Hi,
I want to give you some reference material, this is to see the difference between an “Analog” generator and many function generators.
Let me start with an oldie, this is the Heatkit IG-1272 that I bought and built at the end of 1970.
Early this year I made it working again, just not tuned for low distortion and this can be clearly seen in the photo below.

Many audio devices and transmitters I have measured through with this model.

.
The THD, not so good at the moment...
But more important, look at all these harmonics from de 50Hz power frequency...

.
And now a 1KHZ signal from an Audio Precision “Portable One Plus”
First a picture of the measuring instrument itself, so you know what I'm talking about.

.
This is a picture to kiss so beautifully.
But now what am I seeing below 1KHz....
The grid frequency and its harmonics....
This is one of the few problems I have encountered with this measuring instrument his purchase in 2000.
There is too little shielding applied in the housing.
For certain measurements I would sit with the instrument on a few chairs in the middle of a room as far away as possible from other equipment and power connections.
If I measure with the nice sound card that is in another computer with Virtins' software and I adjust the level a bit for the lowest distortion,
I get a value of 0.00022% THD when I measure up to the 6th harmonic.
This Audio Precision has a wonderful sine wave generator.

.
And now for something I built myself, I have more versions but the other one which has an even lower THD is just quick to connect, this one is not quite finished.
What I show below is a Wienbrug oscilator based on a design by Jim Williams and consists of two opamps.
The second opamp helps keep the commonmode error as small as possible.
Both opamps have been buffered, because of this they do not see any load and therefore deliver their maximum performance.
The buffer ICs are the LH0002(under the heatsinks) and the dual opamp is the OPA1612.
High quality selected parts have been used here, good quality is not free and do not forget about the thick steel housing as shielding.

.
Low spurs are only possible with battery power or an analog power supply with very few commonmode problems.
Maybe nowadays it is also possible with the Switcher operating at tens of MHz, however, I have never tested this.

.
And this is what it looks like in the spectrum measurement, there is nothing at all to see the grid frequency and low distortion.
This is the best I have been able to make with a simple Wienbrug circuit.
Better is possible as the Audio Precision shows, but then with a different construction of the sine generator. and again, with very good shielding!

.
I hope this is enough comparison material to make trade-offs between modern function generators and analog frequency generation.
The spurs and distortion products of modern function generators can be made a lot less by first using a bandpass filter and then a lowpass filter.
This only for a limited frequency range, otherwise it becomes quite extensive.
What, do I hear something?

OK, OK herewith a picture of the generator that is in the Quant Asylum.
The third harmonic is the strongest, but still within specifications which are better than -115dB.

.
Best regards,
Bram