I may have missed it in the thread, but have you tried a different esp device? These things sometimes aren’t well made and they can be, frankly, a POS .
Also when I look at projects that have successfully used the dfplayer they have used different pins. EG this one DIY Video Doorbell with Voice Response | ESPHome / ESP32 Camera / DFPlayer Mini uses a D1 mini with
tx_pin: D6
rx_pin: D5