Lunch with Paul & Darren - The importance of feedback in audio

Cool interview.

Here is Bruno Putzeys paper on feedback: “why there is no such thing as too much feedback”

@Paul and Daren - have you heard Bruno’s Kii3 DSP speakers? They don’t communicate music well in your opinion?

Asking respectfully of course.

Love the video series.

1 Like

Without getting into answering for them, I would point out a couple of things.

The Kii3’s merits (or lack thereof) are not due solely to: A) the amplification or B) whether the amplification has feedback or not.

Paul and Darren were talking about designing a microphone preamp rather than a power amp. EDIT: Sorry - was thinking about another video of theirs!

1 Like

Thanks for your question.

Badbeef is right, my comments are concerning a linear line level amplifier rather than a power amplifier. Bruno is mainly talking about nonlinear power amplifiers. The segment was cut up, so that wasn’t very clear in the video.

You happen to skip the point that I made saying that feedback essentially makes an amplifier “a better amplifier”. My first question to you would be what part of that disagrees with Bruno?

It’s also worth noting that he does not provide any proof or claims that high amounts of feedback results in better subjective audio quality. Bruno has a point: more feedback equals lower distortion and he is not wrong.

I agree from a technical perspective that as feedback is increased - given gain/phase margins are adequate- it will lead to better bench performance.

Let me reword your question:

Why is a high distortion, low bandwidth amplifier that requires a lot of correction not as desirable as one that doesn’t need any globalized correction at all?


This thread has a few directions that are going on.

I’ve heard the Kii3 DSP’s a few times now. It was at the Toronto Audio Show both times and the difficult acoustics in a hotel room were easily overcome by the DSP. They noticeably sounded very good with the bass stack. However, at the prices they were asking, it was a non starter for me. Their target market isn’t the audiophile hobbyist with separates. It’s the urban, cash flush condo dweller.

1 Like

Yeah, I haven’t had the chance to hear a good demo of the Kiis, so I cannot comment.

What I can say is that I highly respect Bruno and his expertise in Class D.

Agreed. I only heard them at a show and could hardly stay in the room for more than 2 minutes because the person giving the demo insisted they were played at over 100dB. I’ll tell ya one thing, they sure can play loud!

Hi Darren and thanks for the reply.

You are correct that Bruno doesn’t talk about better subjective audio quality but that was actually my only question to you in my opening post , i.e.

"have you heard Bruno’s Kii3 DSP speakers? They don’t communicate music well in your opinion?"

So the only question I had in my opening post wasn’t technical/objective, it was only subjective…

In a latter reply, I’ve now seen you haven’t heard Bruno’s Kii3’s. So that answers that.

You should give it a listen! It might surprise you and communicate music really well !

Their target market is both the pro audio world and serious home audio. A number of very excellent mastering engineers, etc. use them. They are also selling home audio as a number of other pro manufacturers have done. Bruno and his colleagues have a lot of pro audio experience.

They do solve a lot of problems, and if you want an all-in-one (or mostly all) solution you can place near a wall, they’re great. A local studio owner I know used the D&D 8c’s for a few months as his studio monitors and enjoyed them - though he didn’t stay with them.

So the thread title was clickbait? It’s not about “the importance of feedback in audio”, or spelling “Darren” correctly?

That’s the title of Paul’s video interview with Darren… :man_facepalming:

Please watch the video referenced in the opening post before commenting…

Your post adds zero value to the discussion.

1 Like

Yes - aware that’s the title. However I guess I was asking, from your follow up post, if it should be entitled “Darren - have you heard Bruno’s Kii3’s?”. When you see a thread title in the topic list, it sorta suggests the thread might in fact be about that. Especially when the first post features a link to a paper that discusses feedback and so on. I’m sure Elk can tell me how often that doesn’t follow.

I did watch the video as soon as I realized it was referring to a different segment of the same Lunch that “Designing the world’s most musical microphone preamp” came from, which is why I made an edit to my earlier post. Apologies.

Back on topic, I see that there is a lot of discussion regarding “too much feedback” regarding power amplifiers. Like everything in life, there are those who stands for that, while the amplifier remains stable, one should use all the feedback that is possible. This would drive distortion to below measurement.

But I see a lot of companies, and IIRC, even Paul defending that the feedback medicine might become poison if used in excess, killing some good sonic properties (maybe dynamic range? but here I might be getting it all wrong). In fact, if I got it right, BHK300 does not use negative feedback, at least on its input stage.

So for me the alternative appears to be:

  • begin with a very good design, that brings little distortion per se, avoid feedback and enjoy other benefits, and one should have to ponder the cost-benefit of such feedbackless design
  • begin with a simpler design, much cheaper, and correct it with as much feedback as you can use
  • any point in between, considering cost-benefits.

@DarrenMyers did I get this right? If so, in your view, what would be the dangers that too much feedback could pose when used on a power amplifier?

Hi jvvita!

The subject of feedback in amplifiers is a deep topic with many details and specifics that consider the type of amplifier, the price point, and the desired subjective outcome.

So first, it’s important that we specify what kind of amplifier we’re talking about. Is this a Class D power amplifier? Is it a Class A power amplifier? Or is this a line-level preamplifier circuit?

Linear power amplifier output stages contribute high amounts of distortion and in the case of the Class AB output stage, highly objectable high-ordered products. Due to the highly nonlinear behavior of Class AB and D output stages, feedback is a must in order to achieve good sound. It’s HOW this feedback is applied that is so critical. The technique and attention to detail, specifically what is going on in the open-loop circuit, is what determines the sonic outcome. Focusing on just maximizing the feedback can certainly be done - and you can bet dollars to donuts that THD will be minimized.

The difficult task of maximizing this loop gain while still providing the amplifier with ample phase margin (>90 degrees) has, in my experience, compromised some of the open-loop characteristics I’ve associated with top notch sonic performance. Can good sound result from high feedback designs? Absolutely. Is it the best we can do? I don’t believe so. Given certain designs such as Class D amplifiers, it most certainly offers the best solution to mitigate the innate compromises of the design.

Now for line level preamplifier circuits - When implemented correctly, integrated op amps offer amazing performance for their size, cost, and efficiency. These op amps use high amounts of feedback and VERY aggressive dominate pole compensation in order to make them adaptable for many environments, designs, and engineering development expertise. This design technique also optimizes traditional audio specifications such as THD, IM, FR, etc. What’s interesting is when you look at the designs of video op amps and how their design method changes in order to adapt to different challenges and spec optimization. There are a lot of hints there and if you’re interested in learning more about this I would recommend looking at how video op amps differ from “audio” op amps - both in performance and topology/design.

Now whats even more interesting is that with line level circuits we can actually make open loop designs that perform extremely well and even have a few superior objective advantages over audio ICs. They may be harder to produce, more expensive to produce, harder to design, run a lot hotter - but in high end audio it must always be about achieving higher subjective performance than the mass produced mainstream stuff regardless of the difficulty or expense. As audiophiles, we are willing to go above and beyond to get to the next level and I think audiophiles deserve to have gear created by designers that respect that philosophy as well.


I’ll make sure to get a good demo of them and as I said, I have a lot of respect for Bruno. I’m sure they’re a great product based on his history and what people say about them. However, I’m not sure how this specifically relates to what I said in the video.

1 Like

Thanks for taking the time. You bring a breath of fresh air to the discussion, keeping away of preconceived notions of right or wrong and keeping an open mind.


For all separates with monoblocks and dedicated power supplies etc. etc. you need:

A large room, if that is not the room where you can install such system you have following options:

  1. Limit the amount of sources, like go streaming only, play digital discs or just vinyl.
  2. Cram all those separates into the limited space and live with the mess as long as the SQ is top.
  3. Put any auxiliary equipment like power conditioners, power regenerators, network switches but also power amps in an adjacent room and engineer a sound cable transfer from one to the other room.
  4. Buy a high end integrated system.

Taking a simplistic look at the subject, integrated components might not do an equal job for cross contamination of unwanted signals and noise as individual separate components.

But on a system level the comparison of SQ might very well be a draw:

  • Separates are not designed to optimally fit to particular loads (amp input stages / speaker drivers), integral systems are. This results in much better interfaces between parts of the system.
  • External cables, no matter how good they are, are antennas for noise.
  • Any external cable introduces extra contacts which complicates the interface impedance, phase distortion and noise are likely to be introduced.
  • Active low power cross-overs cause less distortion than passive cross overs by nature, but need good and clean power.

At the end it all comes down on personal preference. Some people can afford and like experimenting with different separates and have the space. Others get a kick out of a visually sleek impression and how all that stuff is crammed into less boxes yet Mai taming a good level of signal separation.

Bruno started at Philips consumer electronics by the way, Philips initially needed the class D amplifiers to produce cheaper and more compact systems without sacrificing SQ. They wanted to become the more affordable Bang & Olufsen. After a short period their marketing department and financial officers overdid it with their desire to save money and the design of hard plastic shiny boxes. It dramatically limited the R&D budget and finally the targeted SQ. They were not able anymore to distinguish the Philips brand from the new cheap cheap Chinese stuff. I am afraid that Philips is not the only hifi brand that became victim of their own marketing and financial policy. Note that it wasn’t all bad, what they made and still make.

Bruno perfected class D and DSP control with science and algorithms but sure knows how to engineer analog low power stages. Let the Mola Mola Makua and Grimm and Kii input stages be witness of that. Obviously that comes at a cost which only professionals or well heeled enthousiasts are willing to pay, that is how Bruno and many other smart ex Philips engineers became familiar with the pro and high end world.

Designing and manufacturing such integrated devices is equally an art as designing an manufacturing beautifully looking and sounding separates.

Hi Daren

In the video you say that lots of feedback can lead to ultra low distortion measurements but this doesn’t result in communicating music well in your experience.

Bruno obviously uses a lot of feedback, striving for ultra low distortions… Hence my query if you’ve heard Bruno’s Kii3’s and if you thought they communicate music well…

I’ve since seen you haven’t heard them though.

I think you’re taking my statement a bit further than it was intended. Stating that I like Kii speakers would not be enough to conclude that ultra high feedback amplifiers are better than other methods. You’re also dealing with a lot of variables with speakers, way more than just the amplifiers. There is also no way of changing the amplifiers out to see what better electronics bring to the table.

I’ve stated that with Class D you have to have that feedback in order for it to perform well. I’ve also said that I like Brunos work. These statements do not mean that this technology cant be surpassed using other methods. These methods would likely produce more heat and take up more space. In other words NOT a smart choice for a compact active monitor. :wink:

1 Like

Fair points Daren. All noted.

Have you had a chance to hear Bruno’s PuriFi amp, the evaluation kit?