xnor
Headphoneus Supremus
- Joined
- May 28, 2009
- Posts
- 4,092
- Likes
- 227
Regarding crosstalk, I've created some test files: http://rapidshare.com/files/430363572/pink_r_mp3.zip
This kind of confusion is caused by people who make exaggerated claims, amp A is superb and significantly better than amp B. These headphones need to be amped or else you can clearly hear they are not performing at their best. In reality the differences are tiny.
Look at the rmaa test I linked. Without load (like in this setup: cowon s9 - amp - headphones) dynamic range is 90.7 dB, with 32 ohm load it still is 90.4 dB. A difference that small is negligible.
So there is no real loss of dynamic range.
Frequency response will only be slightly rolled-off at the lower end, dunno if you will even notice this.
The only thing that really suffers is stereo crosstalk. If you need/want the left/right channel to be as much separated from each other as possible then adding an amp makes sense.
Quote:This kind of confusion is caused by people who make exaggerated claims, amp A is superb and significantly better than amp B. These headphones need to be amped or else you can clearly hear they are not performing at their best. In reality the differences are tiny.
The thing is, all differences are small at first. A lot of people say you won't be wowed at first listen most of the time. Only by going back to your old setup will you hear the differences. So people who have gone and upgraded to an amp, or a better amp, or a dac, etc. - even cables, seem to exaggerate the differences they hear. Because, from their point of view, the difference really is big, but to someone who hasn't upgraded that far, headphones straight from an amplifier can still sound great. Everything seems to be small differences, the deciding factor is how much extra hassle and money you want to go through to achieve those small differences.
What people are referring to when the mention using a better amp than what's inside a portable player is, the ability to reproduce a highly complex signal and drive a load with frequency-varying impedance in a linear manner.
Music usually is a bunch of sinusoidal waveforms with frequencies ranging from 20 Hz to 20 kHz. That's not complex compared to the signals that are used for measurements.
I don't see a problem with frequency-varying resistance. In fact, amps should be happier with a load that is more efficient and needs less current near the resonant frequency (this is the case with headphone drivers) than compared to a low, fixed resistance, which is a tough load regardless of frequency.
That's one of the reasons why you can assume that an amp will need to supply roughly a quarter or third less current with headphones as load instead of resistors.
This is not the same as a test program playing a frequency sweep through the device (though the measurements do have some relevance). I agree though that the improvement may not be noticeable with all music. Fast and complex music will become noticeably congested when the amplification can't keep up.
The rhythm and speed of music has nothing to do how simple or complex the signal is and how easy/hard it is to reproduce. (see above)
Tracks where the recording has a wide soundstage will sound like everything is coming mostly from the middle. Instruments and voices will be blurred together when you try and visualise the performance. With higher-end equipment the audible soundstage appears wider and individual instruments or performers appear more as distinct entities on the stage. It's not to say, however, that one can't enjoy music with it all blurred together, as this is what most people are used to.
Aka crosstalk or channel separation measurement. And eliminating unnatural stereo separation can make a lot of sense, see crossfeed.
Originally Posted by xnor /img/forum/go_quote.gif
Music usually is a bunch of sinusoidal waveforms with frequencies ranging from 20 Hz to 20 kHz. That's not complex compared to the signals that are used for measurements.
Aka crosstalk or channel separation measurement. And eliminating unnatural stereo separation can make a lot of sense, see crossfeed.