Internet radio and audio file player on ESP32-S3

New software version EVO3 MOD KOL .There is still a lot of room for improvement but you have a look for yourself . Version on Panel ILI9488 connections as before. Enjoy testing.

The visual side looks like this

NEW EVO3 MOD COLOUR TOOLING

WORKS ON PANELS ILI9488 AND ST7796 you have to change the rotation in the code because the image is rotated 180 degrees

And what kind of graphics? In terms of libraries?

>>21900341
LovyanGFX 1.2.21
Audio Version 3.4.6

LovyanGFX 1.2 audio version 3.4

>>21900347 Hi, can you please upload the file, .bin, thanks

>>21900357
I added the bin in the post test don't forget the connections

Connection to ESP32 S3 N16 R8 // -------------- OLED Display - pin definition -------------- #define SPI_MOSI_OLED 39 // MOSI pin (Master Out Slave In) for SPI OLED interface #define SPI_MISO_OLED -1 // no MISO for TFT display #define SPI_SCK_OLED 38 // SCK (Serial Clock) pin for SPI OLED interface #define CS_OLED -1 // CS residual screen to GND - dice on ILI9488 #define DC_OLED 40 // DC/RS screen on ILI9488 #define RESET_OLED 41 // RST screen on ILI9488 ST7796

Hello Robgold, thanks for the new soft version. I have a question, did you add, besides the flag, any pin dedicated for the STANDBY/MUTE type amplifier ?

Ejcon, a question, are you planning to control some LCD backlight pin for Power Off ? Well and the question of adjusting the edibility with the same pin you could PWM control the MOSFET for the LCD backlight.
Do you have Bluetooth RX/TX in this version ?
The idea is to implement audio transmission on the second ESP32 as you had in the previous mod on the OLED ?

Greetings....

>>21900366 well it looks nice, but .bin is for ILI9488 and I have st7796 and have a mirror image

In the web setting you have a rotation setting of 180 degrees, select it and update it, it works fine


Added after 7 [minutes]:

>>21900476
Yes in the softy is the BT control as before all is transferred from the Evo3 mod only on a new panel and here are the changes because everything is not yet completed but the work is in progress just like the previous softy should be the same functions.

>>21900541 yes, I know this rotation works up/down but I also have left/right

>>21900551
A correction must be made in codzie .

>>21900476 The backlight and amplifier control option is already in code 3.20.14.


Clock synchronisation has also been improved in 3.20.14. Now in sleep mode when we have the clock displayed it should show the time correctly. ESP wakes up every 3h and does a re-sync of the clock. I forgot to add this to the list of changes and it is one of the most important things that has also changed.

>>21900299 @ejcon In my opinion all these dynamic screens with bytes are an unnecessary expense, the fluidity is not there at all, and to stare at it- it just tires the eyes, better to harness the resources of ESP32 for smooth audio and network connection. If I were to do any analysis it would only be a screen on parallel communication, here's a comparison: https://www.youtube.com/watch?v=ZtCMIAmLSh8

Edi made a patent on a second ESP32 with an analogue input.

https://www.elektroda.pl/rtvforum/topic4173225.html

Such an analyser works quite cleverly.

Greetings....

>>21901191
Maybe something like this




Connectivity:
2. TWO TFT PANELS ST7789 1.9" 170x320
-------------------------------------
Both displays share the same SPI bus. Only the CS pin differs.

GPIO4 -> CS TFT left
GPIO5 -> CS TFT right
GPIO15 -> RX UART from master ESP32-S3
GPIO16 -> TX UART to main ESP32-S3

GPIO17 -> RST of both TFTs
GPIO18 -> SCK of both TFTs
GPIO22 -> DC of both TFTs
GPIO23 -> MOSI both TFTs

GPIO32 -> I2S BCK
GPIO33 -> I2S LRC
GPIO34 -> I2S DIN

3V3 -> TFT and module power supply
GND -> common groundTFT LEFT:

The mini spectrum code is a reworking of my earlier code. It could have been mentioned...

>>21901400
Sorry but I was writing quite late.
Already adding info.

ejcon wrote:

Sorry but I was writing quite late.

Bottom line.
I've been writing for a long time that this S3 from the radio is sweating from an excess of bits and bobs. A separate ESP for effects is a good direction. Audio over i2c in the spectrum is great as much as it is limited to one device.

>>21901130 I agree 100%. Byers, blinkers ...great. However, this is INTERNET RADIO and the most important thing is that this streamer is able to correctly play back also such HiRes stations with FLAC codec. In 99% after starting the station no one looks at the display. we "fire up" our favourite station and off we go.
This is why I am recompiling our own lwip libraries for the project, so that the radio plays correctly even up to 2500kbps/FLAC.

@efi222 The approach with a separate analyser is spot on.

What do you think about making a circuit as follows
1 main processor ESP32 S3 N16R8 radio Evo3
2.2 ESP32 S3 N16R8 processor Graphic equalizer with spectrum analyzer controlled from the main processor via uart
3 BT AUDIO ESP32 WROM transmitter controlled from the main processor via uart2
A PCM9211 could be added before the equalizer to feed other digital signals.

>>21901465
Point 2 and 3
If the uart is unidirectional one is sufficient Preferably UART2.
In the receiving ESPs the RXs are bundled together.
I have something similar done in another device with two ESP32s. I control them TX from the Esp8266 from the GUI.
Only in the command headers you have to add to which ESP32 the command is flying to. And they receive all the commands and filter if it is to it.

Spectrum software fix.
Cod for the spectrum reworked from the code of colleague efi222.

THANKS FOR A SHOOT OF GOOD WORK efi222
Such a little gadget about the demo every time you run .
A request not to remove this from the code .








Tested

ejcon wrote:

Such a little gadget about demn at every launch .
Request to not remove this from the code

No no... I really don't need to be on TFT at the start. It's enough that my nickname is in the main. You've done a great job too. I'll take a look tomorrow to see what you've come up with there

Ejcon would you upload some video of it jumping 😀

Greetings...

DJCheester wrote:

would you please upload some video of how it jumps

And sweep 64-16000Hz. I'd love to see how much "analysis" there is in this analyser.

>>21899630 Good morning

I compiled a new version yesterday, core 3.5.8, using audio library 3.4.6 (instead of 3.4.5q) and substituted lwIP libraries. I compiled under Arduino 2.3.8.
It works brilliantly, compared to the previous version it runs faster, and on my favourite stations, which by the way were probably the most cutting at startup: PR Two and Radio Chopin PR in aac320 and Cesky Rozhlas D major (flac) stopped cutting.

I don't know what caused this improvement, whether the new core (the radio works more nimbly in general) or the new I2S Audio library, or perhaps the "robgold" libraries

The earlier compilation (3.20.04) on the above stations had problems, especially right after loading the station, I was combining with different versions of the lwIP libraries provided by my colleague Robgold and on these stations (especially Cesky Rozhlas) it was cutting heavily. Surprisingly, other stations, even with higher bitrates (e.g. Linn) were running smoothly. I considered that the problem might be on the side of the broadcaster.

To my colleague Robgold:
I also think that additional fabs like spectrum analysers or pseudo-analogue VU-meters are not necessary, it is better to focus on basic functionality, so as to squeeze the maximum performance out of a chip that was supposed to be used in fridges and other IoT....

Now I am even thinking about simplifying this project (1 bank, only a big station name on the display, maybe a clock, etc.) in terms of a radio for an elderly person, so that its operation is as similar as possible to an analogue radio (you know, except for WiFi configuration).

>>21902025


To improve stl 6.

Ejcon did you compile this as recommended on core 1.04 with the changes as efi222 wrote because somehow these posts go dimly.

In effi's or my compilation it is much better

Take a look

Audio in analogue

Greetings....

DJCheester wrote:

you compiled this as recommended on core 1.04

This low core was because of the high ADC sampling. In higher versions this is messed up. But there is no ADC in the ejcona software.

@ejcon and falling bars do you have every 3 pixels?

That's what you explained to me, but you have to admit that it works without delay. I wonder if this is because of I2S?