They prevent low voltage noise from being amplified by the op amp, thereby keeping you from having to hear it.
For an LED to be on, the anode has to be at a higher voltage than the cathode. For two counterparalleled diodes, they will never truly be on at the same time. However, if the current passing through them is alternating at a high enough frequency, they will both appear to be on at the same time, as our brain is not fast enough to process the time the LEDs are off, i.e. stroboscopic effect. The same phenomenon occurs when you under sample an AC waveform, resulting in aliasing.
If you're thinking about that Ibanez TS-10 Tube Screamer thread, they are for intentional distortion.
1N914, 1N4001, and LEDs are used to change the operating point when distortion kicks in.
I'm having difficulty figuring out how they would reduce low amplitude noise. Probably because I'm thinking of an entirely different circuit than b1u3.
Thanks. Understand the importance of your request; but showing the schematic will give you the largest of headaches and pose a very difficult question on the board.
If you insist, I will post it.
What happened is I designed a circuit about ~10 years ago, built it, tested and worked to perfection, made me proud of myself. Needed its integrated circuit for something else, removed and now cannot remember much; barely what the circuit did, if that much. The design notes lost in two house moves. So it is an obsessive subject am trying to discern breaking it to pieces, and one of the pieces is in the original post.
Diodes in this configuration usually are there to limit (clip) whatever is being fed back to one of the op-amp inputs...I instantly thought of an amplitude-stabilized op-amp oscillator when I first read the post...
I believe audio from a RTTY receiver is fed to pins 8 and 10.
If the receiver is properly tuned, both pairs of leds light up. If tuned only to one tone, only one pair of leds comes on.
The potentiometer either tunes the tone shift of choice being received or one of the tones ... I think. Works on 12V. The headache is cannot remember the integrated circuit missing there. It is not a super common but it is not a oscure one either. The IC was removed to install somewhere else and the design data is lost in moving residence.
The circuit is not copied from anywhere, I came up with it, and worked very well.