GitLab

The TS packets have a maximum size of 188 bytes, AAC frames are larger. It is better to unpack several TS packets, write them into the audio buffer and only then send the data to the AAC decoder

http://mdstrm.com/audio/6237761f64b72676f05cf6ab/live.m3u8
https://live7digi.antenaplay.ro/radiozu/radiozu.m3u8
http://us-b4-p-e-zs14-audio.cdn.mdstrm.com/live-audio-aw-bkp/6237761f64b72676f05cf6ab/playlist.m3u8?listeningSessionID=62a10f0200a9bf00_2098051_cKYJo206__0000001WWlO

new working streams
http://18683.live.streamtheworld.com:3690/XHPIFMAAC/HLSTS/playlist.m3u8
http://26373.live.streamtheworld.com:3690/XHQQ_FMAAC/HLSTS/playlist.m3u8
http://20283.live.streamtheworld.com/XHQQ_FMAAC/HLSTS/playlist.m3u8
http://19003.live.streamtheworld.com/BLUE_FM_100_7AAC/HLSTS/playlist.m3u8

e.g. https://air.pc.cdn.bitgravity.com/air/live/pbaudio001/playlist.m3u8
works with ESP32-S2
ESP32: set in constructor `m_i2s_config.dma_buf_count` higher to 20

as a result, the AAC decoder finds the first syncronword with pinpoint accuracy

if http status code 301 "Moved Permanently" wait for Location: and redirect to new url
correct connection time

https://cdnlive.shooowit.net/rtvalive/smil:channel13.smil/playlist.m3u8
this valid url has a colon, but without a portnumber follows.

now working m3u8 streams:
http://189.240.210.28:1935/setacatlan/radio/chunklist_w1853320237.m3u8
http://stream.revma.ihrhls.com/zc3201/hls.m3u8
http://arcast.com.ar:1935/radio/radionar.stream/playlist.m3u8
http://lampsifmlive.mdc.akamaized.net/strmLampsi/userLampsi/playlist.m3u8
http://6182bfa7cd933.streamlock.net:1935/Radiokriti/Radiokriti/playlist.m3u8
https://cdnlive.codev8.net/canalextremaduralive/smil:channel2.smil/playlist .m3u8
https://cdnlive.shooowit.net/rtvalive/smil:channel13.smil/playlist.m3u8

/Audio.cpp:2254:14: warning: unused variable 'pos' [-Wunused-variable]
Audio.cpp:2538:25: warning: unused variable 'size' [-Wunused-variable]
Audio.cpp:2496:29: warning: unused variable 'seqNr' [-Wunused-variable]
Audio.cpp:2433:10: warning: unused variable 'f_entry' [-Wunused-variable]
Audio.cpp:2434:10: warning: variable 'f_ExtInf' set but not used [-Wunused-but-set-variable]
Audio.cpp:2435:9: warning: unused variable 'pos' [-Wunused-variable]
Audio.cpp:2436:11: warning: unused variable 'host' [-Wunused-variable]
Audio.cpp:2438:10: warning: variable 'f_StreamInf' set but not used [-Wunused-but-set-variable]
Audio.cpp:3440:38: warning: format '%s' expects a matching 'char*' argument [-Wformat=]
Audio.cpp:2496:29: warning: 'seqNr' defined but not used [-Wunused-variable]
Audio.cpp:3100:33: warning: '%s' directive writing up to 511 bytes into a region of size 492 [-Wformat-overflow=]

no new features, but the code is more structured

New TTS provider MaryTTS

and have more SRAM for your projects,
if  ESP32-S3 and have PSRAM: AAC Spectralband Replication (SBR) is enabled for better sound quality