Quote from: ultrabike on October 13, 2013, 06:13:44 AM

FR comparison (seems a Vali+HD558 is a better match than an Valhalla+HD558).

It shows the Focusrite has a higher output impedance. That may also be partly responsible (in addition to the higher level) for its slightly higher bass distortion.

Yes, the Focusrite I have might have a slightly higher impedance.. Here is a thread with results for the Valhalla+HD558 combo, since I was comparing the Vali+HD558 combo to it:

Valhalla+HD558 combo (from above thread for reference):



As stated in that thread, the HD600 is a better match for the Valhalla, but not so great with the HD558.

I understand the mechanism for linear distortion due to headphone output impedance (the "damping" deal or voltage divider between the amp output impedance and the headphone load). I think non-linear distortion may be dominated by the headphone driver characteristics. As far as the amp goes, non-linear distortion might be more a function of the headphone load/sensitivity (current/voltage) demands and what the amp can deliver, than the amp output impedance in these cases... maybe some small contribution is due to the amp's impedance, but not sure about the magnitude.

In an acoustic measurement, the microphone could easily have a higher noise floor than the headphone amplifier.

Possible. Solderdude has mentioned that the mic preamp might also contribute. FWIW, the microphone I use is the EMM-6. I can try and do some single tone THD measurements to see how low the noise floor goes.

Focusrite vs. Vali are too close to call. Vahalla is another story. Lots of D2 - expected since the Vahalla is subjectively a more obviously colored amp (and not necessarily in a good way).

Valhalla+HD558 combo (from above thread for reference):

This seems to have some more significant FR differences in the midrange. But since it is an older measurement, if the headphone/microphone positioning is not exactly the same (in other words, if the acoustic test rig was touched between the amp measurements), the difference could be simply the result of that.

As far as the amp goes, non-linear distortion might be more a function of the headphone load/sensitivity (current/voltage) demands and what the amp can deliver, than the amp output impedance in these cases... maybe some small contribution is due to the amp's impedance, but not sure about the magnitude.

A high output impedance does add some non-linear distortion, basically due to the amp not being able to "damp" the distortion of the driver (which is fed back to the amp output through back EMF) as well. The magnitude of this effect varies, but about a tenth of the acoustic distortion on a sufficiently high impedance amplifier output - with a damping factor of ~2 or less - may be a "typical" level. Perhaps it is roughly comparable in magnitude to the FR effect.

It's not pointless if we are comparing amps with a specific headphone - which is what we are doing here. We've already established with HD558: Focusrite vs. Vali (too close to call objectively, subjectively maybe Vali more detailed, punchier, wider-stage?) and vs. Vahalla (really different - both subjectively and objectively.)

I am of the rather simplistic opinion that a peak level measurement (these plots all are) won't show what's been added by the amp while our hearing evaluates on MORE than just a single aspect (peak levels of certain harmonics) so yes... pointless after all except when the distortion levels are higher than that of the headphone, the amps that do might not sound that good though, or perhaps they do when it is quite euphonic.

The subjective assets you speak of cannot be caught in a few plots nor say anything of how it behaves with dynamic music which isn't just 80dB and 90dB (and or 100dB) especially with valve amps in case they do not have overall feedback.

It's just another way of looking at things. One could make a counterargument that measuring amps by themselves is pointless or misleading because most of the amp's distortion on the analyzer will of little consequence because it will be masked by the transducer's distortion.

I agree that it is another way of looking at things.
One could make that argument you just made but would be discarding how the brain and ears work.
For instance if I drown a single tone in level below a white noise you would see zilch with most measurements, yet with the right techniques, or ears, you can still make the tone out because of certain correlations it has that won't show in the measurements, thereby nullifying the counter argument.

The earlier measurements you made (O2 vs Magni I believe) showed NO differences at all yet the subjective differences were claimed otherwise.
That would have been the perfect moment to conclude that a simple FR/distortion measurement using a headphone+rig won't disclose how an amp sounds.

The fact that you have shown there are differences between O2 and Vali tells me the Vali performs differently from the O2 which is quite logical and obvious.
Fact remains that the distortion plots show it measures 'worse' than the O2 which is contradictory to the subjective impressions which are favourable for the Vali.
THIS I do not doubt for a moment by the way.

Too bad there are no plots of what the actual amp does, which would be more revealing to me personally but seem to stand alone in this.
After all it is a VERY intriguing amplifier being discussed here for a GREAT price and with a support and name that is very hard to rival yet the measurements shown are of a few headphones not every one has.