Audio test equipment are seldom seen in my workshop. I seem to have accumulated spectrum analyzers, modulation meters, RF signal generators, GDO, RF bridges, and loads of other "fun" things. About the only audio equipment I posess is an EPROM AF generator and an oscilloscope. This project rectified that matter, mainly because I wanted to couple out from my computer, for measuring audio levels. It was not just me, I built a stereo version for the multimedia department for which I work. If you listen to most MPEG, AVI and WMV files, you will find the audio levels all over the place. Our department always peak at 100% (32768) and has an average speech level of 30%. This level meter is the tool we use in the multimedia department.
The meter can be coupled to the LINE-OUT of your computer and calibrated with a pair of 30% and 100% WAV (MP3) files. Another use is to calibrate the meter directly in volts, or millivolts, and you have a decent analogue audio millivolt meter. It uses discreet components. All components are "generic", so the transistor types, for example, could just be any old "Silicon PNP" or "Silicon NPN". The NPN could be 2N706, 2N4286, 2N2222, etc. but I used BC547 (NPN) and BC557 (PNP).
The circuit given below is a simple audio-amplifier powered from a 9v battery. The first two transistors are a voltage amplifier, and the right-hand pair form a complementary output pair. The two diodes between the bases of the output transistors give the required bias, but if the output devices ever draw too much current (or get even slightly warm) then increase the 1K0 resistor to another value until they do not get in the slightest bit warm.
The gain of the amplifier is controlled by the negative feedback, which in turn is controlled by the 470 Ohm resistor (R4) and the 4K7 (5K) pot. If the pot is set to about 100 Ohms then the voltage gain of the amplifier is about 5. But, that gain only occurs when the 4-diode bridge rectifier is conducting. If the point of the waveform is less than the forward conducting voltage of any diode, then the gain is probably around 250,000. When the diodes conduct, they rectify the signal and pass the current through the meter. In this way the diode conduction / amplifier transfer characteristic has become more linear. The value of the POT determines how much current actually flows. This circuit works well with a low sensitivity 20mA FSD meter, but works equally well with those cheap 100uA meters you get from most component shops.
If you do not need accuracy, for example, a simple audio level meter application, then a good source of cheap meters is to rob one from a battery tester. You know, the ones you get from those "Same Old Crap" (SOC) shops, along with mosquito repellors, singing coffee cups, dildoes, battery fans, cheap'n nasty radios, pirate Ericsson mobile phone "handsfree" cords that fall appart, nose-hair trimmers and Christmas cards that play stupid squeeky tunes and won't stop playing, even after you have closed the bloody thing so you have to rip out the damn battery to get it to stop, and that means ripping appart the card my mean auntie Julie sent me, with the message "Don't throw this card away, nephew - it's your Christmas present!" - Oh! I am sure you know the kind of shop I mean.
I'm sorry about all this complaining, but I am English, and you have to make some allowances - we're not all perfect, you know.
The project needs no special tools, no setting up (other than a single sensitivity adjustment) and almost all components have an extremely wide range. With low sensitivity meters then the 5K Pot could be replaced with a 500-Ohm pot, or replaced with a 25K pot if you used a 30uA meter. The potentiometer (pot) can be any type of device, from a small pre-set pot glued to the board, to a normal 1950's type volume- control mounted in a hole drilled in the board. I have marked on the artwork exactly where to drill the hole. If there is any really serious interest in this project then I could even make it as a kit, IF I can find a suitable cheap and repeatable source of new meters (I doubt if I could supply a battery tester instead!). I have drawn a PCB foil pattern, and this is to be found on my download section, along with most of my other projects.
The first 1970'2 prototype ... er! ... well! You really don't want to see that! But
the latest version was housed in a project box with two units for stereo operation, as
shown at the beginning of this article. The switch in the photograph selects one of
two stereo input sourcess so it can be used to meter the output of two different
computers independantly. The button in the picture is the battery power switch as
the unit is only used for a few seconds. 
This
particular version is powered continuously with a battery eliminator (a "liberated"
T888 cell-phone charger); the battery is only used for portable working.
The meter scale I removed and scanned, then using a high-quality graphics editor (paintbrush.exe from Windoze 3.11). I coloured the section of 30% (20% to 40%) in green since that was the desired reading in our application. I also took the liberty of adding the text "AF LEVEL". The edited file was printed using a colour printer and the paper scale was glued to the reverse side of the original scale and the paper trimmed. The result is what you see here. If you are making your own level meter then you will need different graphics, so it is not possible for me to provide any of these files here.
The complete instrument does not need screening and has been 100% stable. I first built this particular circuit in the 70's and it has become one of my standing building block circuits. If you look at my amplifiers section then you will see just where I got it from. Incidentally, it also detects low RF - just put a 1uf across the meter and keep all the component leads VERY short.
Have fun with the circuit. Very best regards from Harry - SM0VPO