New estim tracks for CH videos

[darthjj]

qDot is working on adding UI controls to Intiface Central for configuring custom devices. Has been taking a while, though.

I think you can integrate all the audio generation math from restim (restim/stim_math) into your application quite easily, you will have to rewrite the UI components and driver interface. I think it's easier (less dev time) to just connect to restim over websocket. I don't think it makes any sense to maintain two separate GUI's for configuring estim signals.

Indeed, it wouldn't quite be worth it if I would have to reimplement the calibration UI. For me it's definitely not a problem running multiple programs and figuring out how to get them to talk to each other, but writing the previous post with the instructions made me think a little about user-friendliness, and making it easier for less computer savvy people.

I'll definitely do the websocket connection first, but I've been distracted by Diablo 4 and haven't had much progress this week

But back to the topic of restim.. I've noticed that I have to turn up the volume on the stim box pretty high to get a nice balance between left, right and neutral. Is one of the channels always using 100% of the amplitude, and the volume still needs to be this high, or is it possible that there is room boost the waveforms in software? Had a bit of a unpleasant experience when a web browser started playing sound on the wrong audio output :(

[diglet]

But back to the topic of restim.. I've noticed that I have to turn up the volume on the stim box pretty high to get a nice balance between left, right and neutral. Is one of the channels always using 100% of the amplitude, and the volume still needs to be this high, or is it possible that there is room boost the waveforms in software? Had a bit of a unpleasant experience when a web browser started playing sound on the wrong audio output :(

There is room to boost the waveforms in software, but it depends on the calibration parameters. Currently it scales the output such that neither the left, right or center channel clip. I may change it so it only checks the left and right channel, I just checked in audacity that the left/right channel only output at max 50% volume when the calibration parameter for neutral is at -6. It does output at 100% if all calibration parameters are zero or positive.


I concluded earlier that this was a bad idea because it could lead to unexpected changes in volume, but on closer inspection that problem already exists and this change probably doesn't make it any worse. I will investigate.

I usually use WDM-KS as the audio api. This seems to claim exclusive access to the device, which prevents other programs from outputting any audio on the same device.

[diglet]

I made a new topic so new users don't have to shift through 15 pages to get up to speed, 10 of which contain technical ramblings or outdated information. Please continue the discussion there

viewtopic.php?t=26244

[senorgif2]

Hey Fileflax, have you, per chance, perused the excellent Mistress and Control box by phx? I only mention this because the author uses a sound generation software called SuperCollider to create what is, in my opinion, the pinnacle of e-stim signals, thus far. While I know my way around a computer, the complexities of using that program to create something is beyond me. You, however, seem a fair bit more skilled in that department than I am (considering your initial post was about creating your own program). It would be great to see what you could come up with if you can figure out the program. If you are still active, of course.

[edger477]

Hey Fileflax, have you, per chance, perused the excellent Mistress and Control box by phx? I only mention this because the author uses a sound generation software called SuperCollider to create what is, in my opinion, the pinnacle of e-stim signals, thus far. While I know my way around a computer, the complexities of using that program to create something is beyond me. You, however, seem a fair bit more skilled in that department than I am (considering your initial post was about creating your own program). It would be great to see what you could come up with if you can figure out the program. If you are still active, of course.

the supercollider program is based on javascript, and you do not need to know everything. I think you can skip everything up to the point of ~scriptIntro variable from where you can probably figure out what is being done, and then review ~setup which is pretty cool way of adding dynamic events that change current signal in certain way, whatever that current signal is. These predefined routines are then used in ~scriptCh1 (it should be readable easy enough to see how it is used), and you have similar pattern for other chapters.

[phx]

the supercollider program is based on javascript, and you do not need to know everything.

The syntax is somewhat familiar if you know some C-style language. The semantics are more inspired by SmallTalk and Lisp. Javascript is inspired by Self and Scheme which is similar, but there's enough of a difference that the semantics are pretty unique in between the two languages.

For my next project, I'll probably lean more into event triggering rather than low-level scripting a lot of the controls. I think this is eventually going to be a better setup, where a lot of the stuff is far easier to control. As a simple example, I have two functions I call for an edge. One to start, and one to stop. But it would be far better to just have a single function call with a start and a duration. Then a routine can sleep for the duration and automatically stop.

My plan for events is to use (span) markers in the NLE and then export an EDL. Then compile this EDL into SC-events. A marker with a span has a duration, so things will be relatively easy.

I haven't begun on this at all yet, though. Mostly focusing on the video-aspect right now.

[senorgif2]

I found some tutorials for super collider on YouTube that I will start watching if I continue to be furloughed. I am pretty useless without a GUI; typing code into a box is very un-intuitive to me.