Yes! So after writing the last project log (and therefore kind of rubber ducking it) I made the changes I proposed over there:

I made sure I used the correct timings and was connected to the remote input. As the timings are pretty slow for a uC (at least 800us or about 1200 Hz) between setting outputs), I just did it in C++ instead of the PIO-program on the Pico. That way, I better knew what I was doing and was not dependent on my still developing ASM skill. I still want to program it in the PIO later on though.

Next I tried it out, commented out some things I forgot about and tried again and it just worked!

A satisfying muting/unmuting of the amp was heard! Muting was my initial test case, since I would be able to hear a relay in the amp clicking from where my desk is.

But also other commands as setting input source worked flawlessly. 

On volume control

Volume needs some fine tuning though. Sending the volume up/down command works perfectly, but there's some mechanics in play: this amplifier has a physical, motorised volume knob connected to a double 50k-potmeter. So, when you turn the volume up and down by hand, you rotate the potmeter, when you do it by remote, the amp enables a small motor to rotate the volume knob. Which you can also see if you use the remote and which is extremely satisfying, but there is some mechanical inertia.

Quickly a step back in the RC-5 protocol:

The 3rd bit of the protocol is a Control bit. This one should change on each button press. So for example for the volume up key, the binary string would be:

• b11 0100 0001 0000 on the 1st press
• b11 1100 0001 0000  on the 2nd press
• b11 0100 0001 0000 <pause> b11 0100 0001 0000 <pause> b11 0100 0001 0000 on a 3rd press, combined with holding it down. It'll just repeat until you let go.

This way, the receiver can differentiate between different presses or a single press that might have been interrupted by something blocking the signal.

For volume up, that means that with the mechanical inertia, sending 10x a single command of "volume up" (so, with intervals and switching control bits) is not the same as sending a single command of volume up, but repeated 10 times (like a button being held down). The end position of the volume knob will not be the same (given an identical start position).

For now, I've set it to sending out the message 2 times, which gives me sufficiently small control. All other messages are sent only once as these are 'boolean' messages like (un)mute and setting a source.

Remote control circuit

In the last log, I mentioned assuming that the signals that should be sent over the remote connections should follow the RC-5 protocol. This wasn't a random assumption, but was actually based on the schematic of the amp.

In the schematic above, I've highlighted some nets for clarity.

Let's start at the end. The incoming RC-5 signal (purple) connects to the RXRC input of the the control chip (refdes 7401). This active high signal can be pulled to GND (black) by BC547 (7414) on the bottom right.
This transistor 7414 is activated by either the RC-5 input signal (green) through 47k resistor 3462. That is the connection straight from the remote input connector 1401.
Or, transistor 7414 is activated by enabling the BC557B 7412. This one is held high from +4.9V (red) through 10k resistor 3459, but can be switched by the IR receiver 7403 through 10k resistor 3461.

As also shown in the last log, the IR receiver 7403 has an internal pull up to Vcc (+4.9V), when a signal is received it will pull its output to GND.

Lastly, there is a ENTX signal (yellow) that switches both transistors 7409 and 7410, which would enable transistor 7413 and disable transistor 7412 and 7414. This keeps the RXRC signal high and would probably allow the TXRC (light purple) to become active and switch transistor 7411. I don't know when that one becomes active though, as the IR receiver also pulls the RC-5 signal (green, to remote out) to Vcc already. So I assume that the chip 7401 doesn't repeat to TX  the signals it receives on RX.

Pico side of things

On the Pico, actually I did very little. I breadboarded it and soldered a long 2 wire cable to 2 header pins with a diode in between, to protect the Pico as the GPIO's are 3V3 based. Also means that, through the diode D2 below, the transistor 7414 in the Marantz seems to switch reliably with <3V. Which makes sense, as the collector of transistor 7414 is connected to Vcc through a 22k resistor, so the current through 7414 is <0.25mA when active. I'm still contemplating on tapping it from 5V on the Pico side though. But for now, I have a working setup!