Do Pi HAT Dac's need a reclocker?

Lol999

Just a simple question as I recently bought an Iquadio dac. I'm assuming it has the problem of jitter covered in its design like most external dacs which brings me to the Kali reclocker - is it necessary for just the Piano dac by design or would all HAT dacs benefit?

gvasco

Not necessarily, they’ll work fine without one. A reclocker is only useful if you really want to achieve the best possible audio output. It’s not necessary for any RPi I2S DAC but any and all slave DAC’s can benefit from it.

Sent from my iPhone using Tapatalk Pro

Lol999

So something like the Iquadio Pi-DAC pro which is essentially a master dac has its own provision for jitter?

R101

Reclocking the input is not going to make a difference to a slaved DAC which is asynchronous, as it is done in the chip anyway. Reclocking a synchronous DAC input from the poor quality Pi clock (if that is even possible) is unlikely to make much difference either.

If you have a master DAC board which is feeding the I2S clocks to the Pi, you obviously do not have to worry about reclocking. Galvanic isolation might be worthwhile, though.

Lol999

in the context of a raspberry pi and dac hat how is that possible?

R101

You mean galvanic isolation?

You will need an IL717 or a similar isolator chip between the HAT and the Pi, and a separate power supply for the DAC. The clock source has to be on the DAC side of the isolator, so this arrangement is only possible if the DAC is the I2S master. If the DAC uses I2C for setup, as well as I2S for the audio, it all gets too complicated.

janui

Hi Lol999, gvasco & R101,
The bottom line is: can I hear the difference?
Some people claim that the difference can be heard, but they generally provide no measurements to back up their claims.
If you are interested in this topic, read this: http://archimago.blogspot.nl/search?q=raspberry+jitter+measurements+reclocker
And if you are interested in more information, which is a little better related to RuneAudio and the Raspberry PI, lots to read with this search: http://archimago.blogspot.nl/search?q=raspberry+jitter+measurements
According to Archimago’s measurements: With the modern kit, which we are using, it is already very good. There can be a measurable difference, but normal people (like me) just can't hear it.
janui

R101

I agree with janui. If there is a clock source on the DAC board, you are not going to hear the effect of reclocking, as the input clock is not being used for conversion, and it is the conversion clock which matters. You might get a lot of dropped or interpolated samples, though, if the DAC is slaved to the Pi. How many samples have to be lost before it becomes audible is a nice point.

A Pi HAT is low-cost, convenient, and can give good results. It will never match a top-quality USB DAC, though, however much you polish it.

Lol999

Well in context of my listening equipment - sub £100 headphones, I'll probably not hear a difference

Chris63

I just purchsed a new Pi 3b and was uncertain about the subject of on-board clocks. I wanted the ultimate quality output as possible so opted for the HiFi Berry DAC+ PRO HAT as I didn't really understand what this funtion did. I'm still a little uncertain after reading the related posts here. As I'm considering buliding a Pi based player for a friend, is there any benefit then in using this DAC+ PRO card or would the middle band priced card with standard chrome plate RCA Phono's work just as well (i.e. it doesn't have the anti-jitter dual clocks on-board).

I do intend to comment in perhaps a more suitable Forum location, however, regarding the entire installation on the Pi everything has worked perfectly first time including the DAC+ PRO HAT. So all credit to you for a superb plug and play design! Oh, and the sound quality - SUPERB.
Thanks,
Chris.

R101

So far as I know, with the DAC+ PRO the Pi I2S output is slaved to the high quality clocks on the HAT. If all else is equal, it is going to be superior to the versions which rely on a synthesized clock frequency generated by the Pi.

The problem is that, even if you start with a bit-perfect output, all else is rarely equal. I doubt if a good asynchronous DAC with its own high quality clock source would actually sound worse, provided the difference in the clock frequencies did not cause too many dropped or repeated samples.