Why is this sampler design mono, not stereo? - echolevel/open-amiga-sampler GitHub Wiki

Good question. Stereo seems like it would be a desirable feature, but our design is intended to be a practical and affordable one for people who still regularly use their Amiga to make music, or want to get into it. We both have an Amiga/PC demoscene background, and one of us has been composing Amiga music and performing it live for decades. There are various ways of using the Amiga to make music, and some people may need to record and play back samples in stereo, but for the most part the Amiga's polyphony restrictions (4 channels) mean that playing back stereo in MOD or MED music is a profligate waste of resources. Stereo takes up twice as much space on disk or in RAM, and the maximum sample rate is much lower than in mono. It's possible to play back stereo in Protracker MODs, but it involves splitting the signal and playing each side on a separate channel - not unheard of, but pretty rare.

The only area in which the Amiga wasn't G.O.A.T. was the unfortunate design decision to force two audio channels fully to the left, and two fully to the right - aka hard panning. Lots of YouTube retro gaming videos - and even some Amiga music videos - deliberately preserve this hard panning and the result, if you're listening on headphones or a proper stereo speaker configuration is usually appalling. Back in the day, we got around this either by plugging our Amigas into cheap televisions (which usually had single mono speakers or stereo speakers with such little separation on either side of a tiny CRT tube that the signals blended quite well) or by building small mixer circuits that reduced the panning separation to somewhere between 0 and 10%.

Until and even since Commodore's demise, a staggering amount of music has been produced on the Amiga and most respected composers in the demoscene, netlabel scene or games industry (the lines between these categories often blur) compose in mono in the expectation that their music will be listened to in mono. The groundbreaking netlabel Mono211, one of the first netlabels and later known as Monotonik after evolving from its Amiga roots, was thus named partly because their releases were intended to be heard in mono to avoid horrible situations like, for example, all of a drum and bass tune's drums being heard only through the left ear of a pair of headphones, and all the bass being heard only through the right.

So our sampler mixes down stereo input to mono, because it's what we want, and what most musicians want, and it's better. Some people may disagree. They're welcome to add a quad bilateral switch (or similar) to the circuit, and perhaps we'll publish some advice for how to get that working if we have time or there's lots of demand, but realistically stereo sampling on the Amiga has become an edge case; a niche feature.

How did most of the original cheap samplers handle mono sampling from a stereo input? Pretty simple: they only sampled the left channel. It's possible that some only sampled the right, if their resistors were wired up differently to the switch's control inputs, creating an alternative default channel. Remember: the data bus is only 8 bits wide, and this is an 8bit mono AD, so it has no means of digitally mixing the two signals internally and (if we want the highest sample rates possible) we don't want to sample from both channels and then mix in software on the Amiga. So we mix the analogue audio signal where it enters the sampler, with a few resistors to ensure that there's no voltage feedback or weird clipping/signal loss when L and R are merged. Different resistor values (or even both wired to a physical switch, if you're feeling adventurous) can be used to configure the input for headphone vs line level signals.