Re: [Flightgear-devel] Adding engine sounds to AI Aircraft

[Ron H. <rc...@gm...>]

Thanks for the reply, James.

Regarding performance, I've been testing at KSFO in the "morning"... and
there are 1 to 2 hundred or more AI craft, with many departures and
arrivals in the first hour.  I don't see any added performance hit for
adding the sounds, and they all seem to play when intended.

1. Adding <sound> to the aircraft-set.xml or -main.xml is the approved
technique as I understand it, similar to how marker lights were added to AI
aircraft.  My patch is really just proof of concept, but I would love a
mentor to help me make it an official feature.  Using something better than
"airspeed" property to trigger the sounds and shape the volume profile
would be nice, like a "thrust" property, but that's more work for someone,
I understand.

2. Zombie sounds: Yes, it's "easy" to recreate if I trigger the sounds near
speed of zero, when the sound is still playing as the aircraft "parks".
Having them trigger above 10kts makes it happen much less often.  I know
C++ and have compiled 'next', so I can try to find a reliable reproduction,
but it will surely happen eventually if the trigger speed is 0 instead of
10.5 kts.

3. Regarding strange glideslope artifacts on 'next'.... umm.....
*nevermind*.  It turns out that messed up and did "git clone" of the GIT
LAB MIRRORS, thinking they were up-to-date, but... they are NOT.  Today I
did a valid git clone of "next" from sourceforge, and the AI Traffic is
behaving much better.   There is some strange taxi behavior for Arrivals at
KSFO (they veer off to taxi "too soon", but everything else is looking
reasonable for now.

Again: Would love a mentor on this, and I will continue to update this
thread if I discover anything new.

--Ron

On Fri, Jun 23, 2023 at 1:26 AM James Turner <ja...@fl...> wrote:

> Hi Ron, welcome and thanks for sharing this
>
> On 21 Jun 2023, at 20:58, Ron Horn <rc...@gm...> wrote:
>
> So I created a proof-of-concept (my current git patch here
> <https://drive.google.com/file/d/14vggup8MZcdygtw1uT18qpy-pXT8KSpn/view?usp=drive_link>).
> There are a lot of comments in generic-airliner-exhaust.xml to explain it.
>
> I tested this heavily on 2020.3.18 and also patched and built on 'next' a
> few days ago - no crashes or disasters.  (Using Windows 10, by the way).
>
> Being new, I have more questions than answers, but here are a few:
>
>    1. What can I do to help this possibly become part of a future
>    FlightGear release?  I've read most of the fg wiki pages I could find on
>    contributing, but still need some guidance on how to proceed.
>    2. I found a strange "zombie" sound issue that I worked around by
>    making sure I silence the sounds BEFORE the AI aircraft parks at the gate
>    after arrival (sometimes gets stuck at 10 kts).  Feels like something is
>    wrong in how "class FGAIAircraft" and related code shuts down the sound fx
>    when parking.  I checked the bug list and did not see anything like this, I
>    assume I can/should file a new ticket once I have good steps to recreate it
>    (not hard).
>    3. Watching AI Traffic on 2020.3.18 is very stable (I let it run
>    overnight, and only observed issues that are documented, such as gridlock,
>    etc).  But on 'next', I see some very strange problems with arriving
>    aircraft: wrong glideslope on final approach, crazy pitch attitude as well
>    during landing.  I'm happy to file a ticket for this too, if it is not
>    already a known issue (I searched and did not see it, but I'm not very
>    familiar with the bug list yet).
>
> Thanks for reading this far!
>
>
>
> For 1, we need to some planning indeed. I can see you current patch needs
> to modify each AI aircraft,, we need to decide if that’s the right approach
> (probably it is, so turboprops can have different sounds, etc). But in
> principle I don’t see an issue here, although my knowledge of the sound
> code is very basic, hopefully Erik who knows it better can also give some
> feedback.
>
> I’d also wonder what happens (CPU load, etc) at a busy airport with many
> active AI aircraft taxiing, I’ve no idea how the XML sound system handles
> this, maybe you already tried at EHAM or KORD and the answer is it’s just
> fine? Sometimes sound drivers have a limit on the number of active sound
> sources, to give one example.
>
> For 2, there could very easily be a bug here about deleting or cleaning up
> the corresponding FGFX object, would need a simple test case if we can make
> one.
>
> For 3, Keith paterson currently works on the traffic code; there are
> plenty of strange behaviours in the 2020.3 code. The code on next has had
> some particular areas improved, but I don’t know the overall state. Some
> tickets on specific issues you see would probably be good, especially if
> they occur consistently.
>
> Kind regards,
> James
>
> _______________________________________________
> Flightgear-devel mailing list
> Fli...@li...
> https://lists.sourceforge.net/lists/listinfo/flightgear-devel
>