Starke Sound AD4.320 Review (Multichannel Amplifier)

[nstzya]

I agree that the manufacturer could have handled the AD4.320 "half-power" disaster much better in many different ways. I also agree that how that disaster was handled says something about the company and influences my desire to buy from them. But do we need to play Bash the Manufacturer?

The reason I got sucked into this thread was that someone perceived some underhanded action in the way the specs were presented. That's something I've come across before and I really struggle to understand it. If a manufacturer shows their product from its best angle, they're cooking the books. If they're not publishing my pet spec, they must be hiding something and they always hide something for nefarious reasons. It is impossible that the manufacturer didn't publish data for my pet spec because they consider that spec to be irrelevant or ill-defined (take headroom for example). If the reviewed sample performs better than spec, the manufacturer must have shipped a golden unit for the review. How am I as a manufacturer supposed to work within this environment?

Tom

I think if you had the time to read the whole saga, you’d come to understand that this was a particularly egregious example with a dash of international intrigue. It definitely got under several owners‘ skins. And the company’s response was particularly lacking and Inflammatory.

That said, it appears to have had some effect in curbing their wild claims with this model and that is a positive outcome.

Out of curiosity, what is your “pet spec”? And why is it so?

 

[Mack]

Well, the SMPS at least removes all margin of error regarding wrong voltage settings

So are they using the same modules with a switch mode power supply if so can the old model be modified by replacing the power supply

 

[RichB]

Power specs usually don't. The power is measured mid-band (usually at 1 kHz).

Is it? Many Class D amp guys measure the max power at 10 % THD. Some of us measure at 0.1% THD or below.

I would argue that SINAD at 1 W isn't particularly useful. THD at 1 W would be more meaningful or better yet show me a harmonic spectrum for 1 W output. But few do that.

Yeah. A measurement of the noise floor would be better. But that's not commonly done. Also, had they measured SNR with an 8 Ω load they could possibly have gotten an even higher number as the amp is likely able to produce slightly higher voltage into a lighter load.

How so? To me it looks like "we need a spec table but we don't want to put in any effort" with a bit of marketing blather (fan noise, true balance (sic) circuit) added.

Output current: 16.5A/33A (bridge) and higher output power into 2 Ω (BTL) make no sense. I suspect the "BTL" mode isn't actually bridged but rather two channels operating in parallel.

Tom

The decision measure at 1kHz is to game the specs. I don't know how common 20Hz to 20kHz specification is among manufacturers but, personally, I don't trust any manufacturer that does not supply them. Certainly, many quality brands seem to supply full frequency range power and distortion specs.

A spec for 1 kHz at the point where an amplifier is hitting a clipping wall is not indicative of its performance in normal listening. I don't see how that is a controversial statement. The character of the distortion is also important.

SINAD at 1 watt or below are very important for most speakers, increasing with speaker sensitivity.
I have measured the Salon2s in my room and the pair deliver 86 dB (at 220Hz and kHz) with 2 watts (because they are essentially 4 Ohms speakers).
Listening late last night, the SPL was not exceeding 76 dB, so well under 1 watt.

- Rich

 

[nstzya]

So are they using the same modules with a switch mode power supply if so can the old model be modified by replacing the power supply

Probably a better question for Starke (including costs). My guess would be probably not cost effectively…

 

[tomchr]

Out of curiosity, what is your “pet spec”? And why is it so?

I don't personally have a "pet spec". But many others do. You see requests for SNR, THD, or SINAD at 1 W, 5 W, 100 mW for example. "How much headroom" is another one, never mind that headroom is a meaningless spec. I'm not thereby saying that someone's "pet spec" is not relevant, ever. Sometimes it is relevant. But few companies specify THD at 1 W, for example. Does this mean they're hiding something? Or does it just mean that they didn't think to include it?

I don't enjoy seeing people getting screwed. Equipment should meet the published specs. But I also don't enjoy seeing a manufacturer's reputation being completely trashed and accusations such as fraud being made because of a single incident, even if it was poorly handled.

Tom

 

[Mack]

Probably a better question for Starke (including costs). My guess would be probably not cost effectively…

I will inquire and report back

 

[tomchr]

The decision measure at 1kHz is to game the specs. I don't know how common 20Hz to 20kHz speciation is among manufacturers but, personally, I don't trust any manufacturer that does not supply them. Certainly, many quality brands seem to supply full frequency range power and distortion specs.

You can figure out the amount of power delivered at 20 Hz and 20 kHz if you have the 1 kHz number and the frequency response. Why is a separate spec needed?

I don't see how a mid-band measurement is "to game the specs". It's one spec out of many. I agree that it doesn't tell the full story, because no single spec does. The mid-band measurement shows the best case performance. If the performance degrades significantly above or below mid-band the amp could have some issues. This is covered by the THD+N vs frequency and THD+N vs output power sweeps.

The spec table is handy for comparison and for a quick overview. But the real story is told by the graphs.

A spec for 1 kHz at the point where an amplifier is hitting a clipping wall is not indicative of its performance in normal listening. I don't see how that is a controversial statement.

I'm not sure how you deduced that I was claiming there was a controversy. The max output power is a relevant spec. The challenge is how to spec it. I tend to spec max output power at the breakpoint on the THD+N vs output power curve. That works in my case because my amps provide ultra low distortion all the way until clipping. This approach wouldn't work for a single-ended tube amp, for example, because its distortion increases gradually as the amp clips softly. So which point do you choose? Some choose 10% THD, others 1%, a few 0.1%, and then folks like me (and Amir) who use the breakpoint on the THD+N vs output power curve. In doing so I basically end up under-spec'ing the power by a fair amount. In some cases the difference can be 50-100 W between 0.00x% THD+N and 10% THD+N. But all approaches are valid as long as the test conditions are listed.

The character of the distortion is also important.

Not if it's low enough. Overall, I agree, though. That's why I like to see an FFT that shows the distortion profile, preferably at 1 W and also at higher output power.

SINAD at 1 watt or below are very important for most speakers, increasing with speaker sensitivity.

Why SINAD? Why not THD? THD+N? Or IMD? Or multi-tone IMD?

I agree that some measurement that relates to distortion at "normal listening levels" (whatever those are) is relevant. Amir seems to have decided "normal listening levels" is 5 W. Others say 1 W. I've chosen 1 W as well, but that's honestly just as arbitrary as 5 W.

Tom

 

[amirm]

SINAD at 1 watt or below are very important for most speakers, increasing with speaker sensitivity.

SINAD at 1 watt would be dominated by noise so becomes an SNR spec (little to no contribution from distortion). Speakers have infinite SNR so that wouldn't be good in relation to them either.

 

[amirm]

I don't enjoy seeing people getting screwed. Equipment should meet the published specs. But I also don't enjoy seeing a manufacturer's reputation being completely trashed and accusations such as fraud being made because of a single incident, even if it was poorly handled.

Tom, the company left us in the breeze making claims that my measurements were wrong, not anything with the amp. I and a member had to go through expense and effort to test another unit to prove that we were correct. They also attempted to minimize the problem saying just a few units were affected. Yet we also had Audioholics sample so situated. No apology or retraction was ever made and to this date they still have not come clean with what the truth was. We were fed a story until the end (now).

Then there were the claims that this was the first class D amp that sounded like non-class D amp. A lot of people bought their amps based on this statement and stated power. No attempt was made to recall affected amplifiers.

There are a ton of companies out there without these ethical lapses. I see no reason to give them any business. There is "handling" and there is handling. There are few companies whose conduct sinks this low. You have to be very desperate to want to buy their product to do business with them.

 

[amirm]

I agree that some measurement that relates to distortion at "normal listening levels" (whatever those are) is relevant. Amir seems to have decided "normal listening levels" is 5 W. Others say 1 W. I've chosen 1 W as well, but that's honestly just as arbitrary as 5 W.

It is not arbitrary. Speakers have shrunk in size to fit our lifestyles and with it, become very inefficient. So the old standard of 1 watt needed to be upgraded. And that is what I did.

 

[tomchr]

And how did you choose 5 W? From a statistical analysis of a large set of speakers that showed a decrease of 6.99 dB in efficiency compared to speakers of yesteryear? Was there actually a formal standard that required a 1 W test? I'm not aware of one, but that doesn't mean there isn't/wasn't one. I always assumed that the market had converged to include that measurement and went along with it because it made sense.

I'm not making excuses for the manufacturer's bad behaviour.

Tom

 

[nstzya]

And how did you choose 5 W? From a statistical analysis of a large set of speakers that showed a decrease of 6.99 dB in efficiency compared to speakers of yesteryear? Was there actually a formal standard that required a 1 W test? I'm not aware of one, but that doesn't mean there isn't/wasn't one. I always assumed that the market had converged to include that measurement and went along with it because it made sense.

If I understand your argument correctly, Tom, you are saying that 1 watt is/was about as “arbitrary” (perhaps relevant would be a better term?) as 5 watt. And if I understand Amir’s argument (including prior discussion on this topic) he just decided to adopt 5 watt as a more relevant “standard” - for reasons stated. Neither seems to be right or wrong, nor would I think that it would make a load of difference - in a good design. Or unless it was some esoteric 2 watt amplifier.

Interestingly, the Secrets of Home Theater testing of the Fiera4 included both 1 watt and 10 watt testing for similar reasons. It also found rather significant load dependent variations in performance. As well as significant variations in THD and Noise with output - by design. Which the reviewer found pleasing. Obviously he is a proponent of the subjective camp - rather than the more objective leaning crowd here.

I'm not making excuses for the manufacturer's bad behaviour.

For which there are consequences.

And to be clear, it does not appear to be an isolated occurrence. If you read Part 2 of this thread (concerning the second unit tested), there are allegations about the company being international in both ownership and operation and that it has had similar complaints and governmental action in foreign jurisdictions. That’s the foreign intrigue part mentioned earlier…

 

[tomchr]

I'm saying that a measurement at "normal listening levels" is relevant because that's where the amp is likely to be used. At some point someone put a stake in the ground and declared 1 W to be "normal listening level" and measured the performance there. Even if it was never a formal standard, it became a de facto standard.

I highly doubt 1 W was chosen from a statistical analysis of the "normal listening level" of a large and representative set of listeners and their speakers. Hence, I argue that the choice is somewhat arbitrary. Not completely arbitrary, but somewhat arbitrary. Arbitrary or not I'm willing to accept it as a standard (even if it is a de facto standard).

I would further argue that the choice of 5 W is arbitrary for the same reasons. It may not be completely arbitrary, but as far as I can tell the choice of 5 W wasn't driven by science but rather by the perception that modern speakers are less efficient than the speakers that were around when the 1 W "standard" was chosen (citation needed).

As I pointed out earlier, the reason I got sucked into this thread is that someone seemed to have perceived some sort of slight because of a spec table that – in my opinion – seemed pretty normal, albeit somewhat sloppily put together. It piqued my interest because I've seen similar notions elsewhere and it is important to me as a manufacturer to avoid generating negative perceptions of my products.

Maybe I think too much and dare to speak my mind and ask questions. It wouldn't be the first time that's gotten me into trouble.

Tom

 

[tomchr]

And, you're right. There should be consequences for bad behaviour. But people are also innocent until proven otherwise. I don't think it reflects well on ASR to spread claims of fraud at every opportunity. I'd rather stick to the facts (amp failed to deliver specified performance, company acted indifferently when presented with this) and leave the emotions out of it. But maybe that's just me. Like I said... Maybe I think too much and dare to speak my mind. Not always welcome.

Tom

 

[RichB]

I'm saying that a measurement at "normal listening levels" is relevant because that's where the amp is likely to be used. At some point someone put a stake in the ground and declared 1 W to be "normal listening level" and measured the performance there. Even if it was never a formal standard, it became a de facto standard.

I highly doubt 1 W was chosen from a statistical analysis of the "normal listening level" of a large and representative set of listeners and their speakers. Hence, I argue that the choice is somewhat arbitrary. Not completely arbitrary, but somewhat arbitrary. Arbitrary or not I'm willing to accept it as a standard (even if it is a de facto standard).

I would further argue that the choice of 5 W is arbitrary for the same reasons. It may not be completely arbitrary, but as far as I can tell the choice of 5 W wasn't driven by science but rather by the perception that modern speakers are less efficient than the speakers that were around when the 1 W "standard" was chosen (citation needed).

As I pointed out earlier, the reason I got sucked into this thread is that someone seemed to have perceived some sort of slight because of a spec table that – in my opinion – seemed pretty normal, albeit somewhat sloppily put together. It piqued my interest because I've seen similar notions elsewhere and it is important to me as a manufacturer to avoid generating negative perceptions of my products.

Maybe I think too much and dare to speak my mind and ask questions. It wouldn't be the first time that's gotten me into trouble.

Tom

I believe I dared to speak my mind.

One watt is interesting, not that 5 W is not also interesting, in that they represent an attempt to ALSO measure a typical use case. Having measured the SPL at my listening position with 0 DBFS sinewaves and a power meter, I compiled the following table of power usage in my system.

The AHB2 has highly sensitive clipping indicators that corroborate these values.
They clip when expected to clip.

Perhaps, 1 or 5 watts is arbitrary since there are many variables, speakers, room, listening habits.
But given the above measurements that represent maximum power with digital source, l can say with certainty that most often watching a movie or casual listening, one watt peak is going to be about right. Others are welcome to make their own determination. It is not particularly difficult.

I think you will find examples of ASR amplifier measurements that performed much better at 1 kHz than when in the test where multiple frequency power sweeps were measured.

- Rich

 

[tomchr]

I believe I dared to speak my mind.

How dare you?!

One watt is interesting, not that 5 W is not also interesting, in that they represent an attempt to ALSO measure a typical use case. Having measured the SPL at my listening position with 0 DBFS sinewaves and a power meter, I compiled the following table of power usage in my system.

Yay! Data! Love your methodical approach.

Tom

 

[nstzya]

And, you're right. There should be consequences for bad behaviour. But people are also innocent until proven otherwise. I don't think it reflects well on ASR to spread claims of fraud at every opportunity. I'd rather stick to the facts (amp failed to deliver specified performance, company acted indifferently when presented with this) and leave the emotions out of it. But maybe that's just me. Like I said... Maybe I think too much and dare to speak my mind. Not always welcome.

Tom

Always welcome - particularly from experienced informed experts in the field. This seems to be veering far off topic however. I am in full agreement with “innocent until proven otherwise”. I respectfully suggest that you read both parts 1 and 2 of this thread in their entirety and see if you don’t agree that it was proven otherwise. Suffice it to say that numerous attempts were made to encourage the company to communicate and/or provide explanation - all rebuffed. Tends to make people emotional and speak THEIR minds.

 

[tomchr]

Right. But were the claims of fraud proven in a court of law or just in the court of ASR?

Tom

 

[voodooless]

Right. But were the claims of fraud proven in a court of law or just in the court of ASR?

I think there was proof enough due to them pulling the product.

 

[RichB]

I think there was proof enough due to them pulling the product.

There is no compelling choice between fraud and incompetence.

No Matter. Starke should provide a battery of measurements to regain credibility, IMO of course.

- Rich