B&O Tech: “Auto” loudness

#76 in a series of articles about the technology behind Bang & Olufsen loudspeakers

If you look at the comments section to a posting I wrote about ABL, you’ll see a short conversation there between me and a happy Beomaster 8000 customer who said that I had made an error in making sweeping generalisations about the function of a “loudness” filter in older gear. I said that, in older gear, a loudness filter boosted the bass (and maybe the treble) with a fixed gain, regardless of listening level (also known as “the position of the volume knob”).  Henning said that this was incorrect, and that, in his Beomaster 8000, the amount of boost applied by the loudness filter was, indeed, varied with volume.

So, I dusted off one of our Beomaster 8000’s (made in the early 1980’s) to find out if he was correct.

 

The Beomaster 8000 under test. I lied when I said that I dusted it off… (Keen-eyed viewers may recognise the insides of a Beolab 90, screwed to the white board in the upper left corner of the photo. That’s used for measurement-based tests that don’t require listening… The way you can tell it’s a Beolab 90 is the circular PCB at the bottom of the board. That circle is the 72 LED’s that normally sit at the top of the loudspeaker.)

 

I sent an MLS signal to the Tape 1 input (left channel) of the Beomaster 8000, and connected a differential probe to the speaker output. (The reason for the probe was to bring the signal back down to something like a line level to keep my sound card happy…)

I set the volume to 0.1, switched the loudness filter off, and measured the magnitude response.

Then I turned the loudness filter on, and measured again.

I repeated this for volume steps 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, and 5.5. I didn’t do volume step 6.0 because this overloaded the input of my sound card and created the weird artefacts that occur when you clip an MLS signal. No matter…

Then I plotted the results, which are shown below.

 

Remember that these are NOT the absolute magnitude response curves of the Beomaster 8000. These are the DIFFERENCE between the Loudness ON and Loudness OFF at different volume settings.

At the top, you see a green line which is very, very flat. This means that, at the highest volume setting I tested (vol = 5.5) there was no difference between loudness on and off.

As you start coming down, you can see that the bass is boosted more and more, starting even at volume step 5.0 (the purple line, second from the top). At the bottom volume step (0.1, there is a nearly 35 dB boost at 20 Hz when the loudness filter is on.

You may also notice two other things in these plots. The first is the ripple in the lower curves. the second is the apparent treble boost at the bottom setting. Both of these artefacts are not actually in the signal. These are artefacts of the measurements that I did. So, you should ignore them, since they’re not there in “real life”.

 

So, Henning, I was wrong and you are correct – the Beomaster 8000 does indeed have a loudness filter that varies with volume. I stand corrected. Thanks for the info – and a fun afternoon!

  1. Henning Hansen says:

    The guys designing those loudness curves were brave, and they were right. Except for the first few volume steps where the correction was indeed too much, then from 0.5 and upwards the correction was spot on, in my opinion. And allowed us/me to listen at both high levels and “whisper” like levels with a pleasing sound balance.

    So no or minimal treble correction and a really healthy bass correction made a good sound. And that logarithmic volume control with 60 steps incrementing 1.5 dB per step. That was a smooth performer.

    Maybe you should consider two loudness options ( if on ) for the Beolab 50. A “conventionel” and a “Beomaster 8000” setting ?

    Regards and thanks for your detective effort.

  2. Hi Heine,

    Thanks for the info! Looks like an interesting device for someone who is regularly testing amplifier outputs using a “normal” sound card.

    I haven’t gone through the full description, but I wonder if it also has a filter option for testing Class-D amps. Probably not – although that’s certainly something that you could insert in the signal path between the amp output and the auto ranger.

    Cheers
    -geoff

  3. Hi Geoff

    You are right – measuring Class D amplifiers is a special situation requiring probably an AES17 filter and perhaps an additional external lowpass filter AS well. Jan’s autoranger provide neither of these. Shame on him :-). I hope he will create an additional filter for these measurements. Alternative could be a TI filterbox that taget provide. Anyways who makes active loudspeakers with linear amplifiers anymore?

  4. Hi Heine,

    It’s not difficult to find a Class D filter if you need one (AP has some that are simply put in the chain between the amp outputs and the analyser inputs) – and without it, a measurement of such a switched-mode amp is probably a waste of time, not only because you’ll get the wrong response, but also because the out-of-band artefacts may confuse the auto-ranger…

    Linear amps in active speaker? It’s not unheard of… I can think of at least one pro-audio monitor that would argue that this is a good idea… :-)

    Cheers
    -geoff

  5. Adding Kef LS50W (hybrid) and Genelec 8351 (also hybrid)

    Good input – having external filters before the autoranger. Perhaps adding a pre analyzer filter like suggested in https://www.slideshare.net/mobile/shenpei/class-d-and-linear-power-amplifier-testing. A cheap solution could be http://www.ti.com/general/docs/lit/getliterature.tsp?baseLiteratureNumber=sloa107&fileType=pdf although it’s not aes17 compliant otherwise as a alternative to AP http://www.avwidgets.com/AVWidgets/AVW17.html

    Btw great work that you and your colleagues are doing – love to read your blog about how you guys are moving the loudspeaker tech forward.

Leave a Reply