How to change sample rate by use "sgtl5000"

[Shun]

Hi all,
We used Teensy4.1 and sgtl5000 complete transfer the audio data (44.1K) ,but now we need to cut down the audio data to 32K.
Develop by the URL https://www.pjrc.com/teensy/gui/index.html?info=AudioInputI2S. We didn’t see any function about change sample rate on the web page.


Thanks

 

[WMXZ]

Years ago I did some tests and a recorder for sgtl5000 that you could use as a starting point.
tests (T3.x): https://github.com/WMXZ-EU/test_sgtl5000
recorder(T3,T4): https://github.com/WMXZ-EU/record_sgtl5000
programs have not been updated in last 3/6 years, but at that time they were working.
There is also a post https://forum.pjrc.com/index.php?threads/overclocking-sgtl5000.52798/ on overclocking the sgtl5000 (not relevant for your case)

 

[Shun]

Whether I use https://www.pjrc.com/teensy/gui/index.html?info=AudioInputI2S can't change sample rade?

 

[WMXZ]

The basic library does not forsee to change the I2S sampling rate. You have to augment to library code.

 

[Shun]

The basic library does not forsee to change the I2S sampling rate. You have to augment to library code.

I understand. Thank you.

 

[Shun]

I have already read https://forum.pjrc.com/index.php?threads/overclocking-sgtl5000.52798/
Can I use "i2s_mods.h" If I develop using teensy 4.1?

 

[WMXZ]

The code in "record_sgtl5000", even if saying T4.0 should workk for T4.1.

 

[Shun]

Can I change the sample rate to "8K" .

Thanks

 

[WMXZ]

Yes,
take the code and change it

 

[Shun]

Sorry about I read the wrong library .
I thought the minimum was only 32K

thanks.

 

[Shun]

我參考使用 I2S_modification(fsamps[fr],32) 和 SGTL5000_modification(fr) 的範例，但仍無法變更取樣率。
fr=0

 

[Shun]

Then I found this. Does it mean that 8K cannot be used, or that I am missing some operations?

 

[MarkT]

Yes, you are missing SYS_FS and MCLK_FREQ, also in CHIP_CLK_CTRL which sets the clock, this in combination with the RATE_MODE determines the sample rate, see datasheet table 18. I think you'll need to reduce the MCLK frequency sent to the chip to get the lowest sample rates though.

But why not simply over-sample?

 

[Shun]

Because I want to reduce chip load and the audio encryption used requires 8k..