To give an example how not to do it:
In the 1990's, I read an article by Tomlinson Holman and Frank Kampmann, "Loudness compensation: use and abuse", Journal of the Audio Engineering Society, vol. 28, nr. 7/8, July/August 1978, pages 526...536. They explain that loudness compensation should be based on the differences between the equal-loudness contours at different volumes, as the purpose is not to completely equalize your ears, and they wrote a lot about what in their opinion were the then latest and greatest equal-loudness contours. Those were the ones they quoted from their reference 10, S. S. Stevens, "Perceived levels of noise by Mark VII and decibels (E)", Journal of the Acoustical Society of America, vol. 51, 1972, pages 575...601. They showed that only the low frequencies required correction.
Looking at the differences between those equal loudness contours for different volumes, I noticed that the required amount of loudness correction at low audio frequencies increased by about 5 dB for each 10 dB of volume decrease. That is, when you reduce the volume by 10 dB, the low frequencies should only drop by 5 dB. I then came up with this scheme:
The idea was to convert the input signal into two currents, one unfiltered one and one that dropped of at 20 dB/decade from a low audio frequency. The unfiltered signal would pass through two mechanically coupled linear potmeters, while the filtered signal would pass through only one of them, and through a separate potmeter (the one marked "Loudness correction") with which you could set the desired amount of correction. Change the volume, and for every 2 dB of change, the filtered signal only changes by 1 dB.
Because I rather liked designing discrete amplifier circuits, I designed and built this for the voltage to current converters:
The left output is the unfiltered output, right is filtered. I built it in dead bug style on a cut-open tin can; back then, those were actually made of thin steel, and were quite solderable. I had some oscillation issues I could solve with stop resistors and local decoupling, all shown on the schematic.
Now why did I write that this is the way not to do it?
1. When you want a large dynamic range, converting the audio signal into current the way I did it makes you quite sensitive to the noise of the various bias current sources. Hence, I needed to reserve quite some voltage for the degeneration resistors of the current sources, and hence the +24 V and -24 V supplies. I had originally intended to just use +15 V and -15 V.
2. I always had the subjective impression the loudness control didn't do enough.
3. When you increase the volume, there is always a small amount of bass boost left, unless you turn the "Loudness control" potmeter entirely anticlockwise.
4. You need a quadraphonic linear potmeter for setting the volume of a stereo preamplifier. I mechanically coupled two stereo fader potmeters, as I couldn't get a proper quadraphonic one.
5. The 1978 equal-loudness contours are no longer applicable, we now have ISO 226, https://en.wikipedia.org/wiki/Equal-loudness_contour
6. When I got a partner, she didn't like the looks of the preamplifier that this circuit was a part of much:
In the 1990's, I read an article by Tomlinson Holman and Frank Kampmann, "Loudness compensation: use and abuse", Journal of the Audio Engineering Society, vol. 28, nr. 7/8, July/August 1978, pages 526...536. They explain that loudness compensation should be based on the differences between the equal-loudness contours at different volumes, as the purpose is not to completely equalize your ears, and they wrote a lot about what in their opinion were the then latest and greatest equal-loudness contours. Those were the ones they quoted from their reference 10, S. S. Stevens, "Perceived levels of noise by Mark VII and decibels (E)", Journal of the Acoustical Society of America, vol. 51, 1972, pages 575...601. They showed that only the low frequencies required correction.
Looking at the differences between those equal loudness contours for different volumes, I noticed that the required amount of loudness correction at low audio frequencies increased by about 5 dB for each 10 dB of volume decrease. That is, when you reduce the volume by 10 dB, the low frequencies should only drop by 5 dB. I then came up with this scheme:
The idea was to convert the input signal into two currents, one unfiltered one and one that dropped of at 20 dB/decade from a low audio frequency. The unfiltered signal would pass through two mechanically coupled linear potmeters, while the filtered signal would pass through only one of them, and through a separate potmeter (the one marked "Loudness correction") with which you could set the desired amount of correction. Change the volume, and for every 2 dB of change, the filtered signal only changes by 1 dB.
Because I rather liked designing discrete amplifier circuits, I designed and built this for the voltage to current converters:
The left output is the unfiltered output, right is filtered. I built it in dead bug style on a cut-open tin can; back then, those were actually made of thin steel, and were quite solderable. I had some oscillation issues I could solve with stop resistors and local decoupling, all shown on the schematic.
Now why did I write that this is the way not to do it?
1. When you want a large dynamic range, converting the audio signal into current the way I did it makes you quite sensitive to the noise of the various bias current sources. Hence, I needed to reserve quite some voltage for the degeneration resistors of the current sources, and hence the +24 V and -24 V supplies. I had originally intended to just use +15 V and -15 V.
2. I always had the subjective impression the loudness control didn't do enough.
3. When you increase the volume, there is always a small amount of bass boost left, unless you turn the "Loudness control" potmeter entirely anticlockwise.
4. You need a quadraphonic linear potmeter for setting the volume of a stereo preamplifier. I mechanically coupled two stereo fader potmeters, as I couldn't get a proper quadraphonic one.
5. The 1978 equal-loudness contours are no longer applicable, we now have ISO 226, https://en.wikipedia.org/wiki/Equal-loudness_contour
6. When I got a partner, she didn't like the looks of the preamplifier that this circuit was a part of much:
How did you do it - having extra sections on the volume pot driving the loudness stage?
Just as FYI ...
I have crafted a PCB version of the Doug Self preamp featured here ' Doug Self Preamp from Linear Audio #5 ' that uses his December tone controls in place of those in the LA #5 article. I've just got the PCBs back from fab and need prove the layout as functional. Should anyone be interested, I will post to this thread when I have a working prototype.
My cheesy Altec computer desktop speakers
Has Bass and Treble control.
Depending on how high or low my listening levels are.
I just turn those 2 knobs to taste.
Rather effective
Also depends what recording im listening too.
But would also like to play with a circuit which just
adjusts the bass level at low levels.
Its mainly the knob I use.
Has Bass and Treble control.
Depending on how high or low my listening levels are.
I just turn those 2 knobs to taste.
Rather effective
Also depends what recording im listening too.
But would also like to play with a circuit which just
adjusts the bass level at low levels.
Its mainly the knob I use.
