If it is that dramatic, I am sure it is not psychoacoustic. If it is, who cares. You could do a single blind test. Have someone switch the I2S DAC settings. If you can pick 6 out of 6 correctly, that is two standard deviations certainty against psychoacoustic. Listen to it for a week with various recordings you know. Then go back and see if you can stand to listen to it.
With the AK4137, set it for no upsampling, i.e. the same bitrate. It also reclocks the I2S signals. If you upsample, it will be doing something similar to an oversampling chip, and it has a choice of filter settings. You don't want that. The AK4137 can go into a bypass mode, which only does the reclocking, but I am not sure that the board can set bypass mode. The closest would be to set the outgoing sample rate the same as the incoming sample rate.
These boards were being used to convert PCM to DSD, sent to a DSD capable DAC chip. That also might be worth checking out. DSD playback is nonoversampling. All that dithered ultrasonic noise is present in the output signal.
With the AK4137, set it for no upsampling, i.e. the same bitrate. It also reclocks the I2S signals. If you upsample, it will be doing something similar to an oversampling chip, and it has a choice of filter settings. You don't want that. The AK4137 can go into a bypass mode, which only does the reclocking, but I am not sure that the board can set bypass mode. The closest would be to set the outgoing sample rate the same as the incoming sample rate.
These boards were being used to convert PCM to DSD, sent to a DSD capable DAC chip. That also might be worth checking out. DSD playback is nonoversampling. All that dithered ultrasonic noise is present in the output signal.
Hardware: RPi Zero W | Allo Kali | ProtoDAC TDA1387 X8 | PGA2311 | Icepower 500ASP | Harbeth SHL5
Software: Moode 8.3.3
Source: Win 10 NAS
Software: Moode 8.3.3
Source: Win 10 NAS
