Review and Measurements of Benchmark AHB2 Amp

[scott wurcer]

DBT is the sine qua non. Whether you use ABX or a different format is up to you. If the method is not double blind, it is hopelessly flawed.

As usual too much to lose.

 

[zalive]

I guess nothing changes all the verbiage and wiggling to not actually do a blind no peeking test.

Proper blind test requires some equipment, which is not what a typical consumer typically has in his home.
I'd gladly participate in any blind test which someone prepares for me.
I'm too lazy to do such preparations for myself. Besides, I find them completely unimportant.
Should my impressions come from bias, I've no problem with that. As long as I'm satisfied with listening experience. I don't care why it's so.
There are my subjective expectations, there is my taste, all being personal and subjective. So why should bias be a problem, on top of that?

But I'll tell you why I think the bias is not decisive in my case. Because in a numerous occasions I was expecting one result from my bias (expectations from the status of the product or manufacturer, price, topology/architecture, comments from the others), but my listening impressions didn't comply to any.
So when your listening impressions go the other way compared to your biased expectations, you learn to trust your hearing.
Still I would not claim I'm bias free, but as I've said, I've no problem with that.

 

[SIY]

Proper blind test requires some equipment, which is not what a typical consumer typically has in his home.
I'd gladly participate in any blind test which someone prepares for me.
I'm too lazy to do such preparations for myself. Besides, I find them completely unimportant.
Should my impressions come from bias, I've no problems with that. As long as I'm satisfied with listening experience. I don't care why it's so.
There are my subjective expectations, there is my taste, all being personal and subjective. So why should bias be a problem, on top of that?

But I'll tell you why I think the bias is not decisive in my case. Because in a numerous occasions I was expecting one result from my bias (expectations from the renome of the product or manufacturer, price, topology/architecture, comments from the others), but my listening impressions didn't comply to any.
So when your listening impressions go the other way compared to your biased expectations, you learn to trust your hearing.
Still I would not claim I'm bias free, but as I've said, I've no problem with that.

Yes, there's always an excuse. But this forum is devoted to the science of audio, and if you want to have anything you say treated seriously, you need to do basic controls. Lazy doesn't cut it.

 

[zalive]

I think it is difference of viewpoint. The AHB2, which went into protection due to transient response, had a maximum input of + 22dBu and a low load at the same time. Unlike an amplifier used in a concert, amplifiers used in homes and studios are not common in transient input situations. I think it is more appropriate to protect the user's speaker or ear when entering the protection mode when the transient input occurs.

But it is uncertain whether Benchmark's primary concern was to protect users speakers, ears, or to ensure a perfect benchmark (measurements).

 

[Hugo9000]

Do you think a possible bias is the only possible reason for not correctly hearing the difference?



I would need to dig into a particular methodology to find whether it has obvious flaws.
So far I'm pretty certain ABX methodology is completely flawed for the purpose of hearing a difference. I proposed a better one, here.

In addition to bias (unconscious or otherwise), there could be other factors such as mood, fatigue, illness affecting hearing acuity due to pressure or other factors, distractions, and so on. (Naturally, these particular negative factors could also affect the outcome of blind tests, not just sighted ones.)

 

[restorer-john]

I ordered another AHB2 last week and will perform measurements under the same conditions for performance testing. If you have a test method you would like to see, please respond. After the test, I will post the results. However, my AHB2 still can not start shipping due to lack of stock. It will take a very long time to come to Korea where I live.

I'd like to see one third rated power preconditioned for one hour into an 8 ohm resistive load at 1KHz both channels driven. Then do the 8/6/4/3 ohm continuous tests. 5 minutes minimum. Both channels driven.

Hopefully the new unit/version you get will not only meet these specifications, but exceed them.

 

[Blumlein 88]

Proper blind test requires some equipment, which is not what a typical consumer typically has in his home.
I'd gladly participate in any blind test which someone prepares for me.
I'm too lazy to do such preparations for myself. Besides, I find them completely unimportant.
Should my impressions come from bias, I've no problem with that. As long as I'm satisfied with listening experience. I don't care why it's so.
There are my subjective expectations, there is my taste, all being personal and subjective. So why should bias be a problem, on top of that?

But I'll tell you why I think the bias is not decisive in my case. Because in a numerous occasions I was expecting one result from my bias (expectations from the status of the product or manufacturer, price, topology/architecture, comments from the others), but my listening impressions didn't comply to any.
So when your listening impressions go the other way compared to your biased expectations, you learn to trust your hearing.
Still I would not claim I'm bias free, but as I've said, I've no problem with that.

Yes, you do learn that. Unfortunately what you learned isn't really true. It is why you'll think no blind listening is needed. I sometimes go the other way because what I'm hearing is so. The big problem is sighted long term listening impressions are simply not reliable. You'll get it right sometimes, sometimes there was no difference so either preference you had was a myth. Sometimes your bias is turned on by things you don't notice as being the cause. There are differences so large you'd be able to hear them. Usually those will not require long term listening.

Then suppose you are hearing something real. Let us say you inadvertently sometimes clip amp A. You try amp B and it doesn't clip. You listen and it becomes apparent over time because you only sometimes do it listening to exuberant levels to a few tunes. Not all the time. So you decide amp B sounds so smooth, and unflappable you prefer it. Or with other preferences you might decide it is sterile and lifeless and attribute all sorts of subjective descriptive terms to what you hear. Yet even when you might call this hearing a difference, because you did, and it did only occur over longer listening you are not too right. The answer isn't to produce a laundry list of descriptive terms of one vs the other. The answer is to not clip the amp. At which point they sound the same. Swapping gear to find the amp that clips just so would be a curious pursuit especially with so much clean inexpensive power available in the modern world.

 

[reza]

Long term non-blind non-level matched comparisons are pretty good to find your preference.
You'll be listening to pretty various levels and content and you'll get a picture which makes you more satisfied.
So, what is the truth here? Something other than personal satisfaction, preference, taste?

I don't have a dog in this race and have no expertise in audio or acoustic engineering. But I wanted to point out that from a perceptual point of view, the above statement is not without merit. After a few exposure to stimuli from the same family, your brain will quickly learn to approximate its underlying generative process - think of estimating a distribution from multiple samples. From there on, you can quite effectively recognize stimuli that don't belong in the same family. See figure 1 of this paper for a quick demo.

Applied in this context, this could mean (whether it actually does, I don't know) that multiple exposures over the long run may in fact obviate the need for level matching and other control criteria.

 

[agtp]

I don't need to know anything in particular about you to answer this quick reasoning.
If listening to two amplifiers long term under sighted conditions, you may "believe" you prefer the sound of one amp over the other.
But it might have nothing to do with the sound. You may just prefer the looks of one amp over the other and the human bias conditions take over. Under sighted conditions there are any number of reasons you might be fooled into believing one is your sound preference.
I need to know nothing about you to know you suffer the same human weaknesses as us all.

Sal, lecturing others on human bias and being fooled? LOL. Following the “do as I say not as I do” philosophy?

 

[agtp]

They're terrified actually. They know the whole subjective audio world is built on a house of cards and with the removal of just one or two, the whole thing will collapse. They hang together making unsubstantiated claims of DBT's being fatally flawed without ever being willing to publicly test their position. There are at least a couple offers I know of with money on the table that no one will step up to.

Hi Sal,
How about you? Are you “terrified”? Are you willing to put your beliefs to the test? Or do you make exceptions for certain closely held beliefs, for whatever reason? Projecting, maybe? I know of one of those offers of money on the table, you mentioned. Did you step up to claim it? Feel free to PM if you think you have anything convincing, I’d love to hear it.

https://en.wikipedia.org/wiki/One_Million_Dollar_Paranormal_Challenge

 

[Blumlein 88]

Hi Sal,
How about you? Are you “terrified”? Are you willing to put your beliefs to the test? Or do you make exceptions for certain closely held beliefs, for whatever reason? Projecting, maybe? I know of one of those offers of money on the table, you mentioned. Did you step up to claim it? Feel free to PM if you think you have anything convincing, I’d love to hear it.

https://en.wikipedia.org/wiki/One_Million_Dollar_Paranormal_Challenge

Huh???

 

[Blumlein 88]

I don't have a dog in this race and have no expertise in audio or acoustic engineering. But I wanted to point out that from a perceptual point of view, the above statement is not without merit. After a few exposure to stimuli from the same family, your brain will quickly learn to approximate its underlying generative process - think of estimating a distribution from multiple samples. From there on, you can quite effectively recognize stimuli that don't belong in the same family. See figure 1 of this paper for a quick demo.

Applied in this context, this could mean (whether it actually does, I don't know) that multiple exposures over the long run may in fact obviate the need for level matching and other control criteria.

There may be something to this. I think it would apply to rather large signal differences however. I think level matched and blind you can get positive results for smaller signal differences.

An example, over the years I've been surprised at how many audiophiles listen to amps in occasional overload or even mild clipping. I don't have any reason to think I'm especially sensitive or attuned to it. Various amps under various loads clip in somewhat different ways or more accurately recover from clipping in different ways. So I suppose I've learned from various single shot examples and can hear it even when the specifics to something I've never listened to exhibit enough signs for me to perceive, "hey this is clipping now and again". I even know of a fellow who seemed to only stop turning up volume when he began to cause a similar level of overload across several different systems. Like he wanted 1-3% clipping or something.
But clipping is a pretty gross difference in signal. Level matched testing can ferret out far smaller differences than that.

I might even believe it possible for two amps or systems with perceptible frequency response deviations. Then again, such deviations can be much more finely discriminated with good testing or even the right signal. Music is rather poor for that unless the difference is large. Listening to a few seconds of pink noise at moderate levels will quickly let you hear that kind of difference at much smaller differences than you'd hear long term with music.

 

[restorer-john]

Hi Sal,
How about you? Are you “terrified”? Are you willing to put your beliefs to the test? Or do you make exceptions for certain closely held beliefs, for whatever reason? Projecting, maybe? I know of one of those offers of money on the table, you mentioned. Did you step up to claim it? Feel free to PM if you think you have anything convincing, I’d love to hear it.

The topic of this thread is the Benchmark AHB-2, Amir's review with discussions and measurements related directly to that.

Just a heads up, OK?

 

[reza]

There may be something to this. I think it would apply to rather large signal differences however. I think level matched and blind you can get positive results for smaller signal differences.

An example, over the years I've been surprised at how many audiophiles listen to amps in occasional overload or even mild clipping. I don't have any reason to think I'm especially sensitive or attuned to it. Various amps under various loads clip in somewhat different ways or more accurately recover from clipping in different ways. So I suppose I've learned from various single shot examples and can hear it even when the specifics to something I've never listened to exhibit enough signs for me to perceive, "hey this is clipping now and again". I even know of a fellow who seemed to only stop turning up volume when he began to cause a similar level of overload across several different systems. Like he wanted 1-3% clipping or something.
But clipping is a pretty gross difference in signal. Level matched testing can ferret out far smaller differences than that.

I might even believe it possible for two amps or systems with perceptible frequency response deviations. Then again, such deviations can be much more finely discriminated with good testing or even the right signal. Music is rather poor for that unless the difference is large. Listening to a few seconds of pink noise at moderate levels will quickly let you hear that kind of difference at much smaller differences than you'd hear long term with music.

I agree, one-shot or few-shot learning can only happen if the novel category differs from the familiar ones by a lot. For finer discrimination you will need many more examples stretched over a calendar period. I do wonder if casual listening to my own amp over a long time would suffice for such fine discrimination.

 

[amirm]

But I'll tell you why I think the bias is not decisive in my case. Because in a numerous occasions I was expecting one result from my bias (expectations from the status of the product or manufacturer, price, topology/architecture, comments from the others), but my listening impressions didn't comply to any.

Bias is only one factor that interferes with reliable results. The most common error is that you listen differently to A versus B. Your brain has the ability to focus on detail or not. When you know the identity of products, you will tend to fall victim to this thinking one device has higher fidelity, etc.

In controlled testing the same sample can be presented to you twice, catching the above phenomena. Listeners routinely provide different scores to the same content when differences are very small.

 

[amirm]

Above is why if you "hate" a device, you can still fall in love with it in testing. You listen more attentively, you hear more detail, and you think it is a better piece of audio.

 

[maty]

Please don't be ridiculous. Do you think non-level matched listening tells you anything comparing amps for sound? Let me answer that: no. If you disagree, express your reasoning and you can learn from this...

The ridicule is yours because I have not said anything of what you say in that little joke to SIY. You prepend your biased vision to go beyond what I have written.

It is logical that levels have to be equal. Do you know someone who denies it? But also that you have to use recordings and music that allow the listener to appreciate the possible differences in their reproduction. Hence the reference to that Metallica album with DR3.

 

[maty]

An example, rock vs orchestra in a test.

https://www.audiosciencereview.com/...mance-of-my-denon-pma-2500ne.7850/post-189374

by PMA (in diyaudio thread)

[ ...If this distortion is applied to a rock music (I tried well recorded Godwhacker by Steely Dan in 96/24), it is almost impossible to tell that the distortion was added.

If I added the same distortion to Beethoven's No. 9 symphony Movement No. 1, the distortion is immediately audible in the low level passage just after the beginning.

So the test files should be chosen very carefully and they should contain high dynamic range and both loud and silent passages, the music that is always close to full scale level is not good for testing. ]

 

[maty]

OK, so you don't have a single example.

If orangejello, in the same audio system, with only the amplifier -with very good measurements- as a differential element, with equal levels and well recorded acoustic music with great dynamics, he is able to appreciate different depth so we will have a verification.

With current commercial music it will be very difficult to appreciate the difference between those two amplifiers and others that measure worse, I say. Depth, tonality or something else.

 

[Frank Dernie]

Long term non-blind non-level matched comparisons are pretty good to find your preference.
You'll be listening to pretty various levels and content and you'll get a picture which makes you more satisfied.
So, what is the truth here? Something other than personal satisfaction, preference, taste?

This is a widely proposed view but actually I think this serves only as a way to make you feel that you have made a good choice in your own mind but there is actually no way to make comparisons of any validity in the long term IME.
I haven’t changed much of my main kit for 20 years, though I have messed around with other stuff and have secondary (and other) systems.
During these 20 years I sometimes think to myself whilst listening how fantastic it is sounding.
I actually have no idea why, could be a particularly good sounding recording I am listening to (IME the recording quality is massively more important for SQ than the hardware we use to play it on which makes a mockery of Hi-Fi in a way anyway), it could be my mood, some furniture I have moved or something, but the one thing it can not be is a change in my Hi-Fi, since there hasn’t been one.
If I had bought something new I could easily believe it was better than my old system by evaluating long-term and I would be completely wrong.
I used to think the idea had some truth in it but now know there isn’t, one of the things I learned over the last 50 years of playing around with Hi-Fi.
The only way I have found which tells if there really is a difference is precisely level matched direct and immediate comparison.