Writes Remy Cucui, July 2:
What I've found is that the MP3 shield doesn't play files directly or even read them straight from the SD card. There's no "Play MP3" function, you just feed it a buffer and it works with it. We can work with several file types: MP3, OOG, AIFF, WAV, leaving us the opportunity for dealing with uncompressed audio. I tend to think that this makes real-time interleaving a more practical possibility with the Arduino. I want to try that and see how well it works.
The new idea I've had about handling audio isn't quite so spectacular, though. It's actually more of a fall-back method. We could have both of the sound files already interleaved at multiple points in time on the SD card. Several instances where Sound #1 has been playing and then Sound #2 starts, and vice versa. Then, when the second sensor is triggered, it's read from the corresponding mixed file at the same point as the original one. People viewing the exhibition won't know to pay attention to how soon Sound #2 plays after Sensor #2 is triggered, so we wouldn't have to create that small of a gap between the progression of each sound file along the other's timeline. Making all of the sound files won't be hard, we could just write some code that does it repeatedly. The only concern is space; I'm not sure how big the files we're working with will be.
Dr. Marantz replies:
My worry is how you'll mix the current single sound [Sound #1] file to the next sound file [Sound #2] that should be starting from where the first gets interrupted by the trigger of the second sensor [Sensor #2], since you would be accessing a combined sound file. You could momentarily turn down the volume of the first soundtrack before starting the second, but if the sound track is pure speech and the time is off, there may be a momentary echo effect (which might be a good thing?) [AG: Note: there is no spoken word; only ambient machine sound and bird songs]. Though that's something that you can experiment with to get the best effect.
To that, Remy:
Oops! I thought I caught both instances. Fixed! The idea is to start playing the mixed file from the same point at which the original file had been playing, instead of playing exactly where the mix begins. Like going to the subway and catching the next train. We read from the same point in the mixed file as we were in the original, and after some odd hundredths of a second, the other sound will play as well.
AG: My concern is pre-mixing the sound and not allowing the viewers to determine the playback variable.
Philip Gonzalez, via FB: If it's a disc (CD) you are working with, in the program you are saving it on there should be an option to take away the space between the songs before you write the music to disc. On the other hand if you are using a portable device depending on the brand, you will need to sync to exact times which can be a pain. If it is a device that is problem look up the manual and look under "time signatures" or "syncing" for instructions on how to do this.
ReplyDeleteThe sound track isn't music or spoken word. It's ambient machine sound and bird calls, which may be forgiving. However their interleaving may prove to be the challenge, precisely because some sound is continuous, ongoing.
ReplyDelete