DSP volume control or DAC volume control - which should I use?

I am using ADAU1452 for crossover and EQ functions, then I reclock the signal with a FPGA reclocker, then send the signal to two AK4490 DACs, which have built-in volume control and operate on 32bits.

In short, I would like to know if there is any qualitative difference between using DAC built-in volume control, or setting the volume in DSP and sending attenuated signal to the DACs?

I'll provide some details - right now I control volume by using an Arduino connected to both DACs via I2C, but the Arduino is causing unwanted noise, which I confirmed by measuring output with and without it. I tried isolating Arduino from DACs by using I2C isolator (SDA/SCL/GND, supplied on the reclocker) and this would be perfect if not for an additional wire (PDN) needed from the Arduino to turn the DACs on initially, thus rendering the isolator useless. AK4490 has to be in software mode to use the volume control, and in that case it has to receive the ON signal via PDN. If I can use the DSP volume control then I would just leave them in hardware mode and omit PDN and I2C altogether.

I am sending 32bit signal from DSP to the DACs, and use a +42dB gain to account for 32bits (actually +32dB to leave headroom for EQ). You can read Dave Thib's explanation to me here. So I would like to use this slider as main volume control, either by connecting it to an external potentiometer or I2C signal over GPIO.