> this would be a reasonable audio signal generator for someone on a budget with very little room?
For 15 bucks, it's terrific.
Listen to its 1KHz sine output, and then turn your treble control way down. Does the sound change? That type of function generator will have DEAD flat output over the audio band, but several percent sine distortion, a lot of little corners on the "smooth sine" because it is really a soft-clipped triangle. You can do a lot with a crummy sine, just remember that it has harmonics all up the audio band.
I would not like to see it bid-up to 50 bucks/pounds/euros/clams, because we "should" be able to do a lot better for $50-$100. However the next best thing, the Global Specialties 2001, seens to sell for $200 if you can find it in stock.
Nearly any of the good-old H-P, Heath, Philips oscillators 1950-1980 would be a better deal, especially since they can be had $5-$75 as-is, and usually work fine. But they run from big to BIG. And because they were built to last, if they do develop a fault it can be hard to figure out.
> A wide-band voltmeter or a 'scope?
IMHO, you must have a classic Vacuum Tube Volt Meter VTVM. The basic models were not accurate below a fraction of an AC Volt, and some were not flat to 20Hz, but most are dead-flat to far above the audio band (my Heath is perfect at 200KC and isn't falling at 1MC). The resolution does not compare to a DVM, but the dumb analog display will show trends and short peak/dips much better. And unlike these digi-meters, there is nothing to go screwy on high frequency AC: hollow-state rectifier, cap, DC meter.
Whatcha really need is BOTH an audio source and an audio meter that read flat against each other. Set at 1KHz, note the reading, sweep 20Hz to 20KHz, should hardly change. It is possible to live with un-flat gear: one of my 200ABs droops at the top of each band so I remember to add a db at 200, 2KC, 20KC to check general trends, but if you need db accuracy it is too tedious to be adding and subtracting 0.7dB here and 1.25dB there.