bigshot
Headphoneus Supremus
It's a coloration that sounds good to you. Nothing wrong with that, but it isn't balanced. It might not work well with acoustic music.
Is it a colouration? Do we know for sure that headphone manufacturers voice their headphones using zero output impedance sources? Not trying to be argumentative here, just genuinely curious.It's a coloration that sounds good to you. Nothing wrong with that, but it isn't balanced. It might not work well with acoustic music.
I've been experimenting with the different output impedance settings offered by the AT-HA5050H headphone amp. It gives the choice of 0.1, 33, 82 and 120 ohms.
1. I'd assumed going from 0.1 ohms to 120 ohms would turn my HD600s into a loose, uncontrolled and bloomy mess but I was wrong!
2. The bass has indeed plumped out a little, but I'm more surprised by the effect it's had elsewhere. It seems to have improved the soundstage and fluidity, and has pushed the mids back a little whilst giving the top end a sweet and shimmering wetness.
3. I really like the effect, but I'm confused because I thought that output impedance mainly affected bass response, and also headphones that have large impedance swings.
4. The HD600's impedance curve between 1kHz and 10kHz is as flat as a pancake, though it does begin to rise slightly above 10kHz. Does this explain the effect I'm hearing?
5. I feel rather naughty because I am clearly violating the "less than 1/8th" rule between load and output impedance.
1. It's up to the headphone what happens when output impedance changes. Some headphones are more sensitive than others.
2. The bass boost with HD-600 caused by output impedance is: 0.0 dB (0.1 Ω), 0.4 dB (33 Ω), 0.9 dB (82 Ω) and 1.2 dB (120 Ω). So, nothing dramatic really. Higher output impedance seems to have some positive effects on the soundstage, but I don't fully understand why. Because headphones usually have their maximum impedance value around 100 Hz (resonance frequency of the driver) and also have rising impedance in the high frequency due to voice coil inductance, higher output impedance causes frequency response to change boosting 100 Hz area and also the highest octave or so. The sound becomes uncontrolled when the system damping ratio (not damping factor which is a different thing!) is less than one. For Sennheiser HD-600 the system is critically damped (damping ratio = 1) when the output impedance is 150 Ω. For 0.1 Ω, 33 Ω, 82 Ω and 120 Ω output impedances the corresponding system damping ratios are 1.184, 1.130, 1.066 and 1.027 meaning slightly overdamped systems. Again, nothing dramatic. As long as the output impedance is 150 Ω or less you have "controlled" sound from technical point of view.
3. True, but how much is a different question. For HD-600 output impedance of 120 Ω is pretty ok, but for HD-598 it would be catastrophic.
4. Yes, it rises above 10 kHz, because voice voil inductance kicks in. This probably explains partly what you hear.
5. The "less than 1/8th" rule is very general/coarse and following it ensures you don't go completely wrong with any pair of headphones. For HD-600 you need "less than half" rule. Why? Because the mechanical damping and impedance curve of the heaphones say so!
High impedance should basically cause the diaphragm to stop moving less quickly, while damping does more or less the opposite, which is to limit sound reflections and reverberations after the signal has stopped. I wonder if the increase in soundstage is due to the default HD6x0 tuning being overly damped, masking the spatial cues (sound reflections in recorded room) in favor of textural detail.
I tend to doubt that. Channel separation is the main thing for reproducing soundstage in headphones, and that's generally not a problem.
Most modern audio components have inaudible levels of channel spill. That hasn't been a problem since LP records.
Thanks, you are welcome.Thank you for an awesome post! If I may ask you a few questions:
1) How do you calculate the dB bass boost of a particular headphone for different output impedances?
2) How do you calculate the output impedance at which a particular headphone is critically damped?
Most modern audio components have inaudible levels of channel spill. That hasn't been a problem since LP records.
The problem with soundstage is that the term is so often misused in audiophile circles that the meaning has gotten lost. Soundstage is the placement of sound precisely in space. That is only possible with speakers, because headphones put sound between your ears, not ahead of you in physical space. "Headstage" is what we talk about with headphones. That is a straight line through the middle of the ears from right to left. The only way to mess that up is to mix the channels, which never happens.
There are secondary depth cues built into mixes... specifically room reflection and differences in level that indicate distance... but those are no harder to reproduce than any other part of the signal. You would have to introduce an awful lot of distortion to start messing with that, and it would make the rest of the music sound bad just as much as it would the depth cues.
Thanks, you are welcome.
1) You need output impedance Zout and the min/max impedances Zmin/Zmax of the headphone. Assuming all of these very resistive the frequency response error is Lmax - Lmin, where
Lmax = 20*log(Zmax / (Zmax + Zout))
Lmin = 20*log(Zmin / (Zmin + Zout))
2) Zout_critical_damping = (sqrt(L / C)*(R + Rc) - 2*R*Rc) / (2*R - sqrt(L / C)), where
L = compliance of diaphragm suspension on electric sideWhere do I get these parameters? I get the measured impedance curve of a headphone online and simulate it with an equivalent electric circuit. I iterate the parameters until I get a very good match with the measurements.
C = diaphragm mass on electric side
R = mechanical resistance of diaphragm on electric side
Rc = voice coil resistance on electric side
I guess I would assume that the impedance specs are the target for a balanced response. If you're set up for it, you could try testing the response and see how it lines up. Do you have any headphones that you know have a relatively flat response to compare? When you're describing differences, it helps to describe them as specifically as possible. Generalized poetic descriptions assume a conclusion and don't really communicate anything. Better to talk about specific balances in frequency ranges. A tone sweep might reveal what is being affected.