Multiple: Remove RPT_TELECONF and Use ast_audiohook_volume* for ducking #563

mkmer · 2025-03-16T14:27:08Z

Use ast_audiohook_volume*
Remove telemchannel and btelemchannel.
Adjusts volume for ALL telemetry messages.

InterLinked1

First of all, I'm really excited about changes like this! The more code we can remove, the better! (without changing functionality, of course)

I think the commit message needs much more detail. It should explain why the old way was unnecessary and we can use audiohooks instead. For example, I thought these channels had something to do with telemetry, not adjusting the volume. Was that a misnomer? Or something else?

Also, I think for a change of this magnitude (removing channels), we should insist on tests. I would like to see a relevant test that passes prior to this change, and continues to pass afterwards. Once #533 merges, the tests should all be run by the CI.

mkmer · 2025-03-16T20:54:48Z

All that happens between telechannel and btelechannel is the ducking/volume adjustment code... That's it.
Every time we spin up a telemetry thread, we create a new DAHDI pseudo channel and link it either to the RPT_TELECONF or the RPT_CONF. RPT_CONF is connected to all audio moving around the system.

Also in the RPT_TELECONF is the telechannel. We simply copy telechannel to bchannel and in the middle we apply one of the two volume adjustments. Finally, the bchannel is connected to the RPT_TXCONF (which is local audio out)

So - all this does is move local audio connection in the telemetry thread FROM RPT_TELECONF to RPT_TXCONF.

Additionally, telemetry audio that was NOT connect to RPT_TELECONF was not receiving the p.telemduckgain or p.telemnomgain) adjustments.

InterLinked1 · 2025-03-16T21:49:53Z

All that happens between telechannel and btelechannel is the ducking/volume adjustment code... That's it. Every time we spin up a telemetry thread, we create a new DAHDI pseudo channel and link it either to the RPT_TELECONF or the RPT_CONF. RPT_CONF is connected to all audio moving around the system.

Also in the RPT_TELECONF is the telechannel. We simply copy telechannel to bchannel and in the middle we apply one of the two volume adjustments. Finally, the bchannel is connected to the RPT_TXCONF (which is local audio out)

So - all this does is move local audio connection in the telemetry thread FROM RPT_TELECONF to RPT_TXCONF.

Additionally, telemetry audio that was NOT connect to RPT_TELECONF was not receiving the p.telemduckgain or p.telemnomgain) adjustments.

Thanks for the explanation! If you can capture the gist of that in the commit message for posterity, I think that would be good.

I've seen some strange things in this codebase, but yeah, that sounds awfully inefficient. My guess, without confirming, is that audiohooks didn't exist in Asterisk when Jim wrote the code initially, and he may have deemed that an easy way to do that. But now that there are a lot of tools in the Asterisk core at our disposal, it makes sense to leverage those as much as we can.

Remove telemchannel and btelemchannel. Adjusts volume for ALL telemetry messages. Turns out this is a VERY complicated method of changing the volume on telemetry messages. Original design: telechannel is created and joined RPT_TELECONF. As telemetry channels are created the are joined to RPT_TELECONF for local, or RPT_CONF for everywhere. Audio on teh RPT_TELECONF is pulled out via telechannel, scaled via p.telemnomgain when not keyed and p.telemduckgain when keyed, then copied to btelechannel. Finally, btelechannel is in the RPT_TXCONF delivering the audio to the repeater. With this PR, we remove RPT_TELECONF, telechannel, btelechannel and associated code. We now link new telemetry pseudo channels to RPT_TXCONF in place of RPT_TELECONF and adjust the volume using ast_audiohook_volume_adjust().

Allan-N · 2025-03-17T00:36:20Z

apps/app_rpt/app_rpt.h

@@ -272,7 +272,7 @@ enum  rpt_dns_method {
 };

 #define DEFAULT_NODE_LOOKUP_METHOD LOOKUP_BOTH
-#define DEFAULT_TELEMDUCKDB "-9"
+#define DEFAULT_TELEMDUCKDB -9


Our rpt.conf file has :

telemduckdb = -15 ; Telemetry Ducking in dB when local or link voice tx in progress

It's OK for the two values to differ but it would be good for rpt.conf to note the default

-15 is a LONG way down... but I don't see why the shouldn't match

Allan-N · 2025-03-17T00:41:42Z

apps/app_rpt/rpt_config.c

-
-	val = ast_variable_retrieve(cfg, cat, "telemduckdb");
-	rpt_vars[n].p.telemduckgain = pow(10.0, atof(S_OR(val, DEFAULT_TELEMDUCKDB)) / 20.0);
+	RPT_CONFIG_VAR_INT_DEFAULT(telemnomgain, "telemnomdb", 1);


Changing a configuration file "float" value to an integer value?

... and our current rpt.conf file has :

telemnomdb = -3 ; Telemetry Nominal Amplitude reference in dB

Should this default also be the same?
and, it's OK for the two values to differ but it would be good for rpt.conf to note the default

I don't see why not

Allan-N · 2025-03-17T00:44:46Z

apps/app_rpt/rpt_config.c

-	val = ast_variable_retrieve(cfg, cat, "telemduckdb");
-	rpt_vars[n].p.telemduckgain = pow(10.0, atof(S_OR(val, DEFAULT_TELEMDUCKDB)) / 20.0);
+	RPT_CONFIG_VAR_INT_DEFAULT(telemnomgain, "telemnomdb", 1);
+	RPT_CONFIG_VAR_INT_DEFAULT(telemduckgain, "telemduckgain", DEFAULT_TELEMDUCKDB);


Changing the configuration file "telemduckdb" variable to "telemduckgain" :-(

Is there a way to retain the current variable and convert it to the other?

Allegedly, the asterisk volume want's dB, so we don't need to convert to the decimal value, just use dB. Of course if it's NOT dB on the volume, then we need to do something different.

Allan-N · 2025-03-17T00:52:30Z

apps/app_rpt/app_rpt.h

@@ -677,8 +677,8 @@ struct rpt {
 		const char *locallist[16];
 		int nlocallist;
 		char ctgroup[16];
-		float telemnomgain;		/*!< \brief nominal gain adjust for telemetry */
-		float telemduckgain;	/*!< \brief duck on busy gain adjust for telemetry */
+		int telemnomgain;  /*!< \brief nominal gain adjust for telemetry */


We are changing the variable types from "float" to "int" ?

We don't really need a float here... all the examples are int's and what kind of resolution do we really need on the gain for telemetry? (I get radio gains having decimal values to get things "perfect", but telemetry?)

Allan-N · 2025-03-17T00:54:51Z

apps/app_rpt.c

@@ -3659,7 +3613,7 @@ static inline int rxchannel_read(struct rpt *myrpt, const int lasttx)
 				myrpt->linkactivitytimer = 0;
 				myrpt->keyed = 1;
 				time(&myrpt->lastkeyedtime);
-				myrpt->keyposttimer = KEYPOSTSHORTTIME;
+				myrpt->keyposttimer = KEYPOSTSHORTTIME;	


Looks like we gained a trailing space

mkmer added the enhancement New feature or request label Mar 16, 2025

mkmer self-assigned this Mar 16, 2025

InterLinked1 reviewed Mar 16, 2025

View reviewed changes

mkmer force-pushed the use-audio-hook branch from 1326c43 to 935cee9 Compare March 16, 2025 22:52

Allan-N requested changes Mar 17, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Multiple: Remove RPT_TELECONF and Use ast_audiohook_volume* for ducking #563

Multiple: Remove RPT_TELECONF and Use ast_audiohook_volume* for ducking #563

mkmer commented Mar 16, 2025

InterLinked1 left a comment

mkmer commented Mar 16, 2025

InterLinked1 commented Mar 16, 2025

Allan-N Mar 17, 2025

mkmer Mar 17, 2025

Allan-N Mar 17, 2025

mkmer Mar 17, 2025

Allan-N Mar 17, 2025

mkmer Mar 17, 2025

Allan-N Mar 17, 2025

mkmer Mar 17, 2025

Allan-N Mar 17, 2025

Multiple: Remove RPT_TELECONF and Use ast_audiohook_volume* for ducking #563

Are you sure you want to change the base?

Multiple: Remove RPT_TELECONF and Use ast_audiohook_volume* for ducking #563

Conversation

mkmer commented Mar 16, 2025

InterLinked1 left a comment

Choose a reason for hiding this comment

mkmer commented Mar 16, 2025

InterLinked1 commented Mar 16, 2025

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment