Blind Test Part Trois: Schiit Magni Unity vs Topping L50

Earlier in the year, I conducted a blind level matched test between Magni+ and Heretic, both 103db "SINAD" amps. Read here: https://www.superbestaudiofriends.o...tic-vs-magni-with-statistical-analysis.13192/

In theory, I should have not be able to tell the difference, but I did. The test procedure was kind of clunky with an assortment of cables and a primitive switch. I'm sure the "objectivists" cried foul and moved the goalposts as always, pointing out that 103db "SINAD" may not have be sufficient for transparency or the manual procedures used for switching the sources was inadequate, i.e. my kids or wife who did the cable switches telepathically beamed what they did with the cables to my brain. This time around, I wanted to be more sure, so I used a machine, the Van Alstine ABX comparator to conduct the randomized ABX tests.

I did this really quickly. The way the Van Alstine ABC comparator works is that it lets the listener go back and forth from Test 1 to Test 8 at their leisure for a long as they want. For each test, the amp A or B is randomly selected beforehand for the set of 8 once the formal "test" button is depressed. I decided to speed through like mad man, as if I hated the tests. Normally I take practice sessions, but I felt pretty confident I could tell the difference. Modius E, a high SINAD DAC was used as the source. The outputs of the devices were adjusted with a 1kHz signal so that they were within 0.1db. My JAR600 was used for the test. One of the reasons for this is that the L50 seemed to have issues driving 26.4-ohm loads into 2Vrms with one channel, but no problem with driving 300-ohm loads. Both amps were selected because of their high "SINAD". The Unity no better than 115db. The Topping 121db. Per objectivist theory, both are transparent and one should not be able to tell the difference.



In less than a minute I cycled back and forth through all of the the tests and jotted the results down:


I got one wrong, test #6. I guess I shouldn't have rushed.

 

More results. I can't seem to get a perfect score. Always 7 of 8. I did go back and check my work this time around.

 

I don't know how I did so well when I didn't give a crap. Going slower made me doubt myself, or put me into a mood where I tried incessantly to tell the difference, clicking A, B, X, B, A, B, X, etc. Chalk it up to stress or fatigue. I did take a few practice sessions with a quick blind test before I began the format tests. In some ways, I may have made things harder for myself to discern the differences by relying on more subtle rather than more obvious aspects (I like to challenge myself). More on this last though as I offer some observations.

1. The more I focused on any one particular aspect of sound presentation, say soundstage (location of vocal of percussion), tone, microdetail (trailing decays), microdynamics (gradations of sound levels), the less I was able to hear the other aspects. For example, if I focused hard on listening to microdynamics, differences in soundstage would become oblivious to me. If I focused on listening to soundstage, then differences in microdetail would be ignored. I guess this is how the brain works. I noted some aspects are interrelated. For example, the soundstage of the Magni Unity had a stronger and more upfront center. This concentration of the image partially contributed to a difference in tone.
2. The Magni Unity had a more full bodied tone. The Topping L50 was thinner sound, but not by much. Both amps exhibited warmth. The L50 exhibited a hint of grain in the highs, but this as an indicator wasn't anywhere as significant as the Magni Unity's more full bodied tone.
3. After the practice sessions, I realized many misidentifications were the result of music not being static and ever changing. For example, keying on tone in one passage of music would lead to erroneous results if next musical passage changed register. I learned to go back in the music player and start as the same timestamp or do the AB changes within a section of music that didn't change much.
4. The Topping L50 simplified microdynamics. While the Unity isn't as good as the better Class A single-ended topologies, it made the L50 sound stiff. This is a tough one to explain. I've used the words nuanced, expressive. The best music reproduction systems will "dig" into the music, play back an infinite gradation of volume levels, and within even a second of music, exhibit "microchanges" in volume, and do that in an instantaneous way (we are not talking speed of transient response but rather volume level change response). It's this quality that makes me unable to put the headphones down. The opposite of excellent microdynamic rendering or expressiveness is what I often termed "flat and boring". The Magni Unity did better on this. The L50 missed out on those "microchanges". That being said, I found the L50 better than expected in microdynamic performance.
5. Flat and boring is the domain of amplifier topologies with high global negative feedback, nested feedback loops, and inherently high open loop gain. The advantage of this is low distortion. Make no mistake, I do not like high distortion. However, I only need distortion to be low enough. After a certain point, making distortion any lower won't make any difference. I'd rather keep the microdynamic expressiveness with low-enough distortion rather than lose the microdynamic expressiveness with distortion so low where it's magnitudes (10, 100, 1000) lower than I can detect.
6. In the end, I used a threefold analysis to identify the amp:
   1. Listen to differences in microdynamic (and related microdetail) rendering. This worked best when starting from the same track position, and resting ears, before every switch from A to B to X. It was also the most arduous procedure.
   2. Listen to differences in tone. The pitfalls with this approach are documented in #3 above.
   3. Finally, what I would almost consider a cheat: differences in soundstage. The Magni Unity has a strong center image that is upfront with wider sides a littie bit behind. The Magni Unity headstage is shaped like a U. The L50 is more diffuse with a very slightly narrower stage without as much depth, shaped like this -
7. The L50 actually isn't bad sounding. I found the tonal signature agreeable. There wasn't any bad solid-state highs. Microdetail and microdynamics were good enough for the price.

Final thoughts.:

• I very much liked slowly evaluating the amps by employing the formal test mode with the randomized #1 to #8 tests. This allowed me to select a random test and simply listen to enjoy music. I could do so at my own pace, changing the test number from #1 to #2 or #6 to #5 every twenty minutes or two minutes. Whatever I wanted to do. I found this an excellent way for me to pick out the amp that I preferred most. Using this this lower pressure leisurely approach, the less expressive sounding amp, the L50 was very reliably switched out. Sometimes within a matter of seconds, sometimes in a minute. I may do a random switch, and then I realize: wait, what's happened? And when I will switch to another test number. I go will ah, that's better, but then 30 seconds later, oh no it ain't, that sucks. And then switch to another, finally going ahh, that's right. (Keep in mind that the test #s are totally randomized. They are not toggled to go back and forth).
• I am not convinced that ABX testing is the best way for ordinary people to evaluate audio gear. Pressure and fatigue could be contributors to false readings. If I have learned anything about discerning differences in ABX tests: practice helps a lot, it's crucial to stick to a regimen in applying steps (not skip or rush), take breaks - not do rapid AXBXAB switching when things get tough (there's reason why we palette cleanse when comparing wines)
• Really, just hide two pieces of gear behind a curtain, label them Q and R, give people a headphone, a switch to choose the amp, and external volume control. Double blind ABX only makes sense if you have a belief system and/or are self-righteous enough to believe that other people cannot hear better than you. For most people, keep it simple. It's a subjective hobby. It's supposed to be a fun hobby, not religion that deals with absolutes.
• It's very easy to fail the tests. Say for example that I believed that all amps with more than 110db SINAD (using ASR's SINAD parameters) sounded the same. Through confirmation bias, I can refuse to try hard to discern the differences and simply give up, taking wild guesses at it. Or I can simply not be able to tell the difference. We've already seen people pull off this shit: I, the holy one, identified the correct one in the first 4 or 5 tests, but later failed the next five, therefore no one else can identify the correct one.

 

For sure.

 

despite Marv's nice reply, I'm going to ask a few basic questions:

1. Why can't you include the word "please" or a subtly pleasantry to indicate he is not subject to some beckon whim?
2. What benefit does a pair of ears on Marv's head being able to detect deltas between DAC A vs. B impart as a benefit to your audio life?

If having someone detect differences between audio equipment is a popcorn watching, entertainment challenge - fine.
but no ABX score card reveals the sonic descriptors that might help YOU triangulate about DAC A vs. B unless you have the exact same chain and the same track list and speakers, and room, and headphones.....

There is no doubt that better ears here more "differences" that help us mere readers make informed decisions.
but a ABX 7 outta 8 just indicates Marv's skills a more valid "indicator" that deltas can be heard, BY HIM!
not by you. or me. or the next person on to read this post.

am all for more impressions, comparisons and pairing recommendations.
ABX is a better game show during meets to discern who can hear any deltas (not preferences) IMO.

 

The slow leisurely exercise while I put the machine into formal test mode was a form of preference testing, but with stronger safeguards so that I wouldn’t be fooling myself.

I surmise that maybe @Eric Rosenfield wants me to stick it more to the man (the pope of the SINAD religion)?

Anyway, I’d like to point that I do not have any special abilities, only a lot of practice. It’s like driving a car. OMG I thought my daughter was going to kill me when I was first teaching her how to drive. (I taught her on the Jeep which isn’t an easy vehicle to learn on). Now I feel confident with her. Doing well on ABX testing is a learned skill that takes practice. Even then, to do well, one needs to prepare, get into the mood. It’s like taking the SAT or GMAT test.

Yes. I was thinking it would be a lot of fun at meets. Too much pressure, too loud, not enough preparation, not enough time, for valid results. But certainly a lot of fun!

Although keep in mind I purposefully pick challenging stuff. Tonally, the L50 and Magni Unity are quite close.

Preference testing is easily supported by the box in quick blind test mode. The indicator lights A or B get turned off.

At the end of the day, this is for game shows, with Richard Dawson.