• 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!

Review and Measurements of Benchmark AHB2 Amp

Higher capacitance cables can cause high frequencies to rolloff but those are usually runs >100m or if the source impedance is high. Amir measured this before.
Well exactly - are you using 100m runs of cable? And that is only the high frequencies.

This is not an application I've ever seen in home audio. I"ll admit I didn't qualify my statement with "in normal home audio applications"

EDIT - though I did specify "sensible designs". 100m cables for home audio use are not sensible.
 
Last edited:
I'm aware of that but the poster did say my "eyes" affected what I hear.
That is short hand with slightly humorous intent, for non-blind comparisons. The point is your knowledge of what component/cable/device is in use will introduce perceptive bias - in an unpredictable way.


(BTW If there's a clinician here who would like to perform a psychoanalytic evaluation to bring to light these subconscious biases, I'll be lounged in the listening chair ready)

Don't think of these biases as being like prejudices that can be psychoanalysed. It is more analogous to biasing a transistor - shifting the operating point up or down a curve. They are built into the way information from our senses are processed by the brain, and are part of the pre-filtering that allows us to function.

If you've seen the McGurk effect video (use the search top right if you haven't) - the mouth shape seen on screen shifts your perception of the sound from Ba to Fa and back again.
 
Last edited:
In a word, no. The AHB2's noise and distortion levels are so far below audibility that there is no audible improvement that can be achieved.

Anyone looking for "better" than the AHB2 is chasing rainbows and unicorns.

But audibly equal at a much lower price, as I point out above, that is readily available. I didn't include the point that even amplifiers that don't objectively match the Benchmark for noise/distortion etc will sound just as good (because even that level of distortion is inaudible), since my perception of @RobS 's post was he would struggle to accept that. Especially if I pointed out that even low cost amps such as a FOSI V3 mono would most likely be indistinguishable if the comparison were carried out blind, and within its power envelope.
 
But audibly equal at a much lower price, as I point out above, that is readily available. I didn't include the point that even amplifiers that don't objectively match the Benchmark for noise/distortion etc will sound just as good (because even that level of distortion is inaudible), since my perception of @RobS 's post was he would struggle to accept that. Especially if I pointed out that even low cost amps such as a FOSI V3 mono would most likely be indistinguishable if the comparison were carried out blind, and within its power envelope.
Indeed. I have AHB2s to hand, Hypex Nilai 500s, a Topping PA5II and even a Fosi ZA3 with its 88 dB SINAD. Can I audibly discern any differences between any of them?

Nope.

My typical listening levels are around 80 - 85 dB(A). Noise and distortion is below audibility threshold at this level and with a room noise floor of ~40 dB(A) there's no chance of hearing it anyway.

People think they can hear things that don't hold up once sighted bias is removed and listening levels are matched.
 
Last edited:
Don't think of these biases as being like prejudices that can be psychoanalysed. It is more analogous to biasing a transistor - shifting the operating point up or down a curve.
That's only part of it. Add to your analogy a fluctuating supply voltage that responds to unrelated influences.
 
That's only part of it. Add to your analogy a fluctuating supply voltage that responds to unrelated influences.
I'd love to read more on how cognitie/perceptive biases work. Can you suggest any layperson friendly books or web resources that would head me in the right direction?
 
Benchmark, being the heavily embedded in science and engineering sorts that they are, publish very useful application notes and system calculators on their website, this being one of them:


The audibility calculator does make some assumptions about hearing, and to be fair, Amir makes similar assumptions when discussing whether noise and distortion might possibly be audible on measured equipment. The assumptions are that any noise/distortion above 0 dB SPL is audible, even if the peak output SPL is around 110 - 120 dB. This takes into account the absolute perception range of human hearing, however it doesn't account for the effects of masking.

If you're outside in a hurricane, you're not going to detect the sound of a pin dropping.

Masking is complex and I need to do some further research, but with a 30 - 40 dB(A) noisefloor in my listening room depending on time of day, 80-85 dB(A) is more than adequate for me to enjoy a listening session without background noise distraction, which begs me to ask the question, how good a SINAD do we really need such that noise and distortion are inaudible relative to the music?

The ~110 dB or better SINAD yardstick unequivocally guarantees audible transparency, but where is the real threshold beyond which improvement is measurable, yet still inaudible, even to the most golden-eared?
 
Last edited:
I'd love to read more on how cognitie/perceptive biases work. Can you suggest any layperson friendly books or web resources that would head me in the right direction?
It is really not my field and would probably take me as much effort as it would take you to sift one out. Sorry.
 
Benchmark, being the heavily embedded in science and engineering sorts that they are, publish very useful application notes and system calculators on their website, this being one of them:


The audibility calculator does make some assumptions about hearing, and to be fair, Amir makes similar assumptions when discussing whether noise and distortion might possibly be audible on measured equipment. The assumptions are that any noise/distortion above 0 dB SPL is audible, even if the peak output SPL is around 110 - 120 dB. This takes into account the absolute perception range of human hearing, however it doesn't account for the effects of masking.

If you're outside in a hurricane, you're not going to detect the sound of a pin dropping.

Masking is complex and I need to do some further research, but with a 30 - 40 dB(A) noisefloor in my listening room depending on time of day, 80-85 dB(A) is more than adequate for me to enjoy a listening session without background noise distraction, which begs me to ask the question, how good a SINAD do we really need such that noise and distortion are inaudible relative to the music?

The ~110 dB or better SINAD yardstick unequivocally guarantees audible transparency, but where is the real threshold beyond which improvement is measurable, yet still inaudible, even to the most golden-eared?
Note that Amir’s thresholds of audibility are orders of magnitude higher than for people listening to music. He is purposely choosing the highest thresholds to free himself of any question from golden ears, and he is himself trained in the specific (and specialized) techniques used to hear distortion artifacts. These often require things like massive gain-riding during the tails of reverberation as a means of elevating artifacts normally down in the noise to audible levels. There are other techniques, too, that I don’t want to know about.

Also, his thresholds for DACs are upstream of amplification. If DAC signals are amplified by, say, 40 dB, noise at -90 dB gets amplified to -50 dB, which might be audible in narrow circumstances. Speaker/headphone amplifiers are not further amplified and an amp with a S/N ratio of 80 dB already puts the noise threshold well below ambient noise in most environments even when the loud bits are ear/splittingly loud.

At the level of nearly all amps made for a long time, improved SINAD is more a sign of good general engineering and testing that a discriminator when listening through real speakers in a real room. We admire it for the same reason we might admire a sports car with a top speed utterly irrelevant to any reasonable application.

TL;DR,: 70 dB is probably still a strict target threshold for loudspeaker amplifiers in real-world applications.

Rick “can hear quiescent noise that low only by putting an ear to the tweeter with no music playing” Denney
 
There are electrical differences between L-4E6S and 1800F, but these won't result in an audible difference in sound (unless there's something very strange with your DAC).

Both are fine choices for analog line level audio. Microphone cable is completely fine for line level audio, but it's over specified in that application. Simpler, cheaper, and thinner cables are ok for line level use (and no need for star-quad, but it doesn't hurt anything). FWIW, I use the Canare L-4E6S too. It's easy to buy from pro sources and flexible. I don't care that it's overkill.

As an example, Belden 8541 is for line level signals. The foil shield isn't preferred by some since it can cause some issues you don't get with a braided shield. In comparison to both of the above microphone cables, this is only 0.138 in (3.51 mm) in diameter. The microphone cables are a lot bigger.
An interesting analysis that BJC did regarding various cable geometries.

 
Masking is complex and I need to do some further research, but with a 30 - 40 dB(A) noisefloor in my listening room depending on time of day

That seems like a high noise floor to me. But we are on acreage. Right now, there's a little rain dripping in the guttering and I can just hear the refrigerator in the kitchen. Other than that, nothing. 31.3dB(A) slow. 33.8dB (A) fast. SPL meter.
 
That seems like a high noise floor to me. But we are on acreage. Right now, there's a little rain dripping in the guttering and I can just hear the refrigerator in the kitchen. Other than that, nothing. 31.3dB(A) slow. 33.8dB (A) fast. SPL meter.

I'd love for my room to be much quieter. There's a road 15 metres from the front window and we're more or less under Heathrow's flightpath.

Apart from that, half the houses in the neighbourhood seem to be building extensions right now.....
 
Last edited:
I'd love for my room to be much quieter. There's a road 15 metres from the front window and we're more or less under Heathrow's flightpath.

Apart from that, half the houses in the neighbourhood seem to be building extensions right now.....

I finally got around to doing the Acourate driver linearisation measurements today for my active setup. For the bass driver on the speaker nearer the front window, I had to turn on the -20dB PAD setting on the mic input of my Motu Ultralite Mk 5, and max out the volume to knock down the ambient noise and get an acceptable sinc pulse recording.
 
There is a used Benchmark for $3,148.95 at B and H Photo. It has been there for several months.
That is not a big surprise. You can get it direct from Benchmark for 3500

B&H need to drop the price at least 1K
 
Used prices for AHB2 are usually around $2,100-2,200, so yeah that's way overpriced. I just sold mine for $2,000.
Thank-you for letting me know. If the warranty isn’t transferable then the price should be in the low $2000s. I’m sure even without the warranty it is a good value.
 
Back
Top Bottom