Not quite correct to say that the replay of stereo samples requires double of everything.
Stereo only requires two voices (two oscillators or two notes of polyphony if you like), which would then feed through a single filter and amp.
If you have a stereo source that you then convert to mono to send through a single amp and filterā¦ thatās no longer stereoā¦
2 Likes
If you convert a stereo source to mono then it will use only one voice (unless played polyphonically)!
Huh? A mono sample is no longer a stereo sample. What are you trying to say? Stereo samples would require a doubling of the entire path, just as JPM has said.
2 Likes
A single filter that process two channel is like 2 filters with same parameter. As for amplifier. When working with stereo, itās 2x everything.
I agree a mono sample is no longer a stereo sample.
I think my comment is probably too deep in the weeds of sampling (a multichannel sample data is interleaved and any further summing/processing is just a real time calculation that creates a new sample) so I should probably not have responded.
In more practical terms, many Roland rompler synths allow you to create your own drum kits. You can choose to have a mono or stereo percussion sound. The way you do this is simply select a sample for the left channel only for mono or one sample for the left channel and another sample for the right channel for stereo. Regardless of mono or stereo everything else in the sound chain is the same.
Itās not the same.
If I have 44khz sample in mono, I need to run each sample through a process (lets say a filter function).
Each number gets processed by the filter function at a rate of 44000 times per second.
A 44khz stereo sample is actually 88000 numbers. 44000 for each channel left and right.
If I want to maintain the 44khz sample rate, I need to have a filter function running on the each channel.
If we want to use a single digital filter, you MIGHT be able to have a filter which operates at 88khz (double) and do some clever interleaving, but as filter action works in the time domain, youād then also need to store the previous sample for each channel in a buffer, then load the previous channelās sample, do a calculation, then store that, then switch back. This would be even MORE processing.
Again, stereo samples IS double everything. If you want to sum the stereo sample to mono and then process it with a single 44khz filter, you will then just have a mono signal.
If they truly have pushed DT to itās limit, the only solution for DT would be to forfeit 2 audio channels for a single stereo āvoiceā.
4 Likes
Also, correct me if Iām wrong, unless there are some differences on each side, the brain will hear two identical signals (leaving phase issues aside for the moment) as mono right? Hence why chorus FX sounds wider because of the tiny delay to one side or the other, correct? I think thatās right 
1 Like
Yup! The perception of stereo is simply hearing differences between the left and right ear 
Thatās why even passing a single osc/sound source through stereo (analog) filter can result in the perception of width. (depending on the quality of the filter this could be very subtle or very obvious)
2 Likes
You need to remember that digital audio data is always serial. So a mono sample is transmitted serially and a stereo sample is also transmitted serially. So as there is only ever a single line of data you donāt need two of anything to further process the data.
In the case of a stereo sample one byte of left channel audio data will be sent through the āfilterā followed by one byte of right channel audio data through the same single āfilterā and this cycle will continue for the length of the sample.
Also, sample rate values do not equate to sample processing speeds, which are likely to be much greater.
Using a simple time delay is one of the best ways to show that.
Delay haas its uses
4 Likes
Ess said something about this design choice in the Q&A at the end of the Model:Cycles launch party.
(Edit: at 1:10:10.)
1 Like
Just checked: itās at 1:10:10
1 Like
he says due to audio quality / hardware decisionsā¦
no need to make it ALL stereo.
it would be really nice already if they could add at least 1 stereo track, even in lower audio quality ā¦
3 Likes
As for stereo processing, Iām wondering if it could be possible to sync two tracks so every tweaked parameter (except pan) could be applied to the other track.
1 Like
3 Likes
Apropos of this, the last major module I bought was the ADDAC 112 VC Looper and Granular Sampler, which was originally announced as having mono input, stereo output. But the release was delayed by a few months because the developers decided to make it stereo input as well (that is, all recordings would be stereo). The module was plagued with crashes, glitches, and CPU overloading for the next eighteen months, despite many firmware updates. I only decided to buy when it appeared that the stability and reliability goals one would expect of an initial release had in fact finally been met. (Pretty happy with it so far, for what itās worth.)
Too bad this feature was never implemented. The humble Korg Electribe Sampler got this feature right; playback of stereo samples just ends up taking 2 pads and the Electribe Sampler automatically does this for you.
āIt should be simple to codeā imo.