Audio on the Teensy 3.1

[jomac]

Hello,
Im looking for a little guidance and advice here. Im working on a project that needs audio feedback on functions selected for blind users. This is for an amateur radio transceiver. Im not an expert in Teensy audio, and not too sure of the best method to use. One of the functions would be to read out the current frequency selected, ie 14.123 Mhz which would be stored as 7 separate files, and each one called up as needed to generate the full frequency phrase that is required.

What concerns me is the access time required to obtain the 7 files to generate the full frequency chosen. Would i be better storing each phase as an MP3 file, or as a RAW wave file, which would be the quickest and easiest method to use?

Any help would be greatly appreciated.

John

 

[bloodline]

Check out my PWM audio playback code in this thread: http://forum.pjrc.com/threads/25032-Multiple-Audio-outputs

There is plenty of space to store the audio in the Flash as 15,625Hz 8bit audio samples. MP3 might be smaller, but I have no idea how you would decode them with the teensy.

 

[jomac]

Thanks for the info and links, i will have a deeper read later on, however, its not the storage or access to the files thats of concern, but the time taken to do so. If i wanted to for arguments sake, let the Teensy count from 1 to 9 without any noticeable pause between each number, i want to be able to store the audio for each number, 9 separate files, and then access and playback each one, one at a time.

 

[bloodline]

My audio code is asynchronous, there is no delay playing an audio sample from the Flash, you can have a large or as small a gap between the samples as you like.

 

[jomac]

I will have a good read of this and the links in the topic, this may be what im looking for.

 

[PaulStoffregen]

You could also try the new audio library. Here's the closest example:

https://github.com/PaulStoffregen/Audio/blob/master/examples/PlayFromSketch/PlayFromSketch.ino

You could delete quite a bit of that code, since you're not looking to play 6 sounds simultaneously. Perhaps just 1 AudioPlayMemory object feeding into a single AudioOutputAnalog would do the job?

 

[jomac]

Thanks for the help Paul. This idea really grew from using a DUE and colour touch screen to control a popular ham radio set. I looked at using the DAC output on A14 to generate beeps for warnings etc, from there we looked at voice announcements for functions on the radio. This ended up as a request from a ham radio blind group who asked if all functions could be announced by speech. The only problem we see is possibly the access times, the teensy is controlling a display and a frequency synth via a rotary control, and when we pause the tuning knob, we want it to announce the frequency, whatever it is. Any delay in doing this would slow down tuning of the set.

John

 

[PaulStoffregen]

When you actually implement this stuff, I believe you'll discover you actually will need to add some intentional (but small) delays for it to work properly. You may not believe this right now, but let's chat more after you've gotten a little farther into the implementation....

I'd recommend using the audio library's AudioPlayMemory and AudioOutputAnalog objects. They provide excellent sound quality with very low CPU overhead. Most importantly, they fully in the background using interrupts and DMA with very low CPU overhead, which will maximize the time your sketch can manipulate the color display and communicate with the radio.