[donunus]

It is because of the fact that there is no real universal standard for the sound (frequency response) for the different types of headphones that they are more fun and more annoying sometimes when you think in terms of achieving the holy grail with them. More fun cause different goals of different manufacturers produce different sound signatures that are all enjoyable and valid. More annoying if you think in terms of searching for the perfect sound.

With speakers, all high end companies tend to aim for flat response while creating their own signature sounds with their approach...eg. horns, electrostatics, planars, different box designs for dynamic speakers, etc...

 

[JaZZ]

Quote:

Originally Posted by bigshot I had a unique opportunity yesterday. A friend of mine who designs speaker systems invited me over to spend the day listening to his rig. His theory is that if you achieve a perfectly flat response, just about any kind of music you throw at a system will sound good. ... He's spent the last several months perfecting the response of this particular rig, which includes horn loaded bass speakers capable of reproducing cleanly all the way down to 20 hz. He's carefully matched components on his speakers to make sure they cover the whole spectrum of sound evenly. He has an octave equalizer that he uses to fine tune frequencies one by one as he sweeps through them using a tone generator. This ensures that what you hear out of the speakers is *exactly* what is burned into the CD.

Well, the last sentence is a bit of wishful thinking. But let me elaborate why.

Indeed a flat frequency response is the most important precondition for a neutral music reproduction. Now what your friend has measured is a relatively even frequency response at one point of the room and with quite course measuring resolution. A graphic equalizer doesn't allow precise equalizing of resonances, the less so such with small bandwidth. But most likely the course measuring resolution doesn't even make small-bandwidth resonances visible. And it has to be said that those are less obvious when it comes to perceived neutrality, they rather make for hardness or harshness of the sound. Back to measuring resolution: In unsmoothed mode, you would have a hard time judging let alone equalizing the frequency-response curve at all -- considering all the jags. Also a microphone displacement by only millimeters will dramatically alter the curve.

Even if we would ignore the real-world frequency response, it doesn't say everything about the sound. There's harmonic distortion, and there's transient behavior. In principle, an absolutely flat frequency response (with no bandwidth restriction) would guarantee a perfect transient response. But that's just theory -- and only applies to point sources in free-field conditions. A horn system by nature can't be seen as point source, the less so a multiway system. Horns act as reflectors for the original sound waves and therefore -- measurably -- corrupt the transient response. So in contrast to what your ears tell you, horn drivers are in fact slower than classic dynamic drivers.

Moreover: what the microphone measures in your friend's listening room is a wild mixture of direct and reflected sound -- depending on the polar pattern of the speaker. You must know that there's no speaker with just decently uniform sound dispersion, and the radiation characteristic has important impact on the perceived sound. The microphone cant differentiate between direct and reflected sound, but the human ears can. That's why it's possible that two speaker systems with seemingly flat frequency-response curves (at one point of the room) can sound totally different.

With headphones -- believe it or not -- it's even more complicated, because considering individual anatomics you have to take the whole head/ear/headphone system into account, not just the drivers. The headphone developers' attempt to create some sort of a universally valid ear curve is not more than that -- that's why perception and rating of different headphones by different people are so diversified.

 

[Philco]

Quote:

Originally Posted by JaZZ Indeed a flat frequency response is the most important precondition for a neutral music reproduction. Now what your friend has measured is a relatively even frequency response at one point of the room and with quite course measuring resolution. A graphic equalizer doesn't allow precise equalizing of resonances, the less so such with small bandwidth. But most likely the course measuring resolution doesn't even make small-bandwidth resonances visible. And it has to be said that those are less obvious when it comes to perceived neutrality, they rather make for hardness or harshness of the sound. Back to measuring resolution: In unsmoothed mode, you would have a hard time judging let alone equalizing the frequency-response curve at all -- considering all the jags. Also a microphone displacement by only millimeters will dramatically alter the curve. Even if we would ignore the real-world frequency response, it doesn't say everything about the sound. There's harmonic distortion, and there's transient behavior. In principle, an absolutely flat frequency response (with no bandwidth restriction) would guarantee a perfect transient response. But that's just theory -- and only applies to point sources in free-field conditions. A horn system by nature can't be seen as point source, the less so a multiway system. Horns act as reflectors for the original sound waves and therefore -- measurably -- corrupt the transient response. So in contrast to what your ears tell you, horn drivers are in fact slower than classic dynamic drivers. Moreover: what the microphone measures in your friend's listening room is a wild mixture of direct and reflected sound -- depending on the polar pattern of the speaker. You must know that there's no speaker with just decently uniform sound dispersion, and the radiation characteristic has important impact on the perceived sound. The microphone cant differentiate between direct and reflected sound, but the human ears can. That's why it's possible that two speaker systems with seemingly flat frequency-response curves (at one point of the room) can sound totally different. With headphones -- believe it or not -- it's even more complicated, because considering individual anatomics you have to take the whole head/ear/headphone system into account, not just the drivers. The headphone developers' attempt to create some sort of a universally valid ear curve is not more than that -- that's why perception and rating of different headphones by different people are so diversified.

What he said is dead on! Great Post JaZZ !

For the tests you and your friend set up to be really conclusive, you would have to be in a perfect acoustic environment with no reflections or frequency traps, with a perfectly flat mic and perfectly flat processing. Then you would have to get the most perfect EQ on earth to get rid of all the peaks you would find. This is a waste of time in the long term.

The best way to judge for flatness is to listen if instruments you are familiar with sound natural. In this regard, I think you'd be hard pressed to find more accurate phones than the HD600 and HD650. Sennheiser knew exactly what they did when they designed those, they applied the Diffuse Field curves then measured then tweaked then measured then tweaked.....till they got the right response. Other (dynamic) phones that could be as accurate might be the AKG K240DF which is also diffuse-field equalized.

 

[Philco]

Quote:

Originally Posted by donunus Without hearing the headphone, its hard to guess how it will sound by looking at the graphs with the big dips and peaks to simulate accuracy. I'm sure there are ideal Frequency responses for all headphone designs but where are they posted? Who is the authority of these curves? I would like to read more about this myself.

Diffuse Field equalization is your answer.
Moreover, I don't agree that one can't judge how it will sound by looking at graphs. Graphs tell everything, much better than ears can. Remember, everyone's hearing is different, but graphs (if well obtained) are universal.

 

[coffeeaddict]

Durrr... me like music.

 

[12thgear]

Quote:

Originally Posted by donunus EXACTLY! Thats one thing i forgot to mention in my long replies. The sound will supposedly be what the producer intended if you have a flat response but they mix for the masses sometimes and some albums just don't sound good on a neutral system. Besides, we are not even sure if their monitor speakers for mixing and mastering have flat responses

Yamaha NS10 monitors are (in)famous for their hardly neutral sound. They weren't flat, but it seemed that many sound engineers loved them because supposedly, if your mix sounded good on a pair of NS10s, it would sound good on nearly anything.

My buddy has a small recording studio and has some pretty nice near field monitors for the price. They're measured to be extremely neutral. So that's his baseline, but he'll also burn a disc or dump the WAVs on an iPod and give the mix a listen on a boombox and some car stereos.

He often laments that it can be a little frustrating since both he and I spent four figures on decent 5.1 home theater speakers that also perform well in stereo for music. However, they're still sub/sat systems and far from ideal, and far from high-end. We're still talking mid-fi at best. But our systems still blow away what the vast majority of people use when listening to music.

I'm in the same boat as a graphic designer when it comes to web design. Sure, I have a nice, calibrated monitor, and things look true and natural on it. But then I have to take into account nearly nobody caibrates their displays. Most users skimp on monitor quality. Mac native gamma is 1.8 and for Windows it's 2.2. So I get to spend my time comparing on both my Mac and PC while still having to accept that results will vary from that. And let's not even get into older browsers, operating systems, etc. You do your best to account for all reasonable variables, and hope the end user doesn't blame you if things don't work right.

 

[allenf]

Quote:

Originally Posted by JaZZ <Good stuff snipped> With headphones -- believe it or not -- it's even more complicated, because considering individual anatomics you have to take the whole head/ear/headphone system into account, not just the drivers. The headphone developers' attempt to create some sort of a universally valid ear curve is not more than that -- that's why perception and rating of different headphones by different people are so diversified.

I am prepared to take your word on this but...I do find it hard to believe that the head/ear/headphone interface has more variables than the loudspeaker/room interface - or have I misread your post?

 

[dwc]

Late to this thread...

Check out Gerg's thread on the digital eq. On the page of the post linked below, he uses a mic to flatten the response of hd-600's. You can see the before/after sweeps, as well as the shape of the eq he applied:

Gerg's thread

 

[Jahn]

Quote:

Originally Posted by coffeeaddict Durrr... me like music.

You and me both man! This is why I stay out of the DIY forums - my brain starts to spin and me get dizzy!

 

[bigshot]

Quote:

Originally Posted by donunus You said your friends system is setup to be flat and you are trying to mimic the sound of his setup on your system using an eq... You might get good results but it is not likely to sound the same due to the different sound delay characteristics of your room and your speakers design if different from his speakers.

I know that... But I've done some fiddling with the settings again this morning, and it's an improvement at least.

Quote:

Originally Posted by donunus Another thing is... you said he is using an octave eq. Thats around 10 to 12 band i think.

It's a board the size of a small coffee table. I'm not sure how many bands... 48? It's a guess.

Quote:

Originally Posted by donunus The disadvantages in using eq outweigh the advantages unless the headphones you have don't satisfy you at all.

My theory is that EQ is most useful for fine tuning a rig that already sounds good.

See ya
Steve

 

[wualta]

Quote:

Originally Posted by bigshot While we were listening, my friend said, "Let me show you something..." He got up and turned up the volume to a VERY high volume. I could feel the bass hitting the floorboards of the house as hard as someone hitting them with a baseball bat. .. He left it at that volume for a couple of minutes and then he ramped it back all the way to zero. He smiled at me and spoke very quietly, "Your ears don't hurt." ... I had just experienced a sound bigger and louder than anything else I'd ever heard, but there was no ringing or pain. While the sound was blasting me, I didn't even flinch.

You performed a simple crescendo test. And you found out that despite your friend's use of a calibration system that wasn't state of the art even 35 years ago, he's managed to come closer than most home systems that you're familiar with to a flat frequency response, despite the fact that the speakers, microphone, room and equalizer are all very very imperfect. The proof that he did is your reaction to this test, and the other listening tests.

All things are relative, of course. With room treatments (and I mean real room treatments, involving partial demolition, concrete slabs, floating subfloors and lead sheeting), gated impulse-response measurements and further modern bla bla, you could be hearing reproduction that's as much improved over your friend's system as his system is over yours. And as has been pointed out, some CDs would still sound crappy. Crappier, even. You'd have to start making your own recordings... and then you'd be up against imperfections in microphones, preamps... it never ends.

Remember the bumblebee. It shouldn't work, it can't possibly work. Flat response is impossible, an engineering wet dream, a mirage, give up now. Yet improvement happens. It's even measurable. Go figure.


Quote:

Well, I got home and immediately started twiddling with the EQ on my system to get it closer to that sound. I'm going back for the 4th to do some more listening (and hot dogs too!)

Have fun, and eat only balanced hot dogs. Who knows, this may be the start of a whole new nonhotdog, nonpatriotic lifestyle-- it certainly won't be mainstream American. Are you sure you want to go down this road?

 

[bigshot]

Quote:

Originally Posted by Xanadu777 Something to keep in mind too, master recordings in popular music are mixed to sell CD's. The average person, unlike us or other audiophiles in general, are listening to music on car radios, boom boxes or other more inferior gear (Bose). Mixes are made to sound "good" on these mediums, so CD's sell.

The interesting thing about this flat system was that in some respects, just about all of the recordings we were listening to sounded good. My friend asked me to bring along a stack of the best and worst sounding CDs in my collection. The worst by far was Rolling Stones Hot Rocks standard CD. The system showed how the midrange had become all congested, and how the high end had been shaved off, but amazingly the bass was still tight and focused.

We listened to bits of Sgt Pepper and Magical Mystery tour, and I was amazed that I could instantly hear which elements were mixdowns and which were first generation. The mixdowns sounded like they had been run through a bandpass filter, chopping off the top and bottom. Then we listened to a couple of cuts off of Rubber Soul, and it was totally different. There didn't appear to be any mixdowns at all. Norwegian Wood sounded absolutely amazing.

See ya
Steve

 

[bigshot]

Quote:

Originally Posted by donunus Without hearing the headphone, its hard to guess how it will sound by looking at the graphs with the big dips and peaks to simulate accuracy. I'm sure there are ideal Frequency responses for all headphone designs but where are they posted? Who is the authority of these curves? I would like to read more about this myself.

This falls under the heading of psycho-acoustics. There are a bunch of principles involved... but masking is going to be the one that would apply to headphones the most. I'm sure a google search on psycho acoustics masking will bring up some interesting stuff.

See ya
Steve

 

[allenf]

Is this thread really about "flat frequency response" or is it about powerful amps driving sensitive speakers with a great speaker/room interface?
Interesting thread nontheless.

 

[wualta]

Quote:

Originally Posted by dwc Check out Gerg's thread on the digital eq. On the page of the post linked below, he uses a mic to flatten the response of hd-600's.

gerG is da Man. He no fear spooky EQ. Him fearless. Him even try EQ Fostex T50RP. That take guts.

I should explain that Back In The Day, home equalization was another way to say "I'm gonna mess my system up so bad it's gonna make me sit down and bawl like a li'l girl!" Equalizers added noise, distortion and curves that were anything but precise. It was a crapshoot and the odds were heavily against you. Today things are a wee bit better. Which is to say, a LOT better. There's no need to fear EQ any more.

Besides, EQ is a good test of headphone quality. If a headphone takes EQ gracefully, it's a well-designed headphone, all other things being equal. If you boost the bass and nothing pleasant happens, or if you lift the treble and it seems that someone's driving a rusty spike into your ear, something isn't right, and chances are it's not the fault of EQ.