Matching specs

March 1, 2022
 by Paul McGowan

Imagine for a moment two analog gain blocks, each with identical input and output impedance as well as gain.

To make our little thought experiment more interesting, let's also imagine their THD and IM measurements are identical as well.

Now, here's where it gets juicy. Let's imagine circuit A is a traditional high open loop bandwidth amplifier with lots of negative feedback applied to establish its gain. Circuit B is the opposite. Here we have a circuit with very low open loop gain with very little negative feedback.

Would the two sound different if we played music through them and listened on a highly resolving system?

Having made this comparison more than a few times I can tell you my own findings are pretty clear. Indeed they do sound quite different, especially when the type of music we're using has lots of rich harmonics and overtones played in a non-cluttered setting where those overtones can easily be heard.

It's in fact not even a contest. Instantly noticeable and consistently the same even in blind testing.

But, why? What measurements might we apply to see those differences?

A differential null test?

Why haven't we, as an industry, together created a measurement system that clearly demonstrates those differences on a measurable basis?

And even if we did, would anyone care?

Subscribe to Paul's Posts

47 comments on “Matching specs”

  1. If you can hear the difference between negative feedback & no negative feedback with your ears, in this example, then that is good enough because even if there was an industry measurement system that clearly demonstrates those differences, home-audio still comes
    down to how something sounds to the individual listener.
    You can measure & compare home-audio gear until you're 'blue in the face', but ultimately
    it will come down to how the prospective customer hears it & what their preference is.

    1. Having both said exactly the same thing, I do wonder how companies selling direct know what audio buyers want and how they make decisions. Audio dealers like yourself (in a former life) are the focus group leaders and many manufacturers do dealer demo's as much to get feedback as showcase their products.

      1. I'm guessing that there are enough 'professional' pairs of ears at 'PS Audio' to get the 'sound' of a pair of loudspeakers (FR30) or a new amplifier (BHK-650) pretty-much where said 'sound' will be at least acceptable & at most lusted after by the majority of prospective buyers.
        The real 'proof of the pudding' would be to know what the percentage of 30 day trial period returns there are, compared to the finalised sales...on a yearly basis.

        1. I just arranged to send something back from a trial. The dealer is 300 miles away, which is an issue. I spoke to my regular dealer at 11am. He is bringing a different trial unit at 3pm. That's 4 hours, whereas I seem to recall it's taken you 4 months not to get a SACD.

  2. This consumer doesn't buy on measurements, let alone combinations of measurements. I plug things in and give them a try. I think Paul answered the question himself. If the difference is clearly audible, then why the need for measurements? Plus the electronics descriptions are meaningless to me. If my audio dealer said: "Steven, here's a nice high open loop bandwidth amplifier with lots of negative feedback", I think I'd find another audio dealer.

    I'm doing that at the moment with a demo amplification unit. Besides having a couple of issues, clearly audible, I was promised delivery from stock and am now told the thing won't be fabricated for a month, then international shipping. Shall be looking elsewhere.

    1. I had a similar experience. Paid and expected delivery in a few days only to be told no stock available and it would be 10-12 weeks. Then another month. I have it now but not the best experience. Not a PS Audio product btw. The thing is, depending on what you’re looking for, elsewhere can be a diminishing place.

    2. Doing *what*, at the moment with a demo amplification unit, Steven?
      I thought that you are 'head over heels' with your Devialet pizza-box & your CXA-81
      ...trouble in paradise?

      1. The CXA-81 system is sold, replaced by a ceiling system. (No boxes, no wires, yippee.) I'm playing with phono MC pre-amplifiers. There is one in the pizza-box, but plenty of people use external for MC, still use pizza internal for MM.

        The one arriving is a RCM Sensor mk2 from Poland. My dealer has their $50,000 phono amp at home. I'm aiming a bit lower, but they've built quite a strong reputation for phono amps over the last 15 years. RCM are the largest hifi distributor in Poland.

    3. I do it based on listening as well. The problem is the “want” of a binary way of saying some thing is better or worse. Unfortunately, hearing is itself a variable, the differences in speakers and varying components create a synergy variability and finally people just prefer different sound. I listen with several friends, and personal preference is evident.

  3. A very strange question and a comparison of apples and oranges. The measurements were made on a test bench with fixed parameters fixed load resistance). The “measurement” by listening was made in a more or less acoustically treated listening room using a loudspeaker load with resistance/Impedance varying with frequency and using a measuring device (ear/brain system) based on sophisticated data reduction system optimized for pattern recognition and depending on moods varying with daytime and musical genre. Why not perform realistic measurement (at the sweet spot in the listening room with real loudspeakers) or even better measurements based on known aural patterns? This post clearly reveals the lack of knowledge concerning basics of psycho-acoustics.

    1. I would beg to differ, Paul. In this experiment, we imagine the exact same listening conditions and all that changed was the circuit processing the music.

      Now, imagine what you are proposing (I think). Place a microphone in the listening position and record the results, then compare those results between the two scenarios.

      There would be no doubt the two would be different and one could "see" the difference on a scope.

      Then what? Compare the two to the original source to see which is closer?

      1. From a measurement point of view your last statement to Paul makes total sense. The problem occurs when the one that measures closest may not sound the best?

      2. Paul, I always wondered why the microphone measurement option hasn’t been used to solve the interconnect debate. There is clearly audible differences. I would think those differences would also show up. Has anyone ever done that comparison?

      3. My understanding of stereo is quite simple, Paul: a loudspeaker-amp combo labeled “high fidelity” should reproduce correctly an instrument’s or voice’s overtones correctly (correct amplitude and timing) without any audible distortions which will be seen in a perfect step response and flat frequency curve. And now the complexity of the interaction between loudspeaker and room enters the game especially for far-field listening (diffuse sound is dominating). Depending on the loudspeakers directionality or design (horn, omni, dipole, bipolar, planar, open baffle, point source, line source, etc) these interactions can barely be simulated for an individual listening room. And not to mention individual listening tastes and preferences as well as individual hearing deficiencies. My impression is that you like to play with this complexity posting more or less provocative posts and then you start grinning reading the comments. 🙂

    2. As you know there are a lot of differences in performance of O-scopes and spectrum analyzers. "Slow" equipment can miss resonances and or quirks that can appear on the waveform which can be plainly seen on high-end lab grade equipment. Cheap probes can inject more noise than what you are trying to hunt down. All I am saying is just because a person does not see it does not mean it is not there.

  4. In an (audio) world where we’re used to even gear with inferior specs sounding clearly better than the one with superior specs on paper, we’re not too surprised of gear with identical specs sounding different, if the concept to achieve those specs is different.

    The variation in the technical/content depth of posts between most general topics and those of amplifier development is always interesting to observe 😉

  5. If any case while under the large umbrella of this hobby, if I can hear a difference I’d absolutely love to know why I hear a difference and all the technical jargon that goes with it.

    Maybe that is the lawyer in me coming out. 🙂

  6. So it's preferences. What is Paul's flavour?
    Both these are valid yet appeal to different informed audiophiles -- Macintosh or Audio Research?

  7. Alert all,

    This morning I received a spam email from “System at PS Audio”, advising me that I had been picked as one of two users of the month, and to click here to have a special badge placed by my name”.

    I have been receiving a lot of spam email recently from fake sources like Costso, Home Depo, and other obvious misspelling attempts. But a spam email so specific as from PS Audio indicates a breach of someone’s server or a data breach.

    I now return you to your normal programming……

    1. It's definitely not coming from our servers. It's because we communicate a lot over the internet. More than most companies (like what we're doing right now). Our employees are routinely spammed from "me" instructing them to go to Target and purchase $500 gift cards. Instructions of where to send them will follow.

      I wish these low lives would leave us alone. Fortunately, it doesn't take much to figure out it's not from us by looking at the email address of the sender. If in doubt (and always be in doubt) drop us a note. There is no such email address in our domain that is [email protected]

      1. As a public service announcement, and I hope everyone sees this... Never open spam. Don't even click on the message. The hoodlums are working very hard on perfecting no-click infection. Just clicking on the message could install malware on your system. Make sure your mail tool classifies spam and puts it into the junk mail directory. If it doesn't, get a good classifier. SpamSieve works great on my Mac, and I have it put the spam into the same directory the mail tool uses. Inspect that directory to see if there is good mail in it, then empty the directory without clicking on any message.

        Further, make sure your mail and message tools are configured to never send receipts for either receiving or reading a message. Sending receipts confirms your existence as well as your address, without your knowledge. Receipts are evil. And make sure you change these settings on your phone, too.

        One option to consider is configuring your mail tool to use multi-line message preview. That way, you don't need to open the message to see the spelling errors or tipoffs such as "we have a surprise for you". Yea, I'll bet you do.

        One more thing. I keep my Mac up 24 * 7. That allows the classifier to run constantly, which protects my iPhone as well. The classifier moves the junk mail in the background and I don't need to run a classifier on the iPhone.

        OK, now going back to lurking mode.

          1. Nope. I'm for real, I think. I had a long-term girlfriend that was so fluent in Russian the CIA tried to recruit her. She made a career of translating Russian metallurgical papers into English for American Metals Climax (AMAX). Me? I think life is too serious to take seriously. When I pass, I want to be buried with a pair of Focal Utopias, my Walkman and fill all available space in the box with batteries. Learn more starting with issue 150 of Copper. (shameless plug).

            1. Good for you. Are you talking about their headphones? Because those floorstanders are huge. And while I don't imbibe, I laughed a couple of summers ago when Willie sang to us, roll me over and smoke me when I die.:)

              1. I just looked in issue 150 of Copper Jack. And I was one of the first ten folks to feel your writing for the magazine was worthy of comment. [Now that's a shameless plug!]
                Worth every penny. 🙂

  8. Taking this statement and ignoring the last one…
    “Why haven’t we, as an industry, together created a measurement system that clearly demonstrates those differences on a measurable basis?”

    It seems to me that the industry doesn’t see taking profits and throwing them into science to answer basic questions as a wise use of their profits. Having a ‘hidden recipe’ or technique is what separates one audio company from another. Why would you give what you know and have learned to a competitor….. especially if the risk is lower or no profit….

    Each manufacturer has their own way of doing things to compete for the small percentage of the niche called audiophiles. If the market grows more manufacturers will jump in… if the market shrinks attrition of manufacturers will abound….

    There will always be some stalwarts that will remain - mostly supported by those who have tremendous amounts of disposable income. In other words… accessible by the few… drooled over by the majority. Every manufacturer should ask what they want their customer base to be.

    Maybe some manufacturer will figure out a way to make everything audio correlate to measurements. When they do they’ll have nothing but cold hard facts to promote the business.

    1. Secret(?) sauce. Widely used but shhh not talked about. Just add a dash of goog old fashioned second harmonic. Ahh! That’s better.
      You know it’s true.

  9. Re: "What measurements might we apply to see those differences? A differential null test? Why haven’t we, as an industry, together created a measurement system that clearly demonstrates those differences on a measurable basis?"

    Not to be too cynical, but my suspicion is that one reason the high-end industry doesn't do this is the same reason that the alt-med industry doesn't do scientific testing and clinical trials: the results would likely be disappointing and so would not help sell products.

    And you ask if anyone would care. I would care, but I think a lot of audiophiles enjoy the hobby because they are buying hopes & dreams even if they think they are buying better sound. Double-blind listening, ABX switches and null testing take the fun (and the placebo effect) out of it. 😉

  10. Put a thermocouple on the high gain device w/lots of "destructive" feedback and on the low gain - low feedback device.
    I am curious to what they measure.
    Just the warmth of your finger can change the gain of a transistor. All silicon has a thermal voltage.

  11. One of the biggest problems you face is with regard to phase response. From a scientific perspective the combination of measuring phase response and correlating that with audibility is challenging to say the least. On the one hand, strictly scientific tests using synthetically generated test tones generally deliver the result that phase distortion, if it is audible at all, is barely audible. On the other hand, the high end world understands that paying appropriate attention to phase response can result in audible benefits. But, oddly enough, it is rarely viewed in such terms. For example, amplifier designs with "unnecessarily enormous" bandwidths inherently deliver a more linear phase response across the audio band than those with a "sensibly limited" bandwidth.

    As far as null tests go, you have to be very careful. The idea is that the output of a null test represents the presence of a signal in one channel that is absent in the other. A non-linear phase response can (in principle) change the waveform itself without there being any corresponding change in either the frequency response or the THD, IM and noise. Interpreting such results in a null test needs to be done with great finesse.

  12. I am having a hard time understanding the purpose of today's post.

    If you took a 100 people that have serious stereo systems and asked them what is most important in their system I think you would get a wide range of answers ( perhaps a 100 different answers ).

    If you took a 100 different audio amplifier designers and asked them what is most important in amplifier design, again, I think you would get a wide range of answers.

    Finally, if you took those 100 amplifier designers and said that as a group they must come up with a set of 10 or 12 measurements that all amplifiers manufactures must do a publish I doubt that you would get agreement on those measurements.

    1. IMHO It seems to me the reason is know a little bit what is under the hood. From a schematic standpoint there are not 100 different family categories of discrete amplifier topology, to my knowledge anyway. If you sketch a schematic on a napkin its pretty basic stuff... some amplifier designs are better for audio than others.

  13. I'm (along with Nephilim 81) probably an outlier here. I think Paul's question here about negative feedback is very interesting. (In fact, I asked him once long ago on a private thread if there were any measurements that would indicate the deleterious effects of negative feedback; he said he didn't know of any.) With the death of retail, I, like you, cannot actually listen to many different components. So specs are an important indicator. If an amp has crappy IM specs, the odds are I'm not going to like it. And when a reviewer says the amp under review revealed a flaw in their AP test system that no other amp uncovered, I get excited. Apparently, there's pretty broad agreement that heavy use of global negative feedback sucks. Ok, here's a proof-point if you will that specs are insufficient for determining audio quality. (Sorry, Julian Hirsch.) But the scientist in me says there must be something that can be measured there! I'd like to know both out of pure curiosity and to get even better measurements for making decisions.

  14. Where’s Bob Carver when we need him? He can make his amp sound like any amp out there! What a character back in the day!

    He could do it with one set of speakers, but it gets more complicated to do it in production units that are used with many different types of speakers and speaker cables.

  15. Yes Invalid. Bob was a creative Genius when it came to audio. Sonic Holography and other trick circuits, those huge 700 watt amps and all his patents were amazing. How about those Fantastic amps the Silver Sevens that we can't afford but how does one create an amplifier to mimic the sound of another at a much cheaper price? That is simply mind boggling. He proved the experts wrong with that challenge many years ago. I have worked on some of his older amps and by todays standards with lack of protection they can often become Flame Linear amps. And those Cube amps are amazing but very hard to service. I owned one many moons ago and the sheer power for such a small amp was amazing. I did turn it up a bit too loud and burned it up but nevertheless he is an amazing engineer and great contributor to electronics. I do remember Paul telling an interesting story on one of these videos about Bob's wig fell off a long time back and how funny it was at the time.

  16. Those carver amps sound nothing like the mark Levinson ML2, Bob even admitted he couldn't make them sound the same as the ML2 in the production units.

Leave a Reply

© 2022 PS Audio, Inc.

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram