• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Blind A/B test between (Benchmark AHB2+Topping D90SE/Pre90) and Hegel H190

OP
P

phion

Member
Joined
Nov 10, 2021
Messages
58
Likes
100
55-65 dBSPL is above the average for you? Wow lol.

I thought i was on the responsible side for keeping peaks (0dBFS) on about 85dBSPL.

You’re basically running the amplifiers on below 100mW if we take the sensitivity of the speakers as 85dBSPL/W and we took the high end of that 55-65 dBSPL range.

To put that in perspective, Amir’s SINAD measurements are at 5W, your use case is almost in the realm of headphone amplifiers. :p
:) The speakers are 92dB efficient but I think that will not change much in what you said. The trouble is due to cable length limitation I can't sit farther away from them (which I normally do).
 
OP
P

phion

Member
Joined
Nov 10, 2021
Messages
58
Likes
100
55-65 dBSPL is above the average for you? Wow lol.

I thought i was on the responsible side for keeping peaks (0dBFS) on about 85dBSPL.

You’re basically running the amplifiers on below 100mW if we take the sensitivity of the speakers as 85dBSPL/W and we took the high end of that 55-65 dBSPL range.

To put that in perspective, Amir’s SINAD measurements are at 5W, your use case is almost in the realm of headphone amplifiers. :p

Edit: the sensitivity of the speaker is actually 90dB/W.
also I just recorded the peaks of the song using the max tracking function of the SPL meter at 72dB. I think that still sits below 100mW :)
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
I didn't mean to pressure you or invalidate you in anyway btw. All experiences are valid, it just shows how far solid state electronics have gone since 70 and 60 years ago when we were using tube amplifiers.
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
also I just recorded the peaks of the song using the max tracking function of the SPL meter at 72dB. I think that still sits below 100mW :)

Depends on music and level you are listening to and on speaker sensitivity. With some classical music with big differences between ppp and fff, I have recorded peaks up to 100W. I may repeat it and post here, sometimes.
 

amper42

Major Contributor
Joined
Dec 21, 2020
Messages
1,582
Likes
2,284
I enjoyed your comparison testing. As more people take the time to complete detailed blind tests they start to realize when you can hear a difference and when you can't. That can be absolutely eye opening. I have learned not to accept reviews as relevant until my own blind test can confirm a positive audible difference. This leads to my second rule, never buy anything that can't be easily returned. :D
 
Last edited:

amper42

Major Contributor
Joined
Dec 21, 2020
Messages
1,582
Likes
2,284
Yes, I also agree with Kal; though this was intentional. I should probably reword the title of the thread to make this explicit.


SPL for the first track was between 55-65dB with my sound meter (not a very good one tbh) at the listening position, which was about 2m away from the tweeters; Being a relatively low volume listener, this was actually slightly louder than how I usually listen.


As I mentioned, I bought the new gears for a second system in my home office; so I intend to keep both actually. I will be getting Revel M105 speakers for that room, which will arrive soon so I may also repeat the test with those speakers as well.

If you want another test, compare the Revel M105 to the BMR Monitors. I have both speakers. The BMR Monitors are a clear winner in my comparison. They sound cleaner and more open than my M105. The difference is easily recognizable. I have both pair in my office and always find myself gravitating toward the BMR.
 

anmpr1

Major Contributor
Forum Donor
Joined
Oct 11, 2018
Messages
3,722
Likes
6,405
OTOH, I do not think that a typical "magazine or pro reviewer" has any more time (always deadlines) nor the ability to hire help (unheard of, afaik).
Now, don't get me started with "time on his hands." o_O
Early on, Gordon Holt had the best response. He simply told subscribers that they would get the magazine (or what he called a magazine) whenever they got it. If that meant Winter 1970 showed up in your mailbox late Fall, 1971, then that was that! I think his 'help' was his wife. Not sure he paid her. And if he had anything in his hands, it was a Marlboro. :)

An ad financed mag could even splurge for a power cube ( which is outside Amir’s budget ) and see if something odd shows and then go find a real world speaker that provokes the amps under test in that manner .
Peter Aczel owned one, yet in spite of the outrageous cost, it broke down. What good is that? I don't know if he ever got it fixed, or what became of it.

The only other time I encountered 'reviews' incorporating Power Cube measurements was in a Boston Audio Society Speaker. They put the machine on a handful of amps. I need to go back through the issues to find out which one it was. Might not be one of their 'free' downloadable samples.

BAS CDs containing back issues are available, but you might have to subscribe to get it. They make interesting 'historical' reading. Other than that, ASR is the place to go for measurements of newer stuff.

Like most things 'old-school' audio-related, the BAS Speaker has become a shell of what it once was. I don't blame them. Interest in the hobby among old timers has waned. Most of the 'problems' that were once considered important have already been solved; either that, or have been found not to be problems at all, but grounded in imagination. Electrical gear, having converged sonically, is therefore uninteresting for the most part, or if it is interesting (even to look at), it's because of its relative cheapness for what you get, or in rare cases because it is aesthetically appealing..., but then it's typically priced like expensive jewelry.
 

Geert

Major Contributor
Joined
Mar 20, 2020
Messages
1,937
Likes
3,525
This is, of course, not a test to distinguish between the two amps but between two different DAC, volume control and amp configurations. FWIW.
Which makes me conclude that when you don't even notice replacing half of the audio chain it's probably almost impossible to hear a difference between the individual components.

I did a comparable test a few months ago. Switched one integrated amp with integrated DAC for another. 2400€ versus 6000€. The first one with AKM DAC chip and pretty average SINAD, the second one with an ESS Sabre based DAC and being a top SINAD performer. No difference whatsoever. Maybe a small difference in the sub low, but can't tell reliable. No veils being lifted, no darker backgrounds, no lower noise floor and wife in the kitchen couldn't care less.
 

rdenney

Major Contributor
Forum Donor
Joined
Dec 30, 2020
Messages
2,235
Likes
3,856
Reviewed perhaps but this is an apples-to-oranges comparison.
It seems to me that if the amps could be perceived differently, the DACs and preamps would either have to mask that difference or offset it exactly. I have a hard time imagining the latter. If the Topping combination at least is tested to be transparent with no masking ability, and there is no perceived difference in the two stacks, then I’m running out of ways to think of how this test would be skewed by the different front ends. If there were differences, the front ends might explain them.

But then we have the speakers, which could also have masked amp differences, and that seems to me more likely, if we are to believe there must be differences unperceived.

I’ve long believed that quiescent hiss and clipping behavior are the only differences anyone would be likely to notice in a real room, at least with manageable speakers, given that frequency response these days is so flat. But I also think clipping happens more often than we think, and I see few test protocols that actually demonstrate that the amps were operated fully within their power envelopes. And when clipping is evaluated, all amps under test are turned down to that level, which to me is unrealistic. I’d rather test at the level of the loudest use case, but then explain any differences using measurements, including clipping. Determine that an amp is too small for a stated use case is a useful outcome, given how much power affects price. (Obviously, the power issue is not relevant to this relatively low-level test.) A microphone recording of the test would be too flawed to show many things, but the two setups could at least be compared for clipping and compression behavior using a microphone recording.

Rick “balancing implausibilities” Denney
 
Last edited:
OP
P

phion

Member
Joined
Nov 10, 2021
Messages
58
Likes
100
But then we have the speakers, which could also have masked amp differences, and that seems to me more likely, if we are to believe there must be differences unperceived.
I was also thinking about the effect that the speakers could have had on masking any differences. Coincidentally my new Revel M105 also arrived today, so I re-randomized the cables and re-ran the test with the same set of tracks. I increased the volume by 5dB on Pre90 to -28dB gain and set it to 40 on Hegel. Their voltage output match within +/-10mV. I spent more time on certain tiny bits of each track (3-7 second segments) where I thought I would be more likely to hear a difference, going back and replaying much more often than the first time now that I got a hang of it.

1- Grimes | Vanessa | N
2- Nina Simone | Sinnerman | N
3- Air | Cherry Blossom Girl | N
4- Beethoven | Symphony No. 5, Carlos Kleiber Wiener Phil, Allegro con brio | N
5- Celeste | Tell me Something | N
6- Dire Straits | Money for Nothing | N
7- Moloko | The Time is Now | L
8- Daft Punk | Give Life Back to Music | (had to skip*)
9- Michael Jackson | Billie Jean | (had to skip*)
10- Judas Priest | Pain Killer | N

* for some reason Roon was acting up on these two tracks (maybe because of their 88khz sample rate) and not letting me smoothly going back and forth, so I decided to skip them.

The only track where I did have a preference was Moloko's where the segment I repeatedly listened to was right at the beginning and I felt the left system had a tiny bit more ring to a couple of bass guitar notes. I tried this like 4-5 times and I was convinced there was a genuine difference. After that I revealed who's who and it turned out the L system was the Benchmark.

I then listened to that same segment several times with each amp with louder volume, but I am not very sure if the difference is even there when I do this sighted. I will re-randomize the cables tomorrow and recheck this track again in another blind test.
 

Juhazi

Major Contributor
Joined
Sep 15, 2018
Messages
1,717
Likes
2,897
Location
Finland
Which makes me conclude that when you don't even notice replacing half of the audio chain it's probably almost impossible to hear a difference between the individual components.

I did a comparable test a few months ago. Switched one integrated amp with integrated DAC for another. 2400€ versus 6000€. The first one with AKM DAC chip and pretty average SINAD, the second one with an ESS Sabre based DAC and being a top SINAD performer. No difference whatsoever. Maybe a small difference in the sub low, but can't tell reliable. No veils being lifted, no darker backgrounds, no lower noise floor and wife in the kitchen couldn't care less.
So you are listening in the kichen nowdays? ;)
 

Kal Rubinson

Master Contributor
Industry Insider
Forum Donor
Joined
Mar 23, 2016
Messages
5,272
Likes
9,778
Location
NYC
It seems to me that if the amps could be perceived differently, the DACs and preamps would either have to mask that difference or offset it exactly. I have a hard time imagining the latter. If the Topping combination at least is tested to be transparent with no masking ability, and there is no perceived difference in the two stacks, then I’m running out of ways to think of how this test would be skewed by the different front ends. If there were differences, the front ends might explain them.

But then we have the speakers, which could also have masked amp differences, and that seems to me more likely, if we are to believe there must be differences unperceived.

I’ve long believed that quiescent hiss and clipping behavior are the only differences anyone would be likely to notice in a real room, at least with manageable speakers, given that frequency response these days is so flat. But I also think clipping happens more often than we think, and I see few test protocols that actually demonstrate that the amps were operated fully within their power envelopes. And when clipping is evaluated, all amps under test are turned down to that level, which to me is unrealistic. I’d rather test at the level of the loudest use case, but then explain any differences using measurements, including clipping. Determine that an amp is too small for a stated use case is a useful outcome, given how much power affects price. (Obviously, the power issue is not relevant to this relatively low-level test.) A microphone recording of the test would be too flawed to show many things, but the two setups could at least be compared for clipping and compression behavior using a microphone recording.

Rick “balancing implausibilities” Denney
Sure, sure. My comment was a result of my years of experimental design where increasing the number of variables increased the difficulty in obtaining a statistically valid conclusion. If the OP had originally posted this under its current title (in response, I think, to my comment), I would not have made the comment.

Now, it is quite another (although interesting) thing to ask how different two almost randomly chosen device chains would render the sound when placed between a common source and a common pair of loudspeakers in a room. With modern quality electronics, the differences are small and getting smaller.
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
However the sighted test method remains unreliable when speaking about “differences”. Nobody is completely free from (unwanted) bias.
 
OP
P

phion

Member
Joined
Nov 10, 2021
Messages
58
Likes
100
I guess at this point I am looking to find a way to try ABX tests, but I don’t see an easy way to accomplish that. I could ask my partner to connect two of the ports of the AB switch, but it’s a bit difficult to hide that, and I think discontinuities in switching will probably give it away. I could still always first switch to the unused port and then switch to one of ABX, but that’s a bit cumbersome.
 

jae

Major Contributor
Joined
Dec 2, 2019
Messages
1,208
Likes
1,508
A better test would be to compare short dynamic passages. And not necessarily picking preference versus identifying if you can recognize a difference, if any.
 

JRS

Major Contributor
Joined
Sep 22, 2021
Messages
1,158
Likes
1,003
Location
Albuquerque, NM USA
Like others have said, well done! And you identified the one major weakness--once you have picked R, more likely to do it again--confirmation bias? Not that it matters as the conclusion was pffft as to huge differences so many claim.

The test I'd like to see is to take some inexpensive but solid DAC, a Purifi or Hypex amplifier costing no more than 750.00, and give that a run against the Hegel, with random R/L connections.
 
OP
P

phion

Member
Joined
Nov 10, 2021
Messages
58
Likes
100
A better test would be to compare short dynamic passages. And not necessarily picking preference versus identifying if you can recognize a difference, if any.
I should say in my second run with Revels, that’s pretty much what I focused on. In all but one track I could simply hear no difference whatsoever.

But I think ABX is a better format for this kind of test.
 
Top Bottom