The icing on the cake

January 8, 2022
 by Paul McGowan

Look at the frequency response curve of any loudspeaker and you’ll immediately see it is not flat. Not even close. Deviations in loudness at specific frequencies are denoted in terms like +/- 3dB (on a good day). Some active speakers boast +/- 1dB, but even that’s hardly flat.

And those measurements have little to do with the actual in-room response at one’s listening position.

When a speaker designer is faced with the reality of inevitable loudness swings of this magnitude, they can either shrug their shoulders and say that’s as good as it gets, or they can use those deviations to their advantage.

When they decide to go with the latter decision, the process they use is called voicing.

If +/- 3dB is the accepted limitation, designers who understand the art of voicing decide where those deviations are best put to use. Instead of a dip at one frequency, the overall sonic presentation might benefit from a bump in loudness instead.

The deviation range remains the same. The sound does not.

Not all designers choose the art of voicing.

For those that do, it’s the icing on the cake.

Subscribe to Paul's Posts

63 comments on “The icing on the cake”

  1. It’s more than just the icing. Alan Shaw has spent in excess of 1,000 hours voicing a single loudspeaker. It seems to be what he does most of the time, as he’s spent the last 37 years basically finessing a small range of very similar loudspeakers. Nothing is left to chance, with every iteration carefully measured and recorded in books now running to many thousands of pages. He doesn’t get diverted by other products or projects and he does it all on his own. This level of methodical dedication is likely a large part of why his loudspeakers are so popular around the world.

    1. If I wasn’t mainly into Rock ‘n Roll I would’ve bought a pair of ‘Super HL5 Plus XD’s instead of the DeVore Fidelity – ‘O/93’s that I decided on last month.

      Anything Jonny can do, Usman can do better…
      Rain tomorrow…could be a draw 🙁

      1. Paul,

        This is a preposterous statement. I am sure that even Chris would disagree with you.

        Maybe you should get and read Toole’s third edition before criticizing it so misleadingly.

          1. Paul,
            You can change the frequency response of your speaker (in the crossover), you can change the crossover frequency, which could change the response and the dispersion characteristics. But you only know how well this happens by measuring.

            By the way, there is a good description of this process in Erin Hardi’s review of the Sointuva speaker from March Audio. What is interesting of this speaker is the low distortion in the bass, as proposed by the designer of the woofer. This could be a baseline for you to check the FR30, as Chris’ goal was a low distortion in the bass.

            But once you tilt your speaker to your preference, you are walking away from “neutrality” or from what Toole’s research has demonstrated. Even Chris’ speaker starts from “Toole’s ideal”. B&W tilts their speakers FR, while many like D&D, Neumann, Kii, Genelec and others are clearly focused on response based on Toole’s research findings. The BBC sound was a tilted response too.

            I am still surprised that you have not purchased a Klippel device given how committed you state you are to building speakers.

            1. Thanks. A Klippel is a bit more than $100K and doesn’t offer us a great deal more than we have managed make for ourselves (the jigs etc.) to measure. Chris would LOVE one but we cannot now afford it. We’re already a million into speakers and have to be careful until we sell a few and capture back some of that sunk capital. And, as I mentioned, our vendor has a Klippel system and every unit they built gets the full test along with all the drivers Chris designs.

              1. Get in touch with Erin. He would love to test it WITH you. I’m sure he is willing to do this privately.

                What was it that Reagan stole from the Russians? “Trust but verify!”

                The value of systems like Klippel is the test-retest repeatability. You know that the results are reliable. When you generate your own jig, you can’t really compare your results against others, unless you tested other speakers in your jig too. And that is expensive as well if you have to buy Magicos or Kefs.

  2. The other very important ‘icing’ that Harbeth, & I hope that PS Audio, does is to
    ‘pair match’ their loudspeakers.
    Harbeth typically keep their ‘pairs’ to within 0.75 of a dB from 200Hz to 10Khz.
    This attention to detail assists greatly in achieving the sought-after ‘3D holographic soundstaging & pinpoint imaging’ that some of us here constantly ‘bang-on’ about.

    Somewhat ‘Off Topic’ but still within the realm of how home audio will sound:-

    1. Harbeth is a no nonsense company when to comes to their speaker philosophy. Matching pairs is a great idea and I’m sure Harbeth is not the only company has the same practice before they go out the door.

      I see you like OC Mikie FR. He don’t pull no punches.

      1. Hi Neil,
        He has obsessed over a few things that I didn’t think were really worth obsessing over, however, yes, he calls a spade a f*****g shovel; as we say down here 😉

        1. I was reading the private emails between Paul and jb4. Paul’s response this AM said we should be happy and have fun sometimes so I made a recommendation. It may be seen as humorous or outrageously angry. A double edge sword (pardon the pun).

    2. That’s a great video, FR. Should be mandatory viewing for all self-proclaimed audiophiles. He stalled few times in the first few minutes, but really hit stride for the rest of the presentation. I admire guys like this that can just make an unedited presentation without a written script. Thanks for sharing the URL.

        1. All the stuff about mixing and microphones is spot on…buuuuut all we get to play is the end product. Then we can choose which coloration or lack there of and how to play it?

  3. All agreed and I’d say as Steven, it’s not only the icing, but essential.

    But what I’d like to know is, how the room conditions are chosen for voicing.

    Each room is different and to an extent suboptimal. This affects different speakers differently.
    Now you voice a new speaker in a room, of which you have no clue how it will affect this speaker.

    That’s like having to set the fundament of a skyscraper without having any tools like water levels or similar.

    The only reference you have is how other speakers sound in that room. A weak initial situation.

    And even if this problem wouldn’t exist, the speakers‘ voicing in that room would have limited relevance for other rooms.

    Might it be that an unvoiced speaker can accidentally sound better in a home, than a speaker voiced in a different room? 😉

    And is it better for most customers if a speaker was voiced in a suboptimal livingroom-like surrounding than in a fully optimized studio, which differs most from most home situations? 😉

    1. And that’s why I still like tone controls to finesse things by ear. I know, they introduce phase variations that interfere with “proper”, locked in imaging. Don’t care, tone rules to my way of mind.

    2. Mr. Jazznut asks about the room. A good question that can probably never be answered. It seems most voicing would need to be accomplished in the near field.
      As to the Shaw vs. O’Toole conversation, why can’t we give credence and some weight to both methodologies. While O’Toole does seem to get wound up in the weeds of his process, some results may be valuable. I can surely hear the difference in my cables, perhaps Mr. O’Toole can’t due to design differences in his equipment… he may like slightly masked sound.
      As the saying goes, my ears, my money.
      Happy New Year to all.

    1. Barsley,
      Thanks for that.
      I got 17 minutes in to your Floyd Toole link before
      I couldn’t take anymore of his ignorant babble.
      That’s 17 minutes of my life that I’ll never get back.
      Lucky that I didn’t waste more of my time watching
      the whole hour & a quarter.

      Measurements are for the aurally challenged;
      that’s all that Mr. Floyd Toole proved to me.

      1. Dr. Floyd Toole studied electrical engineering at the University of New Brunswick, receiving a BSc, and at the Imperial College of Science and Technology, University of London, where he received a Ph.D. In 1965 he joined the National Research Council of Canada, where he reached the position of Senior Research Officer in the Acoustics and Signal Processing Group. In 1991, he joined Harman International Industries, Inc. as Corporate Vice President – Engineering. In this position he works with all Harman International companies, such as JBL, Infinity, Harman/Kardon, Mark Levinson, Revel, Lexicon, AKG, Studer, DOD and Soundcraft. He also directs the Harman Research and Development Group, a central resource for technology development and subjective measurements in support of product design and manufacturing. He retired from Harmon in 2007. He is a member of the Acoustical Society of America, and a Fellow and past President of the Audio Engineering Society. In 1996 he was awarded the AES Silver Medal for “significant developments in the subjective and objective evaluation of audio devices”. Dr Toole is a Fellow of the Audio Engineering Society (AES) and of the Acoustical Society of America (ASA). He is the recipient of the Custom Electronic Design and Installation Association (CEDIA)’ s Lifetime Achievement Award.

        Dr. Floyd Toole is one of the most respected names in audio alive today, and anyone who has been in the business knows it.

        I can lead a horse to water, but I certainly cannot make it drink. Have a great day.

        1. Barsley,
          That may be, but after all that study & all those years & all that wasted effort that he has ‘apparently’ put in…he still can’t hear properly
          & audio is about hearing…not measuring.

          You can lead an academic to a home audio rig, but you
          can’t make him listen (if he’s partially deaf) I’m guessing.

          1. Wow FR… That reminds me of the guy who claims he is an authority on audio, yet when using one of the most basic of words concerning audio, pronounces “timbre” like felling a tree.. look out… timber! (it’s pronounced ‘tambər’) SMH

            1. barsley,
              No s**t Sherlock!
              So there’s some corollary between the ability to
              hear & pronouncing French words properly?
              Is that the best that you’ve got??

              I’m stunned at your enlightenment, nay, genius, sir.
              You have a great day too…it’s 12:45am here.

              ps. Toole openly admits that he can’t hear the difference between cables…this proves that he is aurally challenged.

                1. Barsley,
                  The irony that a guy who can’t hear the differences in audio cabling wants to pontificate about sound & how we hear…hilarious!

            2. Barsley,
              Please don’t waste your time with these types. For those of us that have read Sound Reproduction by Dr Toole, we know he stands on the top of the podium of audio contributions. He is the peak of the mountain, just ask anyone that has professionally studied loudspeaker design. For example, ask Chris Brunhaver. Of course, you’ll always get those that try to tear down the successful ones, but we know better.

  4. I enjoy doing the icing myself digitally, within the bounderies of the hardware. I select preset voicings and also made some myself. Each source has a default voice setting. TV: news; AppleTV: pleasant; (high res) USB: neutral; Airplay: neutral. For movies I select: movie 1 or 2 (heavy stuff). And then much more options to the liking, depending on music genre. But, to be honest, the neutral is by far used most. This is on Lyngdorf.

    But I guess I am a lazy, new generation audiophile… who does not want to play around anymore tuning synergies between hardware components.

  5. Speakers are just cool. When audio shows were going on, I would go from room to room solely for speakers. No where else in audio is there such latitude for diversity, yet the ability to achieve fascinatingly good results. It’s artistic in many different ways.

  6. A long time ago, in a galaxy far far away, before a time when electronically reproduced music lovers called themselves “audiophiles”, and before they fell prey to subjugation to the puritanical laws of sonic purity, speakers had L-Pads.

  7. WARNING ! totally off topic.

    Yesterday Paulmc.G. sent me a personal email, which I appreciated very much.
    And it’s because of this email and the fact that I respect him very much that I write this little statement.
    In his email he apologized to me for something he wrote.
    I answered his email, apologized to him and wrote that my comment (in response to his comment on my first post) was harsh and out of line.
    I felt guilty afterwards, Paul does not deserve this of course.
    I regretted my comment, but the damage had already been done with the button.
    So, yes, I was happy with Paul’s email. A chance to clear the air, and that’s what we did.
    I can say we’re fiends again, at least on this digital level (we never met in the flesh).
    And although I thought this was something between Paul and me, I have to thank some people for their emails they addressed to me.
    From RCramer I learned a new expression (something with a door hitting me). I never heard of it and had to look it up. Now I know.
    And Fat Rat, who wrote something about his favorite topic, the human arse.
    But it was also these emails that made me realize that I want to refrain from writing comments, at least for a while.
    This is not fun anymore.

    1. Thanks JB4 and I too have been getting concerned about this very topic. We do need to make sure we’re still having fun. This is supposed to be fun and while we (and me) get serious from time to time, it’s all in the service of our community which I hope we can keep with a smile on our faces and a song in our hearts.

      1. Gladiator competation was supposed to be fun for the masses and the Plebs were sure having a good time, but not always. If I were you Paul, the next time a situation like this comes up, suggest that you scream at the top of your lungs:

        “Are You Not Enertained?”


        Maximus McGowanus

        Make sure that when yous cream out, the entire rant is within 3dB for voicing purposes 🙂

      2. Back in the studio to make our latest number one
        Take two-hundred-and-seventy-six
        You know, this used to be fun

        — John Entwistle (1975): “Success Story”

    2. ‘jb4’,
      Take care, be well & do something about those lock-down
      blues or whatever the hell has you so uptight.
      Gonna miss you.
      Please come back here when you feel refreshed, renewed & less aggravated.
      Wallaby 😉

  8. Because of so many variables it would be prudent to select a speaker in the end after having selected the rest of the system. Thus the speaker that works best with the system will be the final choice. Most speakers are passive devices and produce what is fed to them, anyway. Why compromise by choosing the speaker first? Regards.

    1. I agree with Oliver here. I know Paul has said to buy the best speakers you can stretch for and build up to match them. But, with my local dealer I got a bookshelf speaker at the bottom of the Totem line for 1/3 of the price of the class A/B amplifier I bought. The amp could later accommodate just about any speaker I eventually grew to prefer, after painstaking research. Both served me well for two years. In the interim I found a really good, for me, streamer and CD transport and turntable. Only then did I pull the trigger on my floorstanders.

  9. Paul,

    Understandably, transducer designs, driver integrations, crossover development and cabinet engineering are all melded together to meet the overall speaker design criteria.

    Is Final Voicing of the FR30’s (and future PSA speaker systems) to achieve a more linear and flat frequency/phasing amplitude response? Or, does Voicing mean a desired sonic coloration (warm-bright-forward-recessed-open-closed) of the presentation? Possibly, a little of both characteristics??


  10. Speaker icing…

    There’s no more of a personal choice than speakers. The actual sound projectors. Whether they are voiced and or measured is not indicative of how they will sound in any particular room. Just look back a few posts ago when things changed from MR3 to MR2 and the room needed ‘help’.

    It seems to me that voicing is an important tool just like measurements are / can be. The initial character of any speaker’s ‘sound’ is determined in the development and manufacturing phase.
    How that sound is obtained is up to the developer(s). How much you like that finalized sound is up to you.

    The off topic subject today was also an interesting read. I’d say keep it coming. Sometimes things get misconstrued when a personal inflection is read into the written words or things are taken too literal. Sometimes the subject and responses raise your ire. Sometimes they don’t. It seems to me that second someone lets things get personal is when things can bounce off the rails. Stepping back for a few minutes and thinking before splattering out confrontational words can help. In my opinion nothing to do with listening or commenting on audio is a ‘life or death’ endeavor. Opinions abound and some like to be louder and longer than others. Sometimes the “in one ear out the other” method is the best option.

  11. I believe I met Floyd Toole at NRC in Ottawa back in the 80’s when I worked at a local acoustic company also doing research for hard of hearing and telephone handsfree performance.

    Whats your opinion on a sound system with a microphone to measure the room response and then applying frequency equalizer ?

  12. I have a pair of Harbeth M40s that I use as a main stereo pair for my digital pipe organ, supplemented by a subwoofer for the 16Hz to 32Hz low pedal bass. Even though the M40s are a matched pair with probably stellar specs in a perfect room, my room is not perfect, so their frequency response in my room is not perfect. But that’s okay. My digital pipe organ program allows me to voice in the digital domain each and every digital pipe of each and every rank. Just like the pipes of a real pipe organ are individually final voiced in the room in which the organ is installed, I can do the same with the digital pipe sound samples. It takes several days to individually tune and voice thousands of pipes, but it lets me get the even response across the frequency spectrum of each and every rank. When I get a new organ sample set, I always have to spend time regulating the volume of several pipes that resonate too much with my room, For example, certain pedal bass pipes will be too loud and others too soft, so I raise or lower the volume on the offending individual pipes until they sound even and balanced.

  13. Voicing recordings is artificial sweetening, high glycemic audio. Equalization is distortion, and you can’t fix distortion by adding distortion, nor fix vector anomalies by scalar processing.

    Rooms have frequency response worse than speakers; but well trained ears can learn to hear through the response of rooms, like musicians who practice and rehearse in non-flat rooms every day, and audiophiles who break-in their ears to their ‘sweet spot’ seat in front of their favorite speakers.

    What I found odd is that musicians grow insensitive to frequency response deviations, and audiophiles grow sensitive to them. This is well documented by the frequency response training software system developed at Harman International.

    This suggests a far more stringent criteria than frequency response, power response, or even listener position response: flat frequency and phase response at all angles. This is because the designer can’t predict which polar angles will result in first reflections or second reflections at the listening position. Since we learn to compensate for rooms when we use them to converse (cocktail party effect), we can also learn to compensate when we listen to music or audio in different places in different rooms – IF AND ONLY IF the speaker is neutral wrt room acoustics.

    This does not mean the speaker has to have the SAME SPL at all angles; merely that at any angle one chooses, no frequency has different response than any other (i.e. Kii 3, Linkwitz Labs).

    1. When I say I voice the digital recordings of the individual pipes, these are slight adjustments to their volume, pitch and, if needed, tonality of a pipe. This improves the sound and more than offsets any added distortion, which is negligible because the voicing tweaks are small. I would rather hear the beauty of pipes in tune than the dissonance out of tune pipes. If that is “artificial sweetening,” so be it. It satisfies my sweet tooth. My well trained ears cannot tolerate out of tune pipes.

  14. Some speakers are room specific and you need to use some room tweaking to get them to sound good if your room is not what they were meant for. Others are engineered to sound good in a variety of rooms. The character of the sound in the later could change room by room but they still sound good.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

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