the 4.0 upgrades migrates the data (and its location) to the new 4.0 format. What you can do is to take a copy of the SD card before the upgrading. You can do that by starting a bootloader, enabling the USB storage and copying all the files to your computer. Then, when you want to downgrade follow these steps:
downgrade to 3.6.1
open bootloader / enable USB storage
copy the 3.6.1 files back
restart
for reference there is a pristine clean SD card data for v3.6.1 and v4.0.0:
I have loaded 4.0.0g, and tested my VCVRack2 preset (publicly available in the library). Unfortunately I have experienced freezes and hangs where the normal reset didnt work (loading animation completed but my preset didnt start - see log below) and I had to use the hold top left button power cycle to recover. Is there a change related to presets loading saved state ? The E1 froze after seemingly loading preset data. Maybe you could try opening the VCvRack2 preset, go to page 11, click the Save to E1 button then restart the E1 - see if you see the same freeze.
I will also dig further when I have time.
Damn Martin, youâre building a whole new interface layer. That must have been a ton of work. Looks smooth in the video!
As for the Sysex issue, Iâve been holding off since I figured youâve been very busy, but I have had a fully finished new project Iâve wanted to discuss with you since Januari, which replicates this issue. Let me know if youâre interested, even if just for testing the issue.
Hi @martin . On 4.0.0g, it seems that toggle buttons are not toggling from on â off. Off â on seems OK, but once on, the toggle button does not change state back to off when pressed again. I have seen this both on my VCVRack2 preset, and the library Repro-5 preset for example.
Hi @martin More digging on the E1 startup freeze I am seeing. I have narrowed it down to calling pages.display(1) from within a preset.onReady() callback, in my VCVRack2 preset when it is opened automatically when the E1 is started.
When I commented out the pages.display(1) statement, the E1 restarts OK. It seems to be repeatable.
If I put the statement back in again, change to another simple preset then restart E1, then re-open the VCVrack2 preset, the pages.display(1) works. So it does seem this is something to do with when the E1 loads the preset on startup as far as I can tell.
This is new buggy behaviour in 4.0.0g compared with 4.0.0c
Hi @jhh , Iâm using a 14inch 2021 Macbook M1 Pro. I have not yet tested with the new just released version, my last test was with the 4.0.0c firmware.
I do not have the slow upload issue with the current (3.0) firmware. I am sometimes having some instability where after upload the next uploads are ignored or the Electra freezes. Iâve been working around this by throttling uploads. This is not too common though, and may have been fixed with the work @martin is doing now. However due to the slow uploads I have not been able to test that yet.
tackled onReady, there was a race condition. The VCV rack 2 + Orestes is starting okay for me. I will still be looking at it though.
the mysterious slow sysex understood after working together with @jhh. It was caused by forwarding of (fast) USB data to (slow) MIDI IO.
Please re-test this version if you have time for that. I tested with it with Ableton, Bitwig, VCV. Seemed to work just fine.
Originally posted firmware-v4.0.0h.srec.zip replaced with firmware-v4.0.0i.srec.zip (2.0 MB). Extra work on the race condition (better feelings here now)
HI @martin Thanks very much for the quick turnaround for 4.0.0i. Installed and re-tested my VCVRack2 preset , looks much better. It seems to have resolved all the issues I reported on 4.0.0g. Toggle buttons now switch between on and off OK. The pages.display(1) works within the onReady callback.
Interestingly, the rendering or processing of a sequence of sysex controlUpdate commands sent into the device (from my OrestesOne VCVRack module when you select a module) seems noticeable slower since 4.0.0c - I can see the controls being updated top-down, left to right - its still fast enough I think, but I never even noticed before.
Looks like the onPotTouch issue I reported previously has re-appeared since 4.0.0c
I would appreciate if you could share some info on the new onPotTouchChange callback as I asked here in January. When I try this code, the callback never seems to be triggered (this happened in 4.0.0c, and still seems to happen in 4.0.0i):
function events.onPotTouchChange(potId, controlId, touched)
log("potId: " .. potId .. ", controlId: " .. controlId .. ", touched: " .. (touched and "yes" or "no"))
executePotTouchActions(potId, controlId, touched)
end
Funny. For me the toggle buttons work less well than in version 3.x of the firmware. They respond erratically: sometimes not toggling. Sometimes toggling very quickly off and then on again.
Also, assigning âtoggle sensitivityâ to a button doesnât seem to make much difference. Also it is unclear how it works: is the sensitivity increased only while the button is pressed, or is it a real toggle?
Third, for lists controls, if you are on the first item and turn further left, the value jumps to the last one in the list. Is this configurable somehow? It is nice in certain cases, but Iâm not sure I always like this as the standard behaviour.
Also I see erratic ADSR behaviour. When the popup opens, all 4 pots control the first parameter, i.e. Attack. Value updates (sent using sysex command 0x14, 0x0E) also behave weird. The redraw for the first control is larger and above the actual control:
I noticed the wrong assignment of the on-screen knobs to physical knobs already. I have it on my todo list. The broken graphics is weird. I will look at that.
The sensitivity of the physical knobs is decreased while the button is pressed. It goes back to normal when released.
it is kind of a probe on what users think about that. I do not know, till this very moment, if I like it or not. Making it a âconfiguration optionâ seems to be little bit too much.
Hm⌠It seems the sensitivity toggle does not make much difference really (which is why I was unsure how it worked). Also, some graphical feedback (in the info line) perhaps about it being active or not might be useful.