[Agavehound]

Hmph, after reading my post over again I see I didn't give near enough info. The sound level was changing from side to side over very small increments of frequency shift. Almost as if it was playing in stereo...that was the clue. 
 
After thinking about it for a bit I decided it had to be a system issue so I started to change things. It seems that having the Creative Audio Console in 'Audio Creation' mode was causing the shift from side to side. I swapped to entertainment and it tracks perfectly even from both sides. I really don't care for Creative software but...it's what I have for now.
 
Thanks for the reply's gang!
 
Regards,
Chas

 

[Argyris]

Glad you got that sussed out. That's really weird, though, about Creative's software doing that. I've never had it do anything like that, and I run it in Audio Creation mode all the time. It sounds like some strange processing effect was flipped on. Creative's software gimmicks get on my nerves sometimes, though my X-Fi card has been trucking for years without a single problem. Creative's products get bashed around here a lot, and having not had a "good" DAC I can't say for sure, but it doesn't sound bad to me at all. Which reminds me. I saw a screenshot of a parametric EQ in the settings somewhere. I'm too lazy to go check now, but if I can find that I might try to transfer my Beyers' curve over so I don't have to keep loading it in my DAW every time I want to check a mix or just goof around in general. This is of course provided it applies only to playback and not input, which would be pretty boneheaded and would render it useless to me.

 

[Omega17TheTrue]

I have a few questions :

Can the VST plugin works everywhere and not just on special audio player ? I thinks the answer is no but there might be a thing.

Is it possible to change FR of left and right channels independently ? If not its seriously sux to only have a mono equalization and nothing to fix drivers/ears L&R differences.

the thread should be moved to "Computer Audio".

 

[Joe Bloggs]

@Omega: you can get Virtual Audio Cable to make a virtual sound output to route all system audio to
http://software.muzychenko.net/eng/vac.htm
 
then get a VST host... like e.g. VSTHost
http://www.hermannseib.com/english/vsthost.htm
 
to pass the audio through VST plugins before sending them to the actual audio output.  The configuration is a bit hit and miss though.  It totally didn't work on my netbook, kinda sorta works on my desktop.
 
In VSTHost you can choose to chain only one channel to a plugin in the "chain after" dialog for the plugin.  So you should be able to equalize left and right ears independently.  Don't think there's anything like that in foobar2000 though but I could be wrong.

 

[Omega17TheTrue]

Thanks for your link but I tried Electri Q and its total garbage, full of bugs and error, a nightmare, gonna try something better...

 

[AstralStorm]

Indeed, ElectriQ (at least "freeter" edition) has its share of bugs, including crashes. Mostly annoying display bugs and the fact that it can and will clip internally if you use it for any kind of boosts.
 
However, it serves well if you can live with the annoyances. It's not like that eq is meant to be tweaked all the time after all.

 

[Joe Bloggs]

Quote:

Also, as I mentioned before, human hearing is not flat over the whole frequency spectrum, which leads to inaccuracy in balancing the frequencies by ear.  For those of you who got VAC and VSTHost working so that you are checking your EQ on the fly directly from SineGen, I created a replica of the 60dB equal loudness curve:
http://www.mediafire.com/?7b9olaa40qk4lfh

 
What I'm doing now is running two copies of Electri-Q between the VAC input and the speaker output in VSTHost.  The first copy is loaded with the equal loudness curve preset.  With the first EQ alone, a pair of headphones with response at the ear equivalent to a loudspeaker with perfect response will theoretically sound as loud at all frequencies if you play tones at 60dB (loudness of a normal conversation)--if your perception of loudness is the same as the average ISO 226:2003 test subject.
 
With the second copy of Electri-Q, I will tweak my custom EQ preset until all frequencies do actually sound as loud to my ear with SineGen playing through the two EQs.  The EQ should then be the perfect when you remove the equal loudness curve EQ from the chain.  I will try this myself and report back on my findings.
 
If you want to try this yourself, verify that the equal loudness curve preset is loaded correctly (I have to load it twice every time for the curve to be correct;  you would have to set M->Skin->Gain Range to 30dB just to be able to see the whole curve...) and make sure you don't clip the bass frequencies which are equalized above the zero mark on my preset.  That would mean outputting at -30dB or lower on SineGen for going all the way down to 20Hz with a flat second EQ.  If sounds don't reach 60dB at your ears with that output level even with system volume at max then you may have to give up EQing down to 20Hz.

I didn't do the two-pass EQ thing, but I did re-evaluate the EQ of my Philips SHE3580 with the equal loudness contours in mind and here's what I came up with:

 
The dip at 3000Hz was lessened (though I found that undipping it by as much as the contours suggested made the sound unnatural to my ears) and the dips at 5300Hz and 8150Hz were taken further down.  The sound gets ever more unbelievably lush and 3D, and sibilance should be less of a problem than before (this just confirmed by listening to some pop).  You just can't believe this sound is coming from a $10 pair of IEMs.  I tried plugging these like the $7.11 monoprice, but the transformation these undertake from the EQ is so complete that I'm afraid anyone who buys these on my recommendation and just tries them on vanilla would find them to be utter trash.  Or have my ears just been spoiled by the superlative sound I'm getting with the EQ?
 
Anyone wants to try these on my recommendation, I would send the EQ curves to you to try out.

 

[applepancakes]

(http://www.jstage.jst.go.jp/article/ast/24/5/311/_pdf) could also eqing to the 'with pinna' curve. or subtracting without pinna

 

[AstralStorm]

Bah, I found equalizing using curves to be far less accurate than by ear. Even when using good measurements and my own subtracted head transfer function.
 
Best guide is: use 500 Hz tone as reference, set volume so the tone is slightly annoying. Play tones at either graphic eq band frequencies or with bands like: 20-120 Hz, 6 tones; 120-500 Hz - 6 tones; 500 Hz and beyond - tone every 500 Hz.  Equalize until they sound equally loud and are as centered as possible. Also crosscheck octaves like 125 Hz; 250Hz; 500Hz; 1000Hz; 2000 Hz; 4000Hz; 8000Hz.
 
Unfortunately, I can't attach here, so here goes the xgeq file for Foobar2000 Graphic Equalizer plugin, DT1350 and FiiO E7 on loudness 28 and 3-dot bass boost (4.5 dB below 1000 Hz).
Reference 0 dB at 500 Hz, precision 0.5 dB. Total cut -15 dB. (additional -6 dB headroom)
Screenshot:

http://ompldr.org/vYnJ0Zg/DT1350-vol28.zip
 
Default Foobar's equalizer is useless due to lack of interpolation between bands. This deficiency causes nasty audible artifacts.

 

[RPGWiZaRD]

Quote:

Bah, I found equalizing using curves to be far less accurate than by ear. Even when using good measurements and my own subtracted head transfer function.

This, often I've seen quite nice correlations to how I'd interpret it with my ears and a graph but it's not always perfectly the case and well I'm just of the opinion that you should EQ by ear because then you can adjust how YOU think it sounds right, that's what really matters right? You get most satisfied with the result when you just listen to your own ears instead of objectively trying to adjust it in a certain way using a graph as reference because when doing so it might not be what your ears or tastes wants to hear. If tuning for studio use, ofc it's great to have objective reference measurements and guides etc but when EQing for your own good, for music listening pleasure, just trust your ears.
 

 

[AstralStorm]

Nah, especially for studio use equalizing by ear with pure tones works very, very well. Perhaps another kind of equalization would use "pieces" of pink noise at the right central frequency, could be even more accurate perhaps, but the samples for equalization will have to be generated. I should experiment with that some time soon.
 
Minor hotfix to the above equalization:

http://ompldr.org/vYnMwZA/DT1350-vol28.zip

 

[Joe Bloggs]

@AstralStorm: How can plain EQing by ear give correct results when we perceive e.g. 30Hz to be 40dB quieter than 1kHz when they are in fact the same volume?  The result will be an EQ setting that way overemphasizes the bass and treble.  I bet you'd wanted to push those 20-30Hz sliders up past the ceiling but then decided it didn't sound good with actual music so dialed them back after trying to convince your brain that those levels are level with the 500Hz tone when they clearly weren't

 
I equalize by ear too.  I do that AFTER accounting for the equal loudness curve.  If I didn't I would just leave all the bass frequencies flat with 1kHz on my Philips SHE3580 EQ.  And hurt my brain with skull-crushing bass every time I play actual music

 
Also after I went to continuous tone sweeps with Sinegen and continuous FR adjustments with parametric EQs like Electri-Q I'm never going back to discrete tone measurements and graphic equalizers

 

[SanjiWatsuki]

I thought I'd just plop down my style of equalization. Let me preface it by saying this is not for studio use at all -- this is purely for music enjoyment. I really don't think that too many people prefer a perfectly neutral sound, in reality. Also, what's to say that there haven't already been compensations, conscious or otherwise, for the equal loudness curve done when the track was being put together? For just general music listening, "correct" is just what sounds best.
 
My personal style of equalization basically involves taking the qualities of the headphone I don't like and trying to figure out which frequency bands are known for causing it. For example, a boomy bass near 100hz or knowing that the bass drum gains a lot of its body from around 70hz. There are equalization cheat sheets which talk about which frequency bands do what to which instruments all of that sort of fun, like this one.
 
If I have measurements in front of me, first I try to level match to an existing high end headphone which has a similar sound signature to what I want to hear. For example, with my KRK KNS 6400 I level matched against the Sennheiser HD800. This serves as a starting point. After that I just run it through a gambit of music and try to recognize qualities I wish the headphone had, like better percussive attack or a more full bass response. I save a variety of different equalizations and then A/B test against them to try and find which one is best, or what qualities I liked best out of each EQ. It also really helps you get a handle on what type of sound signatures you can expect from a headphone based off of its FR (not to say that this gives a complete picture of what it sounds like, this is just simply an outline). 

 

[Argyris]

How can plain EQing by ear give correct results when we perceive e.g. 30Hz to be 40dB quieter than 1kHz when they are in fact the same volume?  The result will be an EQ setting that way overemphasizes the bass and treble.

He's right. We do not hear high and low frequency bands at anything approaching the same apparent amplitude as those in the middle; therefore, matching amplitude to a reference tone will produce a grossly inaccurate curve. I know all too well about the first part, having messed up my hearing for a few days this week by turning up the volume trying to hear the stuff between 10 - 20kHz "better." (Just because we perceive a range of sound as quieter doesn't mean it's not delivering the same amount of energy to our inner ear--remember that, kids).

The "proper" EQ method (or, perhaps, a proper EQ method) is the one outlined in the OP's first post, which is to bring down peaks to the relative baseline level (using pink noise to reveal that baseline). The baseline will shift downward the higher you go when you're working with the treble (which is where this method is the most accurate), but as long as the notch basically accords with the bands immediately adjacent you'll likely "capture the trend" or else be close enough that it won't matter.

This all isn't perfect because nothing really is. At some point it's always going to be subjective. But it's important to understand how people hear (and what a "flat" hearing curve roughly looks like) in order to get the most out of this method.

And, of course, once you've ironed out the response you can sweeten the sound to your heart's content (add extra bass, liven up the treble, etc.) because you have a clean slate to start from.

 

[Argyris]

In fact, I think it will be instructive to post an image showing people what the human hearing curve approximately looks like. It's not "flat" by any means:



You can probably see what I did--I just took an image of the equal loudness contour (at just one amplitude versus the multiple amplitude chart you usually see) and inverted it. As the constant amplitude gets louder, the chart "flattens out" somewhat (e.g. the rolloff in both directions won't be as steep) and that 3.7kHz bump gets higher. This is at the lower end of the amplitude spectrum, incidentally.

Keep in mind that this is not to be used as an unimpeachable reference--it's just a chart I found after a quick Google search. I got the original image here:

http://upload.wikimedia.org/wikipedia/commons/thumb/d/d5/Perceived_Human_Hearing.png/800px-Perceived_Human_Hearing.png

All I'm trying to do is show, roughly, what sort of curve you should be hearing. Likely you won't hear exactly like this (maybe you won't perceive as much rolloff below 500Hz, especially if your headphones have some extra "junk" down there). Nevertheless, now you should understand why making every frequency sound "flat" is a prelude to disaster. Your focus should be on the stuff happening after that big bump at 3.7kHz (this bump should NOT be EQ'ed out unless it's excessive), and the goal is to hug that slope. You will almost certainly find a feature around 7.5kHz, as this seems to be almost universal (and suspiciously, it's around double the 3.7kHz bump's frequency, suggesting some sort of relationship). Also, you need to do all this at the highest comfortable volume that you will normally listen to your music.