Hi all,

I am new to FMOD Ex (and to sound API in general) and although the API is pretty easy to use, I have few problems to design the audio engine for the indie game I am working on.

I would like to use FMOD for a sort of real-time strategy game. I have hundreds of units on the battlefield and each type of unit hold different sounds (user feedbacks, bullet, explosion, etc…).

Each type of unit can have many instance of them. For example, 30 units of type medium tank, so I may have to play 30 times the same sound (few may be non audible). Basically, I don’t know what is the good way to use the channel.


1) Should I store an instance of each sound (FMOD channel) in each of my UnitType instance ? That mean may be thousands of channels ? Is it the good way ? so basically I call playSound with the paused parameter to true each time I create an instance of a unit ?

2) When I play a sound 1 time, what happen to the channel once the sound has finished ? If I play 1000 sounds, is there 1000 channel created ? after few minutes, is there 100000 channel allocated ?

3) How can I destroy a channel ? (when I not need it anymore)

4) I would like to fade from one music (streamed sound) to another, is there a ready-to-use function ? or should I have to do it myself with a timer and the setVolume function.

Thanks in advance if you can help me to use FMOD like it was designed to be used.

  • You must to post comments

Thank you for the fast reply brett :)

I understand better how FMOD works, but I am still a bit confused about the (1) and (3). (3) says that when sounds are not playing, the channel can be used again, so I cannot do what I said in (1). That mean “a channel is an instance of a [b:10z4xf9q]playing[/b:10z4xf9q] sound” ?

Q) So it is wrong to call playSound() with paused and to store the channel when a unit is created, and then call channel->play() each time the unit has to play a sound ? (because if I create another unit later, FMOD could returns the channel of another unit which is not playing a sound).

I must store a pointer (or handle) to the sound and call sound->playSound() each time the unit has to play a sound. Here I can get the channel and sets the 3D attributes but I don’t need to store it, right ?

Q) But for looping sounds (ambient sounds like forests, ocean, etc..), I can store the channel and I am sure that it will not be reused by another playSound() call ?

Q) About virtual voices. Do you only compute distances for the ‘playing’ sounds ? Does it help if I do not call playSound() when I am sure that the sound will be non-audible (the sound is far far away).

Thanks again for you help,

  • You must to post comments

😳 don’t know why I imagined a play() method in Channel 😕

Thanks again for the clarification brett, it is pretty clear now.

Have a nice day.

  • You must to post comments
Showing 2 results
Your Answer

Please first to submit.