[Glmoneydawg]

hmm, the Harman Curve for headphones:

these guys spent a lot of money testing hundreds of headphones on hundreds of people to find what was the most pleasing sound combination across as many people as possible, then they came up with the curve that represents that sound signature.

any headphone coming closest to it, should in theory, produce the most pleasing sound, be it cheap or expensive. i actually think it's true, i mostly listen on n5005 and i love the sound signature -- maybe there is a tiny bit of of too much bass for my taste, i suppose i can swap out the included filters -- but otherwise i'm a big fan of the sound across many genres of music.

interestingly enough, for truly portable situations i use my beats wireless -- and i'm saying it only here because it's sound science and i feel safe saying it -- i think they're extremely underrated headphones and sound amazing. i looked up some of their sound profile and when compared to the Harman curve they were very close. there are internet posts that you can look up if you search for "headphones closest to harman curve".

This is basically +/-6db...outrageous specs for any audio component...even a room lol....a 12db span :0

 

[bigshot]

The Harman Curve is actually pretty close to my ideal too.

 

[manueljenkin]

Assuming your transducers are capable of producing detail, then achieving it is primarily related to frequency response balance. All recordings sound more detailed when there are no spikes or dips, and there's no auditory masking. Timing errors aren't generally a problem with digital audio, neither is distortion. Good quality transducers should be fine for both timing and distortion as well. The wild card is the response. EQ can work wonders.

More detail might reveal more noise in the recording, and that might not be desirable. Introducing a bed of high frequency hiss might help recordings without upper frequencies, like pre-hifi recordings. But those are different subjects.

While i do agree that fr is one important aspect of detail, i wouldn't call it the only attribute. Fr is calculated by continuous sine sweeps. There is no reference to how it behaves from an inertial reference when say hitting a cymbal crash. CSD, needle plots etc give a much better idea of how a system residual noise is. Residual noise can completely overshadow low level detail and lack of transient attack can affect front back localization.

Also i don't agree with the Harman curve. I don't think there's an ideal curve since we are looking at a system that doesn't have crossfeed which we experience in real life. Depending on whether recording is center channeled or side panned, i swap between the tone of srh940 and srh1540 for my preferred tone. When i want to analyse something from a "I'm at the position of mic" point of view, i resort to the tone of srh940. Binaural is so damn realistic on that.

 

[bigshot]

If a system can reproduce a 20kHz tone, why couldn't it reproduce a cymbal crash?

The Harman Curve doesn't include crossfeed. It also doesn't include time delays from room reflection, noggin metrics or any other distance or directionality information. It's just frequency response. The biggest problem with headphones is frequency response. They do distortion and timing error better than speakers do. When calibrated, they also do frequency response more precisely because of the lack of differences caused by the room. If all that is the case, why to speakers sound so much better than headphones? Because the room is the part of the sound that sets speakers apart... not the specs. The room is just as important in a speaker system as the speakers are.

 

[SP Wild]

I think great details can be broken down into two elements in a headphone.

1. Precise pressure control

2. Sheer driver speed

Precise pressure control requires a total air control system with rigid and 'torquey' motors that can push the mids and bass into seamless cohesion and excellent seperation of all audible elements.

Sheer speed drivers to enable the upper mids and treble to obtain that smooth crystal clear water like texture which gives a sense of total transparency.

Then to bridge the two elements into a single indistinguishable entity.

The problem with low mass systems is that they sacrifice rigidity and power. One has the speed, but not the torque...and vise versa.

Other designs will warp the driver surface and create undesirable response somewhere else in the audible range.

Physics is a zero sum game. Perhaps when we can merge our minds into cyberspace and bypass physics, we will have perfect audio reproduction!

 

[bigshot]

That doesn't seem right. Speakers in a room have no control over pressure and the transducers are much slower than headphones, yet good speakers in a good room sound better than headphones. Do you have any links you could give me to info that supports your theory?

 

[castleofargh]

That doesn't seem right. Speakers in a room have no control over pressure and the transducers are much slower than headphones, yet good speakers in a good room sound better than headphones. Do you have any links you could give me to info that supports your theory?

He talks about speed as speed. A distance per period of time. Not sure speaker being slower than headphones is globally true in that sense, as they usually travel a much bigger distance when creating the same frequency. A big fat driver will roll off sooner in the treble than many headphones because it's heavier and has more air resisting its movement, but we can't just apply a 1 to 1 relation based only on the frequency achieved when the speaker will still have a lot more distance to travel to output the same volume level for the listener. Also in practice, many speakers do maintain their FR well into the ultrasounds, when many headphones do not. So depending on tech and size, the total displacement(speed) is what we're looking at and there might be some surprises.

Clearly, in this case the need for speed is only until we're able to reproduce a high frequency at a desired amplitude. Being able to go faster than that serves little to no purpose. And to completely stop, damping will be king(however we manage to achieve it).

 

[bigshot]

How would the speed of sound be different from one transducer to another? Wasn't he talking about the speed of the transducer (i.e.: accuracy of the exclusion)? If that's the case, speakers in a room would be a lot "slower", yet they sound better. High frequencies are only needed to the point that people can hear them. A lot of good headphones can do that well enough. And "pressure" would be different for open or closed cans, but there are good examples of both. I guess I'm not understanding what he's talking about.

 

[Harry Manback]

These are helpful replies. I currently have some mid-level offerings and am wondering about pursuing the higher end cousins of them. One is the Fostex THX00. I know it's not particularly linear, but as preference I love the sub-bass I get with them. The treble is a little hit and miss and I get what I would say are a decent level of detail out of them. This has me wondering if I really like the sound signature of these, should I go for the TH900. This got me wondering just how different the drivers in the phones were, or if it were more a matter of tuning, since the only real difference i can see from stats is a stronger magnet used in the 900 vs the X00. Both are the same 50mm biocellulose design.

I’ve owned both. TH900 are more “speaker like”. Much more controlled and capable.

 

[billqs]

At this point I own the TH900 mk1. I also owned the EMU Rosewood which would set up as the midpoint in this set of Fostex Woodies, with the THX00's as the entry model. The EMU have more control, versus just gobs of bass, and the 900 have the fullest signature low to high with bass that's also controlled. But the differences are not huge, though IMO the EMU sound closer to the 900.

When I started the thread I had an unspoken hypothesis that many higher end phones tuned lots of treble into their response in order to add "detail" which is the reason why HD800, STAX 009, and other TOTL phones can have disturbing mid-treble sibilance.

Now, I don't think I took enough into account of good driver construction- at least enough driver construction that could yield a well damped reasonably flat FR. I want to do some really careful EQ when possible, though I've acquired some awesome headphones now.

 

[bigshot]

Even midrange cans can give a balanced frequency response if you EQ. It's possible to EQ decent cans to sound significantly better than they are.

 

[billqs]

I agree with you. I plan to engage in a serious effort of EQing shortly, I hope. This thread has been helpful.

 

[gregorio]

[1] How would the speed of sound be different from one transducer to another? [2] Wasn't he talking about the speed of the transducer (i.e.: accuracy of the exclusion)? If that's the case, speakers in a room would be a lot "slower",
[3a] yet they sound better.

1. It wouldn't be.

2. Yes, I presume he meant the speed of the driver.

3. No, they (the voice coil + membrane/cone assembly) would need to be faster. They have a greater distance to travel (due to a higher output amplitude) in the same amount of time (to maintain frequency). This increases the physical problems/difficulties (such as inertia, etc.) that would cause distortion.

3a. Define "better". All else being equal (which isn't really the case), speakers in a room would have more distortion/less accuracy and therefore sound worse. On the other hand, the mix/master would have been created with speakers in a room, therefore speaker/room distortion has been accounted for in the distributed recording and reproducing the intended speaker/room distortion would sound "better".

G

 

[bigshot]

Speakers in a room may be less accurate- higher distortion, room reflections, not a pinpoint sound source... but they sound loads better than headphones... better soundstage and sound placement, more immersive, kinesthetic bass, more natural sounding overall. My point is that accuracy is a worthwhile goal, but it isn't all that is involved in getting great sounding music playback. When people say that headphone transducers might not be tightly controlled enough, I just point to speakers in a room that are much less tightly controlled, yet they sound clearly better.

 

[manueljenkin]

Speakers in a room may be less accurate- higher distortion, room reflections, not a pinpoint sound source... but they sound loads better than headphones... better soundstage and sound placement, more immersive, kinesthetic bass, more natural sounding overall. My point is that accuracy is a worthwhile goal, but it isn't all that is involved in getting great sounding music playback. When people say that headphone transducers might not be tightly controlled enough, I just point to speakers in a room that are much less tightly controlled, yet they sound clearly better.

Sounds like shifting goalposts. Something mastered on speakers will sound "pleasant" with speakers, since good headphones will show the issues with the recordings, especially the janky panning artefacts. Take any good binaural recording and play it through a headphone with quick attack (rarely does anyone measure this parameter) and CSD and you've got something that will nail the balls on any speaker for realism or spatial accuracy.

I thought this place was about sound "science".