Reverse of the old Loudness Control

[Don Hills]

Schematic: See attached. Ignore the type of the opamp, I just grabbed one from the symbol library. Note that all 3 pots (6 for stereo) need to be ganged.
Its operation is quite simple: 3 cascaded +-15 dB stages, the first two are frequency independant, the third one is the bass half of a standard Baxandall tone control. Note that the third stage pot has to be wired "upside down" compared with the first two, to reduce the bass gain as the main gain increases.

Frequency response: see other attached.

I used 2 gain stages, you can get 30 dB boost/cut in a single stage but the control law doesn't match that of the bass boost/cut stage. It might be possible, but opamps are cheap.

This is still proof of concept, waiting on finding some suitable high quality digital potentiometer chips. Also, the time constant of the bass control stage is not finalised. I'm in the process of recalculating the ISO E-L curves into more conventional bass cut curves to calculate the capacitor value required.
There's some Matlab code available on the Web that calculates the actual curve values. I ported it to REXX and I'm modifying it to do the calculations.

[Don Hills]

Originally Posted by pheonix358

Would someone please be kind enough to post the first section of the schematic of the July '96 parametric equalizer, it's a shelf/subsonic filter. It's over 13 years old so copyright would have expired.

Terry

I think I have it. I'll take a look. Remind me if I haven't done it by the end of the week. Copyright has not expired, I thnk the 13 year term is for a patent.

[pheonix358]

Hi Don, I don't believe you need such a complex circuit to achieve what it is you need to achieve. A single stage parametric shelf would do what you want, and since you know the speakers you are using the frequency and width controls could use preset pots with the last control, the gain, linked to the volume with two simple gears, one on each shaft. This sets the slave pot in reverse. Simple is often better.

Terry

[Don Hills]

It may be simpler, but is it going to track? That is, when you reduce the main volume by (for example) 10 dB, is it going to reduce the gain at 20 Hz by 5 dB? (In other words, will the parametric EQ provide 5 dB of boost?) That's the problem I ran into - trying to match the parametric equalisation law to a log law(*) volume control.

(*) A log law volume control is rarely a log law. It's usually two different values of linear track, joined at about the 10 or 12 o'clock position.

I'm starting with a general model, that will work with all speaker systems without modification. LF shelving to suit specific speakers can then be added on a case by case basis. In many cases it will only require changing the value of the capacitor in the bass gain stage, but you can use more complex networks instead of just a single capacitor if you want a specific curve. And you can add a complementary output for a powered sub, so that it progressively takes over the bass duties from the OB as you wind the volume up.

[pheonix358]

Interesting questions! I don't think any system will track in such absolute terms. The question is do you really need them to. Generally we listen to music at certain listening levels e.g. background, normal, loud and 'rock the planet, I love this song'.

If you use 'real log' pots then is it simply a matter of using identical pots with one in reverse? In my experience these systems work quite well. I don't sit on the volume control while listening, it defeats what the sound engineers designed the music to be like.

I suggest you may want to sit down and think what you want this control to do in the real world. List the must haves, and the would be nice to have, and then the silly things (also refered to as marketing hype). See what you come up with.

Terry

[Don Hills]

Even if the laws don't track exactly, you still need to have them track in a predictable and desired manner. You effectively need to be able to set / specify equalisation curves for each of the range of levels you normally listen at, then arrange to automatically switch between them or smoothly transition from one to the next.

Are 'real log' pots as excruciatingly expensive as they used to be about 20 years ago?

But even if you use log and anti-log (or geared log) pots, you still have a control law problem. Take a look at the graph of control resistance versus rotation for log and anti-log pots. Log pots have most of their resistance change at the anti-clockwise end of their travel. Anti-log pots have most of their resistance change at the clockwise end. They don't "track". in practice this means you'll get too little bass boost at low volume, and too much bass cut at high volume.

I've thought very carefully about what I, personally, would like this control to do. I'm also trying to keep Jamikl's requirements in mind. So far, it looks like I can meet his requirements as well as mine with variants of one basic design.

Anyone motivated to see if they could use such a control can try this experiment:

Place some sticky labels around your bass and volume controls so you can make temporary scale marks.

Play a wide variety of familiar music at various settings of the volume control. Mark and label (number etc) the settings. For each setting, adjust the bass control to give what you consider to be a pleasing (preferably natural sounding) result. Mark the bass control positions with matching labels.

Now you will need a source of test tones and a dB meter (a multimeter with a dB scale will do), or a simple voltmeter if you have basic math skills to convert voltage differences to dB.

The test tone source can be a signal generator, or a PC with a tone generator program, or a test tone CD etc. You need 2 tones, one at about 20 to 50 Hz and one at about 1 KHz. (Pick the LF tone frequency to match the LF rolloff point of your speakers.)

Connect the meter instead of (or in parallel with) one of the speakers.
For each marked setting of the volume control (and matching setting of the bass control), measure the output at the two frequencies. It's OK to adjust the level of the test tone source to avoid overloading the meter as you adjust the volume control, but adjust it by a known amount (say 10 dB) and adjust your results by the same amount.

The end result should be two columns of numbers. If you plot the results on a graph, you should see a definite trend. You'll very likely see that at as you lower the volume level, you'll be turning the bass up further and further. It'll be very useful to see if a consensus emerges as to how much bass boost is needed as we turn the volume down (and vice versa), or if we all prefer wildly different amounts of "loudness compensation".

Jamikl, I'll be particularly interested in the results if you choose to try this. It'll give a much clearer picture of the amount of equalisation you would need for your particular system.

[pheonix358]

Thanks Don, that really specifies things nicely. I rarely use my tone controls, in fact getting up to look at where they are set, the're not in circuit.

I guess that is for two reasons, I either listen at background levels and don't care much or I listen at a good strong level and have never wanted to use tone controls. My speakers f3 is 32Hz so I have more than adequate bass.

I will be interested in following your developments.

Terry

[Don Hills]

Yeah, I can see how the proposed control would be of limited usefulness in your specific environment.

I'm doing it because it's something I've thought of doing for quite a few years, and no-one else appears to have done it. This is diyaudio forum, after all. It just took Jamikl's suggestion to push me into getting started on it.

[j.michael droke]

Originally Posted by pheonix358

Interesting questions! I don't think any system will track in such absolute terms.......

I suggest you may want to sit down and think what you want this control to do in the real world. List the must haves, and the would be nice to have, and then the silly things (also refered to as marketing hype). See what you come up with.

Terry

Suggestion: When dealing with tapped horns (TH), it is essential to have a circuit which cuts off the bass signal about 3-5 hz lower than the TH's low corner, to prevent exceeding driver Xmax.. no bass bost is exceptable below this point.

[pheonix358]

IMHO the same applies to port output in a BR. In fact, all loudspeakers benefit from control of low energy. It is difficult to get amplifier guys to think this way. They seem to want their creations to be flat to DC. Looks good on spec sheets, but it really does limit max speaker output, especially with modern music or smaller speakers.

Terry