutral it is hard to get back. you kind of realise it sounds natural, while when you were used to "boomboxes" it "just sounded good because it has nice bass" for example.
it's like with todays pictures with those stupid filters. people who grew up with it don't see nothing wrong, but if you grew up with more realistic pictures it just seams off

Yeah. It's like sound calibration for the brain..

 

[ozzy9832001]

Just a personal side note of anecdotal (questionable) quality; I've come to realise that a NEUTRAL response often sounds "DULL" compared to, lets say, a peaky one. An eye-opener in this regard, for me was when I started out with REW and a UMIK-1 and did a ****-ton of measurements in my learning process. One can easily get used to a certain frequency response, whether that be neutral or +12 dB bass boost etc. So I figured that I may as well get used to as neutral sound as possible to, you know, get as close to reproduction 1:1 as possible..

I always think a neutral or slight increase from like 120hz and below always sounds best and works with a wide variety of music. Mastering quality, artistic vision, etc all really push for a more neutral response.

 

[tmtomh]

Apologies if I'm being dense, but isn't the point of all the so-called "Harman" and/or "Toole-inspired" target curves that an anechoically flat speaker tends to produce non-flat response in a typical room, and that most listeners prefer that non-flat in-room response, even if trained and untrained listeners prefer different degrees or variations on that response?

If so, then it seems to me we can't just say there are "good" and "bad" speakers and therefore you should't EQ above the Shroeder frequency of your room, because if you need to do that then what you really should do is get different speakers.

Of course there are bad speakers out there and it's best to get rid of them rather than trying to use room EQ to "fix" them. But we're all familiar with "good" passive speakers with decent directivity that nevertheless can benefit from some EQ because the treble is a little too hot, or there's an area of some attenuation around a crossover point, or there's a cabinet or port resonance that mars otherwise good performance, and so on - yes?

I suppose one could use room EQ below the Schroeder frequency and then manually apply PEQ to address any issues above that frequency - but it seems to me that's not inherently a more sound approach than just letting the room EQ address those issues.

Thoughts?