Now with Sound!
14/09/06 15:05
I've finally done what I consider to be a reasonable
implementation of 'capture system audio'. It works by
redirecting all system audio to the soundflower, and
then redirecting the output of that to whatever
device you had selected before. The later part means
that you can hear what's being captured.
I was going to write my own reflector driver based on Apples example code - but then found that Soundflower did the same thing, so I used that. Spent quite a bit of time integrating it (installer, transparent reselection of output device, play thru), but I think it's looking good. It'd be nice if there were a better way to record all system audio (listening Apple?), but alas - no API exists to do that, so it ends up being what in my opinion is "a bit of a hack".
I don't mean "hack" in that it's untidy, but it'd be much nicer if I didn't have to install kernel drivers and bounce audio between the Soundflower and the speakers. The biggest 'cotcha' in all this I'm betting will be that users get confused about what iShowU actually captures from. There are various applications out there that don't respond to a change of default system output (VLC being a prime example), so if you start them up - and then start iShowU - they don't send their audio to the right place, and it's not captured. Bit of a pain really.
I must confess that I think the interface is getting a little complex, but with the introduction of playthru and multiple audio channels, there needs to be some way to do things like set channel volumes and monitor the mix that would be produced. Well, I'd be really interested in what you have to say about how it feels, how it works. The general user experience interests me!
I was going to write my own reflector driver based on Apples example code - but then found that Soundflower did the same thing, so I used that. Spent quite a bit of time integrating it (installer, transparent reselection of output device, play thru), but I think it's looking good. It'd be nice if there were a better way to record all system audio (listening Apple?), but alas - no API exists to do that, so it ends up being what in my opinion is "a bit of a hack".
I don't mean "hack" in that it's untidy, but it'd be much nicer if I didn't have to install kernel drivers and bounce audio between the Soundflower and the speakers. The biggest 'cotcha' in all this I'm betting will be that users get confused about what iShowU actually captures from. There are various applications out there that don't respond to a change of default system output (VLC being a prime example), so if you start them up - and then start iShowU - they don't send their audio to the right place, and it's not captured. Bit of a pain really.
I must confess that I think the interface is getting a little complex, but with the introduction of playthru and multiple audio channels, there needs to be some way to do things like set channel volumes and monitor the mix that would be produced. Well, I'd be really interested in what you have to say about how it feels, how it works. The general user experience interests me!
|