He seems to think that DSD and PWM outputs are unipolar - i.e. soft sounds have few or narrow pulses. The reality is different - soft sounds have close to 50% duty cycle.
I guess that I've been lucky, they Class-D amps I've used have not been fatiguing, far from it. I agree with the video that ICE is very smooth and not fatiguing.
I have only listened to one ICEpower amp, it wasn't fatiguing but over time I noticed it being a little bit over bright at the top end. My TDA8932 classD amps don't have that extra brightness.
I'm not clear what Paul is attributing as the cause of fatigue here - he begins by saying its down to the output filter. Clearly it cannot be that as all classD amps have an output filter and we all agree that not all classD are fatiguing. My suspicion for what causes fatigue is the fast edges of the output stage getting back (maybe radiated?) into the input stage. Not all input stages are susceptible to that - as Paul hints (but doesn't say explicitly) valve input stages tend not to be.
I'm not clear what Paul is attributing as the cause of fatigue here - he begins by saying its down to the output filter. Clearly it cannot be that as all classD amps have an output filter and we all agree that not all classD are fatiguing. My suspicion for what causes fatigue is the fast edges of the output stage getting back (maybe radiated?) into the input stage. Not all input stages are susceptible to that - as Paul hints (but doesn't say explicitly) valve input stages tend not to be.
Funny, i found the ICE to be very neutral at the top, and friend of mine said it was "too soft" at the top and you found it a bit over bright. 
I saw a lot of Tripath amps that had poor PCB layout, resulting in that very coupling of noise back to the inputs that you mention. It's easy to get wrong, and can often sound "sour." Perhaps a lot of other Class-D amps suffer from poor layout.
I did notice that Paul says that the filters are aggressive, which isn't what I've seen, most are just 2nd or 3rd order, IIRC.
I saw a lot of Tripath amps that had poor PCB layout, resulting in that very coupling of noise back to the inputs that you mention. It's easy to get wrong, and can often sound "sour." Perhaps a lot of other Class-D amps suffer from poor layout.
I did notice that Paul says that the filters are aggressive, which isn't what I've seen, most are just 2nd or 3rd order, IIRC.
I think there are a number of incarnations of 'ICE' - it goes back many years. I heard an early generation one at B&O's premises when I was working in the industry and didn't much like how that sounded.
I'm with you on disagreeing with Paul on 2nd order filters being 'aggressive'. That's just another classD myth he's propagating methinks. I've only seen 2nd order in the flesh but did meet a schematic a few days ago with what looked to be 4th order.
I'm with you on disagreeing with Paul on 2nd order filters being 'aggressive'. That's just another classD myth he's propagating methinks. I've only seen 2nd order in the flesh but did meet a schematic a few days ago with what looked to be 4th order.
I think what Paul is saying is that all Class D amps have an output filter, but not all output filters are the same. Depending on how they are designed some are fatiguing, but others are not.
I didn't get the impression Paul was saying that. I don't recall any comments he made in respect of how an output filter is designed but I might have missed them.
That seems to be exactly what he is saying starting at about 1:30 into the clip. And then again at about 3:45.
Last edited:
As I am busy with a MA12070 design and test this practical design it becomes clear that with this particular chip the output filtering is largely responsible for audible improvement. More than I previously thought to be honest. What also helps is to keep input wiring far from output wiring (as usual, no surprise there). A nice way of testing is to place metal shielding at various places. I had to make metal shielding for the volume potentiometer to have it optimally performing.
So I treat it as a RF transmitter than also can playback audio 😀
So I treat it as a RF transmitter than also can playback audio 😀
Last edited:
I need a lot of salt around me when listening to this guy.So much marketing...but youtube algorithms like marketing!From time to time he admits other's audio equipment is better than his , but only when its price is higher than PS Audio equipement which is such a rare event...
My first encounter with Bruno's amps was in 2014 when I listened to some 10k bucks near field(1m) MM27 Barefoot speakers.They sounded well but they were eating 800 Watts to sound from 1 meter like a horn eating 5 watts.Then in the same room we listened to some 1956...60 Tannoy red 15 monitors driven by @greierasul's Sara 2014(the first version) with about 10 ...20watts .The cabinets were open to wire the speakers who's original amps were defective and the base on Tannoy was somewhat less powerful below 50hz, but other than that the 60 years old coaxial measured identical to the Barefoot , a brick wall from 50 hz to 20 khz with the Barefoot having more linear base below 50 Hz.
I can understand the Hype though 🙂
My first encounter with Bruno's amps was in 2014 when I listened to some 10k bucks near field(1m) MM27 Barefoot speakers.They sounded well but they were eating 800 Watts to sound from 1 meter like a horn eating 5 watts.Then in the same room we listened to some 1956...60 Tannoy red 15 monitors driven by @greierasul's Sara 2014(the first version) with about 10 ...20watts .The cabinets were open to wire the speakers who's original amps were defective and the base on Tannoy was somewhat less powerful below 50hz, but other than that the 60 years old coaxial measured identical to the Barefoot , a brick wall from 50 hz to 20 khz with the Barefoot having more linear base below 50 Hz.
I can understand the Hype though 🙂
That hype will be everyones reality within time. Class D is only now becoming mature and the seizing of production of discrete transistors will be the nail in the coffin of discrete DIY designs.
If there is a time to buy the semis of that famous class A/AB design then it is now.
If there is a time to buy the semis of that famous class A/AB design then it is now.
Unfortunately some of the class d amps, while very cheap to produce are sold as expensive as class A gryphoon and mark levinson which I think is a very mischievous hype.
They can be. I have a small AM/FM portable radio; when placed close to the speaker cables of a class D amp, it plays the music too. Might give that one a try sometime. Especially with these designs having the 603 sized SMD inductors...
Try to design a class D design the old fashioned/optimal way and then you'll see that it will also be costly. I think it takes a lot of effort but they can be made very good. Too bad interest in designing optimized complete "bang for the buck" designs is near to nil. Low cost and features dominate.
One looses interest in ready made stuff quite quickly when listening to most devices on todays market. I visited an audio shop last weekend that had no HiFi separates anymore. Quite some cheaply built boards can outperform ready made devices so this seems to create a new DIY branch, the PIT DIY amplifiers. Put It Together DIY.
My view is to have a good looking minimal DIY class D amplifier with all analog optimal choices for a reasonable price and no digital Spielerei in plain old stereo. Like when it all started with HiFi, today that is a niche in a niche 🙂
One looses interest in ready made stuff quite quickly when listening to most devices on todays market. I visited an audio shop last weekend that had no HiFi separates anymore. Quite some cheaply built boards can outperform ready made devices so this seems to create a new DIY branch, the PIT DIY amplifiers. Put It Together DIY.
My view is to have a good looking minimal DIY class D amplifier with all analog optimal choices for a reasonable price and no digital Spielerei in plain old stereo. Like when it all started with HiFi, today that is a niche in a niche 🙂
Last edited:
You can't take anything that Paul at PS Audio says as meaningful he talks so much nonsense like it's scientific fact.
I listened again - around 1.30 he claims that the fatiguing quality is due to the output filter. But that's just a claim which he doesn't back up with anything at all so I don't give any weight to that claim, it doesn't make any sense. Then at 3:45 he again claims that the filter is 'primarily the culprit behind it' and then he wanders off on a tangent into talking about Bruno. So is the 'it' in that context the fatigue? If it is its just another claim without any substantiation or explanation so I'd not give it credence myself.
So I guess you accept those claims at face value and then interpolate that he must be meaning there's some quality of the output filter involved? Whereas I dismiss them as just claims without supporting reasoning/rationale.