No announcement yet.

Thoughts and measurements of on on-axis, off-axis & power response on sound quality

  • Filter
  • Time
  • Show
Clear All
new posts

  • Thoughts and measurements of on on-axis, off-axis & power response on sound quality


    For years, I've wondered how off-axis response affects the listening experience. Also, what power response or in room response are. And why certain speakers seem to have a great on axis response but tend to be "shouty" in the midrange in a real room. This post is all about finding answers to these questions.

    The idea

    A couple of months ago, I asked a question here:

    "What is better, a 10" woofer with a waveguide, or an 8" with a very robust tweeter?"

    There weren't any real answers, so hey. I decided to build both and see how they performe (from an objective and subjective point of view). I also added my own commercial speakers to the mix.

    If I knew about on-axis, off axis, power response and in room response, and compared every speaker side by side, I thought, I'd understand what it was all about. And indeed, I learned about it.

    My findings.

    Results first, in case you don't want to read through all the test.

    1) You cannot design a speaker with one or two off-axis measurements and an on-axis one. Likewise, if you design for on-axis response, you'd better pray your off axis is correct. And it usually won't be.
    2) Power response and in room response are basically at least as important as on-axis response.
    3) Designing for power and in room response (downward sloping lines) gets you a "neutral" balance, with no colorations or "shortness". Designing for flat on axis response has the most impact on how you perceive the sound of the speakers.
    a) Confused? Allow me to explain. My two speakers have a downward sloping power response and they sound balanced and uncolored, without any major issues. But the flat on axis one is clearly superior, with more perceived detail.
    4) Driver selection, baffle width, waveguides and driver design are extremely important. Although, as you'll see, my favorite speaker was the one that I actually thought "would never work".

    A couple of notes.

    1) Measurements are as real as they get. I don't have an anechoic chamber, but I do have a room 20 feet above the ground, and with no walls around it. I place the microphone at 1m away and I have resolution up to about 25 Hz.
    2) I did take every off axis measurements and checked phase at each step.

    The contenders.

    1) The JBL Clone.

    Click image for larger version  Name:	8inch.jpeg Views:	4 Size:	306.6 KB ID:	1477149
    A clear (at least to me) homage to JBLs of the past, this is the 8" (SB Acoustics) woofer, mated to a Peerless Corundum 32TX tweeter. I thought this one could borderline work. With a 6th order L-R symmetrical crossover at 1.2 KHz. it wasn't cheap. Or easy to build.

    It also has lots of compromises regarding FR.

    Click image for larger version  Name:	Captura de Pantalla 2021-10-07 a la(s) 14.03.12.png Views:	4 Size:	515.3 KB ID:	1477150

    Yes. That is the on axis response with a mega BBC dip on it. Because the tweeter is extremely directional at around 2 KHz. As you can see on the on axis graphs, I basically had to compensate for its response off axis by massaging the on-axis response, and this helped get a somewhat decent power response.Phase tracking is basically perfect.

    2) The old school clone

    This one shouldn't work at all. An old Morel 10" woofer I had around, an old Dayton RS-28 aluminum tweeter, and a Visaton WG-148 waveguide. The previous speaker was built with all new components. This one was built with things I had around the house. Plus, IMHO it looks weird. Almost ugly. Although many people have said it looks better than the previous one.

    Click image for larger version  Name:	10inch.jpeg Views:	4 Size:	393.3 KB ID:	1477151

    Click image for larger version  Name:	Captura de Pantalla 2021-10-07 a la(s) 14.17.47.png Views:	4 Size:	579.4 KB ID:	1477152

    The crossover point is around 1.25 KHz. It's asymmetrical. 6th order on the woofer, 4th order on the tweeter. As you can see on the transfer function, it has a couple of corrections to the FR of the tweeter. The waveguide is bviously not part of the tweeter, so response is not exactly perfect. But with a couple of notches, everything is fine. Directionality is great, off axis is very good, and up to 75ª, response is pretty decent.

    I was surprised it worked as good as I did. I wasn't expecting a 10" woofer to mate with a 1. 1/8" tweeter, and yet, it did.

    3) The commercial speaker.

    Focal Aria 926s. I didn't do any measurements on them. But here's some data.

    Other than a slight flare starting at 4 KHz off-axis, I couldn't see anything out of the ordinary.

    Coming up... in room measurements.

    Line Array: IDS-25 Clone, FE-83.
    2-2.5 Way:
    Zaph Audio's winning entry: ZA5+SB29. - Microliths: RS125+RS28. - Small Bangs: TB W4-1658SB+SEAS 27TBFC/G. - Monoliths: Peerless 830884+SEAS 27TBFC/G.
    3-3.5 Way:Miniliths: SEAS P21/CA21REX+Neo8 PDR+Neo3 PDR. - Megaliths: 2xDayton RS270+2xT-B W4-1337SB+SB29. - ZDT3.5 +: 2xDayton RS180+Dayton RS52+Vifa DQ25. Reflexos: OB 4xDayton RS150 + Neo3 PDR.

  • #2

    I should have done measurements second and listening tests first. But due to how I was placing the speakers (by hand, etc) I could not avoid looking at the measurements before listening to the speakers. Still.. patterns emerged.

    In room measurements

    I adjusted SPL's and took a REW sine sweep of every speaker in the same location at the same SPL's. Here are the results.

    Red: JBL clone
    Green: Old school clone.
    Orange: Focals.

    Click image for larger version  Name:	Graphs.jpg Views:	0 Size:	118.9 KB ID:	1477154

    First thoughts. Wow. Those are pretty similar! But the graph is also pretty squished. So...

    The JBL clones have more deep bass than every other speaker. They also have a slight excess of energy in the presence region (1.5 - 2 Khz) - really slight. Other than that, and a bit more energy in the mid bass, they are pretty good.

    The old school clones are a sealed design, so they don't have as much bass as the other two. Starting from 150 Hz and down, they are down at least 3 dBs than the other two. But hey have a terrific, flat response from 500 Hz to almost 9 KHz. Other than a bit of a flare at 4.5 KHz or so, they are the best of the trio.

    The Focals don't have as much energy as they should from 1.2 KHz to 4 KHz, and they have a bit more energy from 4.5 KHz upwards. They have great bass, but not deep. They must be ported somewhere around 40 Hz. The JBL clones are ported at 24 Hz. I was surprised either one of my speakers were flatter in room than the Focals, actually.

    Listening impressions.

    I took 10 tracks I'm very familiar with, and listened to them in sequence on each speaker. Took some quick notes and then compared them. This is what I found - ranked

    3) The JBL clones

    * Most inefficient speaker of the group. By far. Really hard to drive.
    * Tone is correct, and bass goes extremely deep. Easily my favorite bass response of the three.
    * "Creamy sound" which turns out to mask a bit of detail, especially on female voices.
    * When sounds like horns come in, sometimes they can hurt a bit. They are a bit more agressive in the midst than the others.
    * Sounded great with jazz music and as a background speaker.
    * I could live with them, but they lack detail in the mids, and sound a bit muffled and recessed when compared to the others.

    2) Focal Aria 926

    * Extremely efficient. Crazy efficient. I can drive them with my McIntosh MC-30s to loud levels. GREAT!
    * Tone is mostly correct, but they tend to become a bit too airy for their own good. The extra energy up on top is really noticeable. It's not bad - only when compared to the other two.
    * Bass punch is great, but extension isn't. They don't have deep bass. But the quality of it is really good.
    * Midrange is more present than the JBL clones, especially on instruments like piano.
    * Horns sound perfect and recessed. I loved how this speaker never sounded aggressive, but clear and airy.
    * I could definitely live with them. The 1.2 KHz - 4 KHz dip is not as objectionable as I thought. The extra energy on top is like a cherry on top of a great cake. Really liked them!

    1) Old school clones

    * Treble is extremely good. It's not as airy as the Aria, but it sounds a lot less emphasized.
    * Bass is not as present but it's extremely good in quality, possibly the best. Very well defined.
    * Midrange is just amazing. Female voices jump out of them when compared to the others. I absolutely love how they render human voice.
    * Horns, etc, are not tiring. They are more "forward" than on the Focals, but not tiresome.
    * I just EQ'd them with a bit extra bass (+3 dB) and they are my favorite of the bunch.


    I now understand a bit better why it's important to look at power response and on-axis response.

    I tried to "flatten" the on axis FR on the JBL clones. They were incredibly shouty and tiresome. I could not live with them. I had to adjust the on-axis curve to lessen the impact on the off axis curves. Tonality was correct, but I lost a lot of detail on the mids.

    The Old school clones shouldn't work at all. A 10" woofer going to a 1 1/8 tweeter? Impossible. And yet, the waveguide and a well behaved woofer did their magic. I could find no objections to it. Even though the Focals have a bit less distortion, the waveguide and off axis response make these, with a bit of EQ, my favorite speakers of the three. Flat on axis, smooth off-axis - Floyd Toole and Sean Olive couldn't be wrong, and they weren't. That's the trick.

    The Focals are great. Simply great. They seem like a speaker that was designed with listening in mind. They are so fun to listen to, if not entirely neutral. Even on pink noise you can hear the extra bass and treble. I guess it's a bit of the California sound, with a French interpretation.

    Next steps.

    A two piece speaker, with a very narrow top module built out of a SEAS DXT tweeter and a 4" mid-woofer, and a bottom module with 2 8" woofers sounds like something that could be in my future...

    Line Array: IDS-25 Clone, FE-83.
    2-2.5 Way:
    Zaph Audio's winning entry: ZA5+SB29. - Microliths: RS125+RS28. - Small Bangs: TB W4-1658SB+SEAS 27TBFC/G. - Monoliths: Peerless 830884+SEAS 27TBFC/G.
    3-3.5 Way:Miniliths: SEAS P21/CA21REX+Neo8 PDR+Neo3 PDR. - Megaliths: 2xDayton RS270+2xT-B W4-1337SB+SB29. - ZDT3.5 +: 2xDayton RS180+Dayton RS52+Vifa DQ25. Reflexos: OB 4xDayton RS150 + Neo3 PDR.


    • #3
      Looking at that last comparison FR plot really brings home how subtle response differences can cause such large auditory differences/ perceptions in the listener.

      Excellent work! Thank you for sharing it.


      • #4
        The Old school clones look great to me! Nice work all around.


        • #5
          Very interesting experiment! Are you gonna share the crossovers designs? Also what morel woofer? I think I might have a pair stashed around here.
          keep up the great work!!


          • #6
            I'm also curious about the design of the Old schools, I have found Morel woofers to be excellent. I am looking for a sealed design for my garage since I don't want dust getting in a port.


            • #7
              Hi Xavier,

              Great post. I only read the first post and jumped here to comment. May I make a few comments?

              From your measurements I predicted that the Focal Aria, followed by the second design (10" woofer with tweeter and waveguide) would be the most preferred, and the JBL clone a distance third.

              Now I'm going to shoot myself in the foot here, if I get it wrong, but allow me to explain...


              • #8
                This is what I know, about loudspeaker measurements, as of 2021...

                Well worth watching from beginning to end.

                But the summary is:

                Click image for larger version

Name:	image_92539.jpg
Views:	218
Size:	184.9 KB
ID:	1479721
                Click image for larger version  Name:	Sean Olive and Erin Study Limitations.jpeg Views:	0 Size:	163.0 KB ID:	1479722