Bus Extract prim problem: Not updating channel name..

[Spogg]

Bus extractor bug.fsm

This may not be a bug but it's bloody annoying!

(7.74 KiB) Downloaded 891 times

Hi guys!

I’ve found an “issue” with the Bus Extract prim. The attached FSM shows the problem well.

The prim doesn’t update its channel name from the selector as I would expect, unless you click on the S input pin. The selector IS sending a trigger, as shown.

If you type into a String prim, the channel name updates, once you’ve finished editing the string (also shown in the FSM).

This came up with the modulation matrix I’ve made, when I want to clear all the sources and targets. The clearing works fine, the indexes go to zero and the popup prim outputs the correct string, but the Bus Extract prim keeps the old settings, unless you click the S input.


It seems to me that the schematic has to re-compile with a changed string for the Bus Extract prim to update correctly, but I can’t see a way to force a recompile by any other mechanism. Clear Audio prim doesn’t do it and there’s no Re-compile prim available.

Or it could be that I’m missing something fundamental about “string theory”.

EDIT: I should add that the behaviour is the same in the latest 3.09 test release.

Please help an ageing Vulcan!

Cheers

Spogg

[ChrisHooker]

Hi Spogg,

It appears to be an issue limited to the redraw of the prim only. While it does not show as updating the channel, if you connect a bus, and extract data, the output does indeed show the correct value. (At least on my computer.)
Attached is a modified schematic that functions correctly for me, even though the prim's display issue still exists.

[Spogg]

Chris, you are absolutely correct and I thank you for taking the time to respond.

I’ve discovered that the behaviour I’m experiencing only seems to affect streams. With the floats in your example it’s fine.

I also made another discovery. If you change the signal type connected to the Bus Extract output, it changes the input signal type on the Bus Create prim, when it’s selected. This means for sure that the system re-compiles to update the schematic. It kinda makes sense now but I feel the stream handling isn’t ideal.

I do recall Maik saying he corrected the Bus system handling of streams while developing 3.09, but I assumed that was just a bug fix with the new code and not an inherent problem from before 3.09.

In any case, I’ve re-worked my project to use a selector-based matrix now. This should also have the advantage of not killing sounding notes with loud clicks when changing routings. I hope: I haven’t tested it yet as I write this!


Thanks again

Spogg

[ChrisHooker]

Hey again!

Regarding the stream connector type, you may just need to have the audio engine running (turn on a Direct Out / ASIO Out) in order to see the update. I tried placing a Mono Readout module on the output of the Bus Extract, in order to see an update on the float values at the input, and they didn't change without turning the audio engine on. Once on, though, the streams are running, and "looking" at what is being fed to them, so all is successful.

Attached is another example schematic where (when the audio out is turned on) it does update the Bus Extract stream output to the selected channel (waveforms in this case, so you can listen to clicking on switching).

I've also included a selector as an alternative to check the clicks/pops on switching, and unfortunately, it seems to do so as much as the bus selector. (I'm guessing the same switching method is used inside the bus extract prim as the selector.)

[Spogg]

Well that’s both fascinating and frustrating! I agree it works fine in your example and thanks again for your input.

In my Bus-based matrix I used poly streams exclusively and it was a large array with 22 sources and 16 targets. Each of the 16 routing strips had the option of choosing one additional source of 22 to multiply, add or subtract.

The problem I described is repeatable; the Bus Extract not only doesn’t update its channel label but also it doesn’t update its selection. Maybe it’s specific to my method of resetting all the preset modules to zero in order to clear the routings all to “no source”. But this method does give the index of zero and the correct string. It’s just ignored by the Bus extract prim, in my big schematic.

Anyways, I’ve now successfully made an alternative version using wireless connections, selectors and multiplexors, tested it, and this does behave as per my plan and expectation.

So again, thanks for your interest and help.

Cheers

Spogg

[tiffy]

Bus extractor bug.fsm

Hi guys!

I’ve found an “issue” with the Bus Extract prim. The attached FSM shows the problem well.

The prim doesn’t update its channel name from the selector as I would expect, unless you click on the S input pin. The selector IS sending a trigger, as shown.

If you type into a String prim, the channel name updates, once you’ve finished editing the string (also shown in the FSM).

This came up with the modulation matrix I’ve made, when I want to clear all the sources and targets. The clearing works fine, the indexes go to zero and the popup prim outputs the correct string, but the Bus Extract prim keeps the old settings, unless you click the S input.

It seems to me that the schematic has to re-compile with a changed string for the Bus Extract prim to update correctly, but I can’t see a way to force a recompile by any other mechanism. Clear Audio prim doesn’t do it and there’s no Re-compile prim available.

Or it could be that I’m missing something fundamental about “string theory”.

EDIT: I should add that the behaviour is the same in the latest 3.09 test release.

Please help an ageing Vulcan!
Cheers

Spogg

Hi, Spogg. I tested it here on my side as well and I do not find any difference when using your setup when the Bus-Extract Primitive's Channel Names Input are being fed via a Selector or via a String Primitive - in both cases the Channel Names does not update immediately until you click with the mouse on the Bus-Extract Primitive's Channel Names Input Terminal OR somewhere else on an empty space in the schematic. It thus appears that this issue was there all the time, it is just that no-one noticed it until you did.

However, the Bus-Extract Primitive still immediately extract the correct Bus Data from the Bus-Create Primitive. Since in an exported schematic the "Channel Names" are not visible it should not be a major issue...I know it can be irritating at times though when testing schematics etc.

Just thought I bring it to your attention

Regards.

[Spogg]

Guys!

Thanks for your help and interest.

I finally fixed the issue in my schematic. I added an additional trigger from the Reset button to the S input of the Bus Extract. Since it was connected after I connected the original preset reset it means the preset index is set to zero then the bus extract is triggered immediately afterwards. The Bus Extract then updates its routing correctly (but not the displayed Channel name -which is of no consequence).

From a technical aspect it seems maybe this is a timing issue with triggering in my huge schematic. I actually don’t see why I have to send a trigger when the string changes, but hey, whatever.

I shall stick to my new selector/multiplexor matrix system though, because it’s much faster and there is no re-compile, so any sounding notes aren’t killed. You just get a few little clicks which I can live with.


Once again, many thanks.

Spogg

[ChrisHooker]

Glad you got it worked out. I have no idea how Poly streams work, so I couldn't be of any assistance on that aspect. I personally have never used them in any projects (haven't really done much programming for MIDI-generated voicing, which I guess is where they're used?).

One thing I have noticed is that you can't connect the poly output of one module/primitive to multiple poly inputs, nor can you drag-move an existing connection to change from one module's poly input to a different module's input. You must first disconnect the destination input, then you can connect to the different one. No other connection type seems to have this limitation (that I know of). Perhaps this is the root of the bussing bug. (The trigger on the Bus-Extract's S-input somehow resets the links?)

Anyway, cheers and happy holidays!

[Spogg]

Glad you got it worked out. I have no idea how Poly streams work, so I couldn't be of any assistance on that aspect. I personally have never used them in any projects (haven't really done much programming for MIDI-generated voicing, which I guess is where they're used?).

Yes indeed! Polyphonic streams are used in synthesisers where more than 1 sound channel will be required. Each new note played creates a complete copy of the poly circuit in parallel and is closed when the last envelope is finished. If there is no envelope controlling the channel then it will close abruptly on note release (not much use in practice).

One thing I have noticed is that you can't connect the poly output of one module/primitive to multiple poly inputs, nor can you drag-move an existing connection to change from one module's poly input to a different module's input. You must first disconnect the destination input, then you can connect to the different one. No other connection type seems to have this limitation (that I know of). Perhaps this is the root of the bussing bug. (The trigger on the Bus-Extract's S-input somehow resets the links?)

I'm puzzled by this finding; I've not had that. I can connect a poly LFO to several targets without issue. What version of FS are you using?

Anyway, cheers and happy holidays!

You too mate.

Cheers

Spogg

[ChrisHooker]

Well how 'bout that! lol (This is what happens when I try to open my mouth about things I know little about.) Looks like my issue is only happening when trying to switch from a Poly-to-Mono input to an Analyzer with Poly input. (Or if you try to connect to both.) (Schematic attached)
This may not even be practical or sensible signal flow, like I mentioned: when it comes to Poly, I have no experience... I was just trying to throw something together to test the bus issue with poly connections.

Edit: Oh, and I'm using FS 3.08.1