janneman said:
Then as you increase the load, you increase the noise.
Jan Didden
Found an interesting paper at Cirrus Logic about their CS4397. Think it could be valid for AKM too. They compared noise with 1k and 10K. Going from 1k to 10k lowered the dynamic range by 5dB. So a differential transformer load of 3-5kohm could be a good compromise.