OpenAudio Library for Teensy

*** Special Note *** 6 January 2021- This library is undergoing revision to make Teensy 4.x compatible and to add functionality. The Tympan Library files and associated classes output_i2s_f32.h, output_i2s_f32.cpp, input_i2s_f32.h, input_i2s_f32.cpp are now ready to be used for T3.x and T4.x. There are some restrictions, particularly this should be used with 16-bit I2S codec data. Codec sample rates can be varied. Variable block size is supported, but be sure the settings option is used. More on this to come. Work on the F32 i/o routines is continuing. Thanks to Chip and @jcj83429 .

Purpose: The purpose of this library is to build upon the Teensy Audio Library to enable new functionality for real-time audio processing.

Approach: I am attempting to follow the structure and approach of the Teensy Audio Library so that coding techniques, style, and structure that works for the Teensy Audio Library will also work with my library. synth_sine_f32

Send Help!: While I'm hoping to follow the pattern of the Teensy Audio Library, I am not a well trained programmer, so I have difficulty chosing the right path. If you see ways to make my code better aligned with the Teensy Audio Library, send a pull request!

Contents

As I said, the goal of this library is to extend the functionality of the Teensy Audio Library. My extensions are focused in these areas:

Floating Point Audio Processing: The Teensy Audio Library processes the audio data as fixed-point Int16 values. While very fast for the Teensy 3.0-3.2, it is unnecessarily constraining (IMO) for the Teensy 3.5 and 3.6, which both have floating-point units. To enable floating-point audio processing, I have created:

• **AudioStream_F32: Following the model of Teensy's AudioStream class. In addition to the new AudioStream_F32, this file includes a new audio_block_f32_t in place of audio_block_t, as well as a new AudioConnection_F32 in place of AudioConnection.
• New F32 Audio Processing Blocks: Using this new floating-point capability, I've written some new audio processing blocks that operate using floating-point audio values instead of fixed-point values. Examples include AudioEffectGain_F32 and AudioEffectComperessor_F32.
• F32 Versions of Existing Blocks: To maintain continuity with the Teensy Audio Library, a number of the Audio Library's Audio blocks (which process as Int16 data) to operate on floating-point (ie, Float32) data. Thanks to other people's contributions, examples include AudioMixer4_F32 and AudioMultiply_F32. Feel free to convert more blocks, if you'd like!
• New Hardware Controls: The Teensy Audio Library does a great job of wrapping up the complicated control of hardware elements (such as the SGTL5000 audio codec) into a nice easy-to-use classes. But, sometimes, certain features of the hardware were not exposed in those classes. Here, I've extended the default classes to expose the extra features that I want. See AudioControlSGTL5000_Extended as an example.

After installing this library into your Arduino->Libraries direction, you can have access to any of these capabilities simply by including the following command in your Arduino sketch: #include <OpenAudio_ArduinoLibrary.h>.

Installation

Download (and unzip) this library into the Arduino libraries directory on your computer. On my computer, that means I put it here:

C:\Users\chipaudette\Documents\Arduino\libraries\OpenAudio_ArduinoLibrary

Restart your Arduino IDE and you should now see this libraries example sketches under File->Examples->OpenAudio_ArduinoLibrary

Dependencies

This library extends the functionality of the Teensy Audio Library, so you'll need to install it per its instructions.

The floating-point processing takes advantage of the DSP acceleration afforded by the ARM M4F core inside the Teensy 3.5/3.6 processor. Therefore, it uses arm_math.h. This dependencies is installed automatically when you install the Teensy Audio Library, so you don't need to take any extra steps. It's there already.

Status

The following table summarizes the status, as of 24 November 2020. I will try to keep this up-to-date as work continues, but tere will be lags. Thus there may be more T4.x compatibility than shown here. A blank entry for status does not indicate a problem, but rather a lack of verification. Further examples will come to test some of these objects. Update 29 Jan 2021: Replaced WDRC2 with AudioEffectCompressor2_F32. See Example testCompressor2.ino Update 18 Feb 2021: Added analyze_fft1024_F32.h and .cpp that parallels analyze_fft1024.h and .cpp in the Teensy Audio Library, but F32. Update 8 March 2021 Complex input FFT classes have been added for 256, 1024 and 2048 input points. These see use in Software Designed Radios for double Fs spectral analysis with I-Q inputs. The example files (from the Examples directory) are

1. AudioTestAnalyzePhase_F32

2. AudioTestPeakRMS

3. AudioTestSinCos

4. BasicCompressor_Float

5. BasicGain_Float

6. FineFreqShift_OA

7. FormantShifter_FD_OA

8. FrequencyShifter_FD_OA

9. LowpassFilter_FD_OA

10. MixStereoToMono_Float

11. MyAudioEffect_Float

12. OscillatorWithPitchmod_Float

13. ReceiverFM

14. ReceiverPart1

15. ReceiverPart2

16. Repair audio mixer_F32

17. SignalNoise_float

18. Switches_float

19. TestEqualizer1

20. TestEqualizer1Audio

21. TestFIRGeneral3

22. TestFIRGeneralLarge4

23. TestFIRGeneralLarge5

24. TestInputOutput_float

25. Tympan_TLV320AIC3206

26. testWDRC2 (not supported)

27. testCompressor2

WORKING NOTES - For now, here are some temporary notes:

1. synth_sine_f32.h & .cpp allowed both KINETISK and__IMXRT1062__ to support T4.x. Tested 3.6, 4.0
2. Fix synth_pinknoise_f32.h & .cpp built in I16 to F32. Now T3 and T4.
3. Fix synth_whitenoise_f32.cpp to include IMXRT1062 and built in I16 to F32. Now T3 and T4.
4. Added Gaussian White Noise F32. This is T3 and T4.
5. Renamed 3206 codec to .xxx to stop collisions with Teensy I16 library. Needs work.
6. Added new examples/SignalNoise_float.ino to test white, pink, Gaussian noise and sine. Audio out.
7. Disabled output_i2s_f32.h, .cpp by .xxxfile type. These need rework for T4. Use Convert_F32toI16.
8. Disabled control_tlv320aic3206.h, .cpp by .xxxfile type. These collide with Teensy I16 Audio.
9. Added new analyze_peak_f32.h and .cpp that parallel similar classes in the Teensy I16 library.
10. Added new analyze_rms_f32.h and .cpp that parallel similar classes in the Teensy I16 library.
11. Moved AudioSwitch_F32 from Tympan, added 8-channel version.
12. Added /examples/Switches_float.ino for 4 and 8 channel switches.
13. Moved FFT_Overlapped_F32 files from Tympan, revised for T4.0, added _OA name isolation.
14. Moved Overlapped FFT LPF .INO from Tympan to OA. Working T3.6 and T4.0
15. Moved Overlapped FFT AudioFormantShifter INO and related from Tympan to _OA Open Audio. Tested 3.6 and 4.0.
16. Moved AudioFrequencyShifter INO and related from Tympan to _OA Open Audio. Tested 3.6 and 4.0.
17. Working on SdFatWriter and related SdFat files. DO NOT USE.
18. Modified AudioEffectDelay for T4.0.
19. Moved in radio function AudioFilter90Deg_F32.
20. Moved in radio function RadioIQMixer_F32.
21. Added Example INO, FineFreqShift_OA.ino with Hilbert shifter and delay stereo.
22. Brought in support stuff mathDSP_F32 and sinTable512_F32.h.
23. Brought in AudioAnalyzePhase_F32 and Example
24. Brought in AudioFilterEqualyzer_F32 and 2 Example
25. Brought in AudioFilterFIRGeneral_F32 and 2 Example INOs
26. Brought in RadioFMDetector_F32 and Example
27. Brought in synth_sin_cos_F32 and test example
28. Brought in RadioNoiseBlanker_F32 and Example
29. Created output_i2s_OA_F32.h and .cpp to have F32 input. SEE 31 below
30. Repaired Audio Mixers 4 & 8. These are adders, NOT RF DBM's---see Multiplier_F32.
31. I2S Input and Output made T4 compatible, both supporting variable sample rate
32. Added AudioEffectCompressor2_F32 quite general audio compressor.
33. Added analyze_fft1024_F32 that does audio in and rms volts, power or dBFS out.
34. Added analyze_fft256_iq_F32 that does complex (I-Q) in and rms volts, power or dBFS out.
35. Added analyze_fft1024_iq_F32 that does complex (I-Q) in and rms volts, power or dBFS out.
36. Added analyze_fft2048_iq_F32 that does complex (I-Q) in and rms volts, power or dBFS out.