Class D vs tube at the listening position

[dougi]

After recently purchasing a New (Very Old) Stock Audio Research amp from a store and noting how I subjectively liked it in comparison with my existing amp (March Audio P502 Hypex), I thought I would take some comparative measurements at the listening position to look for any differences. Subjectively, punchier bass and sweeter treble.

My system setup is as per the below figure.

For measurement, a PC uses the MiniDSP USB input in place of the streamer, and a Focusrite Scarlett 2i2 interface with a Sonarworks calibrated mic at the listening position.

The listening position is about 4m from each speaker. I estimated power levels to be around 1-2W out of each amp, levels roughly matched only.

For these tests, DIRAC was left on between each amp, then differences looked for. Ignore SPL figures, they were un-calibrated.

FREQUENCY RESPONSE

In both, you can see the high output impedance of the ARC and the speakers impedance causes a softening in the upper treble and upper bass. Likely the main cause of the possibly perceived difference. Does a depressed upper bass cw low = punchier? I don't know.

DISTORTION

At the modest levels, which are my typical listening levels, there is a tad more distortion with the ARC in the bass, but none in the treble. Note the lower treble hump in the right is due to differences between my two speakers.

AFTER DIRAC FOR THE ARC (old DIRAC for March, new DIRAC for ARC)

After re-running DIRAC with the ARC in place, with simlar but not identical targets, the FR difference largely go away, as expected. Subjectively, perhaps with biases, the sound between them is closer too.

OTHER
I didn't note any significant differences in the time domain. For information a typical DIRAC-corrected step response for the Lintons is below.

 

[DMill]

Wow. I’m a bit surprised. I’m certainly not an expert but it all looks very comparable between the two. I notice when I push my tube amp hard vs. SS I hear the latter being cleaner. I’d be curious if you see that if you dial it up a bit.

 

[restorer-john]

Great investigation, write-up and analysis.

Very nice ARC amplifier you have there too.

 

[dougi]

Wow. I’m a bit surprised. I’m certainly not an expert but it all looks very comparable between the two. I notice when I push my tube amp hard vs. SS I hear the latter being cleaner. I’d be curious if you see that if you dial it up a bit.

Yes, I'm sure the distortion on the ARC may be more in the treble if tested at higher SPLs. I may do that at some point, with better level matching. This time, it was to see any differences at my normal listening levels.

 

[DMill]

Yes, I'm sure the distortion on the ARC may be more in the treble if tested at higher SPLs. I may do that at some point, with better level matching. This time, it was to see any differences at my normal listening levels.

Please post if you do test. But all said, thank you, and congrats on a nice ARC find.

 

[lashto]

After recently purchasing a New (Very Old) Stock Audio Research amp from a store and noting how I subjectively liked it in comparison with my existing amp (March Audio P502 Hypex), I thought I would take some comparative measurements at the listening position to look for any differences. Subjectively, punchier bass and sweeter treble.
...

very useful test @dougi, thank you!

If you feeel like doing more tests, a HD-spectrum comparison would be very interesting: i.e. a H2-to-H5 harmonics-graph like shown in ASR speaker tests. Or just a 'simple', standard THD graph at 1kHz.
(maybe you can produce such graphs from the measurement data you already have.)

Why do I think that the THD-spectrum comparison would be relevant? The THD-spectrum differences should be the "subjective differences" that you are hearing (see the spoiler below)

Spoiler: math

A bit of quick math to check/confirm the THD-numbers you measured at the listening position: i.e. those THD numbers should be ~same between your tube and classD amps.
To make the math somewhat easier we can round/simplify the THD numbers a bit:

• THD speakers: -40dB (simple average between the -20dB in the bass area and -60dB in the treble area of a typical/good pair of speakers)
• THD tube-amp: -60dB. (btw, I assume yours is something like this.)
• THD classD-amp: -120dB

THD at your ears with the tube-amp = -40dB (speakers) + -60dB (tube-amp) ~= -39.9 dB THD.
The sum could actually be anywhere between -39.9 and -40.1. But anyway, it's almost same as the speakerTHD number.

THD at your ears with the classD-amp = -40dB + -120dB ~=39.9..9x dB THD.
Pretty much same as the speakerTHD number.

AFAICS, the math fully confirms your THD measurements. But it's always very helpful to verify. And much easier to read a graph .. many thanks again!

Many people call tube-devices "distortion factories" but the above math and the OP measurements show that 'name' to be quite wrong: i.e. tube-amps will not change the distortion/THD numbers that reach your ears.
Calling tube-devices "harmonics factories" could be a lot more appropiate: i.e. the harmonics/THD-spectrum comparison should show big(er) and easily visible changes between your tube and classD amps.

And those THD-spectrum differences should be the "subjective differences" that you are hearing.

P.S.
I am not an EE and in case I made some errors, math corrections are highly welcome!