-
Notifications
You must be signed in to change notification settings - Fork 4
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Audio scope filter shape disappears #105
Comments
It seems like it's not properly retrieving the current mode, so it doesn't know how to draw the scope edges. If you can reliably reproduce it to figure out what causes it, I'd greatly appreciate it! |
It seems to happen when I first switch to Memory mode from VFO mode, it's consistent when I switch back and forth but I can get the filter shape to appear again by refreshing the browser tab and from then on switching VFO and Memory mode doesn't seem to lose it. It would seem to be something that doesn't initialise on server start, I can stop the server, restart and essentially the same sequence occurs. |
I have just seen another effect of what seems a similar bug. I restarted the server and even in VFO mode when I started the radio UI I had no filter shape and no LC and HC. Changing to Memory then gave me a very narrow filter at extreme shift, so I think it was 200Hz at 2800Hz and it was showing width and shift although the radio itself showed no difference. Changing to VFO in the browser corrected this, and it stayed correct when back to Memory again. I managed to capture this, here it is showing the point at which the audio stream starts up: [info] JOINED radio:audio_stream in 10µs HTH |
Thanks! That should help quite a bit when I get around to looking at this. |
The same issue also happens when Disconnecting and then Reconnecting. Swapping VFOs fixes it immediately. |
… so we know where to put them. Refs #105
@Tyrbiter I think I've tracked this one down, and pushed a couple of fixes to |
Well it starts, but as soon as I open the browser tab for Radio UI I get errors, here is the full output from starting the server: mix phx.server open890 is now running. [info] Auto-starting connection id fa8467e1-de2d-4a33-9923-49c72ea79123, "TS890" Be watching in static Processing css/app.scss... Waiting for file changes... |
Hmm, this shouldn't have caused this error. It looks like maybe you got the connection id in the URL bar incorrect somehow, maybe accidentally deleted a character or something? Does it work when you navigate back to |
Right, restarted server and then closed Radio UI tab, then reloaded connections page and reopened Radio UI. Working again, some sort of browser caching thing? I now get the filter shape when stopping and restarting and when changing VFOs, however the filter shape is gone when I switch from V->M and back when I go from M->V. HTH. |
Great, I think I'm on the disappearing filter, then. I'll add a little bit of logic on the radio page, that if for some reason it can't find the requested connection ID, it will just redirect you back to the connection index screen, which should prevent the big nasty error you saw. |
Sounds good, I can see from looking at the commands that some clever tricks are needed to keep track of filter settings. Really appreciate the effort to fix this up. |
Could you tell me specifically what your VFO frequency/modes are, and likewise for the memory channel you're switching between? And whether either the VFO or memory channel are in SPLIT? I saw the disappearing filter briefly, but then managed to stop seeing it. |
VFOA is 20m USB, VFOB is 80m LSB, Mem channel is 60m USB, none set for SPLIT. |
Great, thanks! |
@Tyrbiter After a bit of a hiatus, I'm picking this back up. I've pushed a couple commits to develop that should help. I wasn't able to completely replicate the filter disappearing, but I did notice that the filter was displaying the "wrong" sideband in a couple cases. If you've got time and are able, would you mind checking out latest |
OK, I have just made a quick test, I only got time to do this just before I head for bed :) I still see the filter shape vanish as soon as I select Memory instead of VFO. Both LC and HC also vanish but the : that follows them is still there. Here is a complete dump of the console output, please let me know if you want any more information: [bdm@deangelis open890-develop]$ make up open890 is now running. [info] Auto-starting connection id fa8467e1-de2d-4a33-9923-49c72ea79123, "TS890" Be watching in static Processing css/app.scss... Waiting for file changes... |
And of course no sooner do I send the previous message and after I restarted the connection in the browser, now it doesn't happen and the filter shape stays put. More console output which follows on from above: [debug] Bandscope LV: RX connection_state: :starting |
Yeah, this is a weird one. In your first log, it looks like the radio is not sending the operating mode (the
Here it looks like you are switching from VFO to memory, and then back to VFO mode. But in this span, the radio doesn't send the operating mode, which is really annoying. But in your second log, it does:
That last I was hoping to squash this one for the next release, I'll keep messing with it a bit tomorrow, but if I can't reliably get it to work, it may just linger. |
I'm not expecting miracles, but it's clear that the radio responses do seem to vary depending on how it's setup and what span is set for each VFO. My test involved pressing only the M/V button. I'm going to assume that all TS-890s behave the exactly same way, but if you need me to test and send more debug please ask, it would be embarrassing for Kenwood to have to admit that the firmware is non-deterministic :) To avoid any doubt I have fw 1.04 |
Just noticed this: If TS-890 is in VFO mode when I connect in the browser tab then switching to Memory mode loses the filter passband shape and LC/HC annotations. If I disconnect while in Memory mode, then reconnect, the filter passband shape appears together with the LC and HC annotations but then switching to VFO mode (and swapping VFOs) loses them shape and annotations again. Here is a complete debug dump of this after running make up to get to the develop branch immediately after 1.0.4 in main: mix phx.server open890 is now running. [info] Auto-starting connection id fa8467e1-de2d-4a33-9923-49c72ea79123, "TS890" Be watching in static Processing css/app.scss... Waiting for file changes... |
Note for myself: I think the solution is to send |
That sounds like a plan :) |
That seems to work, not too much code changed too :) |
Woohoo! |
I have noticed a few instances of the disappearing filter shape and I think the LC and HC labels, the values are still shown
This gives long strings of this in the console window:
[debug] Unimplemented case for audio_scope_filter_edges for mode nil
It gives the impression of something being lost, possibly on swapping VFOs or maybe switching from V to M or the reverse.
Refreshing the browser window fixes it, is it maybe a caching issue?
The text was updated successfully, but these errors were encountered: