Review and Measurements of Benchmark AHB2 Amp

[RichB]

Yea, but there's a fly in the ointment, we don't have a choice in much of the music we love. We have to take it the way it comes.
Very much a large part of the criticism of the sound of early CD, all of a sudden we were able to hear how bad many recordings really sounded. Spitty, sibilant vocals, cymbals that sound like bursts from an air hose, etc, etc.
Kind of the double edged sword that has been the path of HiFi over the last 70 years in general. The better our systems got, the worse much of our cherished recording were revealed to sound. Then to make it worse, over the last few decades although much has improved, the loudness wars has murdered both the remastered classics and modern recordings.
You got to be a masochist to be an audiophile.

Compression and digital clipping is rampant in popular recording, regardless of bit-rate and depth.
I'd like to see audio follow video's lead and introduce HDR (High Dynamic Range) and other master quality metrics.
There were some great CD recordings done in the early 80's before the loudness wars.

Mastering by Mongo (Blazing Saddles).

Adele Hello CD

Adele Skyfall HDTracks (96/24)

Unfortunately, the recording industry has managed to produce another "standard", HD Audio, that provides no indication or apparent correlation to quality.
You'd think digital clipping would be disqualifying.

It's no wonder folks are buying vinyl.

- Rich

 

[GrimSurfer]

I think that in-home listening and comparisons have value.

In the end, we all do it. Even the "objectivist". It's just that we feel comfortable in the rightness of the choices.

- Rich

I think this is right. The objectivist in us will pour over specs, confirm these through independent sources, and create a component "short list" within a given budget. If done correctly, the theoretical performance of each of the components on the list will be within 5% of each other. Listening tests then help narrow down to "the one".

Even if one uses a listening test (sighted or blind) to choose the runner-up, the objective differences are likely to be small as long as a disciplined approach to listening is followed.

 

[GrimSurfer]

Compression and digital clipping is rampant in popular recording, regardless of bit-rate and depth.
I'd like to see audio follow video's lead and introduce HDR (High Dynamic Range) and other master quality metrics.
There were some great CD recordings done in the early 80's before the loudness wars.

Mastering by Mongo (Blazing Saddles).

Adele Hello CD
View attachment 28290

Adele Skyfall HDTracks (96/24)
View attachment 28291View attachment 28292

Unfortunately, the recording industry has managed to produce another "standard", HD Audio, that provides no indication or apparent correlation to quality.
You'd think digital clipping would be disqualifying.

It's no wonder folks are buying vinyl.

- Rich

Agree... and Love the Blazing Saddles reference.

 

[maty]

Unfortunately, the recording industry has managed to produce another "standard", HD Audio, that provides no indication or apparent correlation to quality.
You'd think digital clipping would be disqualifying.

It's no wonder folks are buying vinyl.

- Rich

One hour ago:

Ólafur Arnalds - Re:Member (2018), Vinyl, Mercury KX, Europe

https://www.discogs.com/Ólafur-Arnalds-Remember/release/12430618

-> https://www.amazon.com/gp/customer-...ef=cm_cr_dp_d_rvw_ttl?ie=UTF8&ASIN=B07D57FB2W

Update: Amazon kindly exchanged the disc and regrattabky similar problem with the second disk. Meanwhile Kaismos CD that arrived at the same time is just fine. So definitely poor production quality.

Original review: The factory that was contracted to manufacture the CD ("Made in the EU") has very poor quality controls. The CD is unbalanced and skips with some of my CD players, while notably vibrates in the portable ones. Just to make sure, I checked my other six Olafur CDs (from the Erased Tapes times) and they are fine. This is quite disappointing that great music is pressed into bad discs.

 

[RichB]

One hour ago:

Ólafur Arnalds - Re:Member (2018), Vinyl, Mercury KX, Europe

https://www.discogs.com/Ólafur-Arnalds-Remember/release/12430618

-> https://www.amazon.com/gp/customer-...ef=cm_cr_dp_d_rvw_ttl?ie=UTF8&ASIN=B07D57FB2W

If you can rip it once, you are all set. I don't spin music anymore. Music library management software has arrived.

- Rich

 

[Sal1950]

But I actually don't think that you yourself truly believe that sighted listening is completely unreliable, Sal... As I have understood it you have a preference for horns, based on things like perceived dynamics, immediacy, etc. Is this based on measurements and DBT's? Or is it because you have listened to horn systems, and figured out that you like them?

Oh sighted listening is absolutely reliable if your looking to suite a preference. That is something we all do at a minimum when we pick our speakers. We all know and can very easily show how different they sound with measurements and gain some insight into their "real world, in-room" response.. In 2019 audio, and for at least a few decades prior, we have pieces of a audio kit that for the most part have been transparent for a long long time, and it's the debate on these components where if you want to make claims, you have to close your eyes and bring science to the forefront.
After that, I not only pick my speaker by preference, but also my house curve via dsp, if and what software I might use to upsample 2ch stereo to 5.2.4 immersive, theres more. Some will detest these choices, others agree. But that is MY choice to make. I don't want hardware manufacturers making them for me, and I want scientific methods used to determine if a DAC or amp has a sound of it's own (some do by design) not some keyboard kowboy who says, "I've heard it, so it is so".

 

[oivavoi]

Oh sighted listening is absolutely reliable if your looking to suite a preference. That is something we all do at a minimum when we pick our speakers. We all know and can very easily show how different they sound with measurements and gain some insight into their "real world, in-room" response.. In 2019 audio, and for at least a few decades prior, we have pieces of a audio kit that for the most part have been transparent for a long long time, and it's the debate on these components where if you want to make claims, you have to close your eyes and bring science to the forefront.
After that, I not only pick my speaker by preference, but also my house curve via dsp, if and what software I might use to upsample 2ch stereo to 5.2.4 immersive, theres more. Some will detest these choices, others agree. But that is MY choice to make. I don't want hardware manufacturers making them for me, and I want scientific methods used to determine if a DAC or amp has a sound of it's own (some do by design) not some keyboard kowboy who says, "I've heard it, so it is so".

Cool, I understand your position better now. Agreed.

 

[amirm]

At the same time, I think it`s too simple to just discard sighted long-term listening out of hand. Humans have managed to develop quite a few things by using ear alone. Musicians, conductors, piano tuners etc - they all work (mostly) by ear, and honestly it has worked quite well. Last time I checked, the world of acoustic music was relatively well-functioning. This doesn't mean that anybody's listening is infallible, of course. But sighted listening does provide data points, at least. For those of you who are AES members, here's a fascinating paper by a couple of the engineers in Genelec: https://secure.aes.org/forum/pubs/conferences/?elib=19621

I just read the paper. It doesn't say much about sighted listening being any good. It does start with this telling graph which I have mentioned countless times:

Out of megabits/sec of of audio data, the brain only captures a few bits/second. No way, no how does the brain capture for the long term minute details in music. You would go crazy in real life if you had to remember such nuance in my voice as we were talking for example. Or the background noise in your home or environment.

So anyone who says they listened to something for weeks, then switched to something else for weeks, is just wrong.

As to the thesis of the paper, the conclusion does it in: that most of what they are talking about is good practices summarized in ITU BS1116, "Methods for the subjective assessment of small impairments in audio systems." That is, familiarity and training matters when it comes to testing.

And yes, "slow learning" is super important. But it has another name: training. It took me about 6 months to get trained to hear the smallest amount of lossy compression artifacts. This does NOT at all say that when conducting listening tests, I need 6 months, or even 6 minutes to hear differences. Indeed, switching times over 1 second severely limit my ability to detect small artifacts.

The long process for hearing small artifacts may be needed to find the right artifacts. Sometimes it takes me half of hour of comparisons to figure out what to focus on. But once there, the actual comparisons occur in milliseconds, not "long term."

And again, there is nothing in there about sighted listening being more correct. Yes, you can do the training sighted, but the final test better be blind to mask other factors getting into the test.

 

[oivavoi]

Thanks Amir, that's a great response!

But I think you may be selling them a bit short with regards to how much their recommendations are already implemented:
"Based on a limited perceptual bandwidth and 8 hours of dedicated listening per day, getting to know a room and equipment in any detail would take at least a week, but assuming years would be safer. This would also be compatible with what we have discussed with many audio professionals, and a story multi award winning mastering engineer Bob Ludwig told at the 2015 AES convention in New York: He avoids listening at different levels because of the time it takes to mentally re-calibrate. Instead, clients are invited to use headphones in case they wish to listen louder. Elaborate aspects of trained listening are given in [94], and ITU-R BS.1116 provides more extensive practical recommendations with regard to subjects, screening, training etc. than the casual ITU-R BS.1534. However, we might still underestimate the time required for pre and post listening learning and fatigue assessment, or at least do not emphasize the importance of various temporal elements strongly enough."

I don't think I have read many published psychoacoustic tests where the testees spent that much time getting trained. Might that be a reason why many of the positive blind testing results for "difficult" stuff like high-rez, amplifier difference etc usually have been made in smaller and more intimate listening environments, with testees who know the testing equipment intimately? I'm thinking about you, LTS, and others.

And I certainly didn't mean to imply that "sightedness" itself was a positive thing. I just meant that we can't rule out that familiarity with auditory input over time can actually help us to hear things we didn't notice immediately. Once that familiarity is established, though, I agree that blind listening with short excerpts is the best way of determining whether the difference is actually there.

 

[GrimSurfer]

Out of megabits/sec of of audio data, the brain only captures a few bits/second. No way, no how does the brain capture for the long term minute details in music. You would go crazy in real life if you had to remember such nuance in my voice as we were talking for example. Or the background noise in your home or environment.

And yes, "slow learning" is super important. But it has another name: training. It took me about 6 months to get trained to hear the smallest amount of lossy compression artifacts. This does NOT at all say that when conducting listening tests, I need 6 months, or even 6 minutes to hear differences. Indeed, switching times over 1 second severely limit my ability to detect small artifacts.

The long process for hearing small artifacts may be needed to find the right artifacts. Sometimes it takes me half of hour of comparisons to figure out what to focus on. But once there, the actual comparisons occur in milliseconds, not "long term."

A few considerations before ppl start making some very liberal assumptions:

1. Diction is important. Does the brain capture only a few bits per second or does it process a few bits per second? The minute vibration of the hairs inside the cochlea results in the transmission of electrical signals to the brain. These signals continue regardless of whether the brain processes them or not.

2. The brain may only process a few bits per second, but which bits? Different ones each second? This is important because it implies a variability and specificity into to equation.

Agree with training. Very important because it facilitates critical listening.

Now, without training how indicative is ABX testing likely to be? Probably not very inductive, which explains why testing untrained people won't lead to very much.

 

[SIY]

Now, without training how indicative is ABX testing likely to be?

Depends on what you're looking for. I'm not trying to be flippant- without training, I was able to identify pretty small frequency response, level, and pitch changes. For things like data compression algorithms, I did better using a sorting test format.

 

[GrimSurfer]

Depends on what you're looking for. I'm not trying to be flippant- without training, I was able to identify pretty small frequency response, level, and pitch changes. For things like data compression algorithms, I did better using a sorting test format.

Understood. Keep in mind that you have a high level of technical awareness and experience.

I think the word "training" is often misinterpreted as coaching. It may be more useful to think of it as calibration.

If you had a test device that was dunked in water daily (bath/shower), supplied with energy of varying quality and quantity (nutrition), bumped, vibrated, placed in hot and cold chambers (car, aircon home), periodically exposed to the elements, and subjected to age related deterioration, it would be reasonable to recalibrate (train) it before expecting it to be used with any degree of precision, accuracy, or reliability.

 

[RayDunzl]

40 bits per second?

That seems unrealistic to me, unless they are really big bits.

Maybe it's slangified. "I know a bit about what I did yesterday."

Bits of what, anyway?

Are we digital?

I think the diagram is stupid, so, there.

Or maybe it is just missing some bits.

--

Looking again.

Five senses, 40bps, 1 byte per sense per second?

I don't think so.

 

[GrimSurfer]

40 bits per second?

That seems unrealistic to me, unless they are really big bits.

Maybe it's slangified. "I know a bit about what I did yesterday."

Bits of what, anyway?

Are we digital?

I think the diagram is stupid, so, there.

Or maybe it is just missing some bits.

--

Looking again.

Five senses, 40bps, 1 byte per sense per second?

I don't think so.

I was just about to reformat my digital music library from uncompressed to 40 Bps. Imagine all of the space I would have saved at no cost to fidelity.

 

[RayDunzl]

I say, call your colleagues at Boston Dynamics and see what the overall input data rate is for Atlas...

 

[amirm]

That seems unrealistic to me, unless they are really big bits.

My bits are always big bits:

 

[amirm]

I don't think so.

What parts of the sound spectrum did you capture from the last person who said something to you?

 

[RayDunzl]

What parts of the sound spectrum did you capture from the last person who said something to you?

Having inoperative HF acuity I don't need as many kibbles or bits to store the bands I heard.

Wait a minute.

What was the question?

 

[Sal1950]

My bits are always big bits:

How does it taste?

 

[zalive]

How does it taste?

Bit me.

My biased expectations for Amir's response which won't likely happen