Update to play_sd_wav/raw with method to use different FS

[mjs513]

While playing with MTP/FS Audio we modified the play_sd_wav and play_sd_raw with an addition method based on what @wwatson did for Franks Codec library: https://forum.pjrc.com/threads/68816-Now-for-MP3-with-FS to use different FS as well. So after a bunch of debugging by @KurtE and @Frank B we now have all of Frank's codec library working with the T4.1 (+audio shield). With this change we are now able to play wav files on FS device - tested with memboard chips, QSPI, SD Cards and USB. All seem to be working well.

Basically just added a second play method: bool AudioPlaySdWav::play(FS *fs, const char *filename) and change using SD.open to wavfile = fs->open(filename);

nothing else changed.

Don't have any midi equipment or fancy audio stuff to test so fair warning.

[PaulStoffregen]

Frank changed his code's license back to GPL, so I can not merge any code which is derived in any way from Frank's work.

[mjs513]

Actually, this change is only to allow use of play_sd_wav and play_sd_raw to use LFS, or SD or USB which is totally independent of Frank's stuff - it was just the impetus and test method. Could easily just have use an example without Frank's codec.

Really don't understand the license issue I guess.

[FrankBoesing]

+if you want to merge the new waveplayer, you can still merge the MIT version from the history. No code change happened since then.

I that really happens, i'll be happy to change the licence back. After that.

[PaulStoffregen]

@FrankBoesing - Would you be willing to create a branch within your public repository from the last commit before the license change to GPL?

[FrankBoesing]

https://github.com/FrankBoesing/Teensy-WavePlayer/tree/MIT

[PaulStoffregen]

Thanks. I've created a fork.

At this moment my workbench has a complicated USB setup to investigating the USB packet loss issue with MTP, so I probably will not touch any audio stuff for a while.

[KurtE]

@mjs513 @PaulStoffregen - There is a current thread up on the forum: https://forum.pjrc.com/threads/71716-recording-playing-back-audio-on-multiple-SD-cards

Where they would like to play back a raw capture file that they may store on different SD cards. Where they have maybe 3 SDs setup.

I know we were playing with this PR and playing media (mostly wave files) stored on other storages.

If I remember correctly, this PR worked fine.
I also showed a minimalistic set of changes on the forum where we can: pass in a reference to an FS object like:

class AudioPlaySdRaw : public AudioStream
{
public:
    AudioPlaySdRaw(void) : AudioStream(0, NULL) { begin(); }
    void begin(void);
    bool play(const char *filename, FS &fs=SD); 

And likewise in the cpp file change:

bool AudioPlaySdRaw::play(const char *filename, FS &fs)
{
    stop();
#if defined(HAS_KINETIS_SDHC)
    if (!(SIM_SCGC3 & SIM_SCGC3_SDHC)) AudioStartUsingSPI();
#else
    AudioStartUsingSPI();
#endif
    __disable_irq();
    rawfile = fs.open(filename);
    __enable_irq();

where the fs. was SD.

If this way is preferred, would also do it to the other class as well

[h4yn0nnym0u5e]

If this way is preferred, would also do it to the other class as well

This seems a much cleaner approach to me - it's what I've done for my buffered version and it works fine, and makes maintenance of a single play() function much easier.

One further comment ... is there a way of determining whether the FS in use actually needs SPI? As-is, the code assumes it does, but I'd guess (haven't tried it...) that a filesystem on USB Host or littleFS doesn't do so. Not sure about SDIO, either. Are there any drawbacks to leaving AudioStartUsingSPI() in?