Digital Audio Output Differences? Is USB Audio Really The Top-notch?

[Preston8452]

Hello guys,

So after some some struggles between spending time in optimizing my PC audio functionalities and adopting a dedicated music streamer, I opted for the second one.
I purchased the Munich M1T by Silent Angel, given it's good reviews in UK and Germany and Switzerland, I thought it wouldn't hurt to give it a shot, but frankly speaking, the mainly reason was because the price fits my wallet more hahahah...

Anyways, so there are several digital outputs on its back panel, AES/EBU; I2S(HDMI), Coaxial, USB Audio, hell hahah, I have literally no idea which one is better or which can provide a special feature or something, I mean it does show that the USB Audio supports up to up to PCM 768KHz and DSD 11.2M (DSD256), but does that the higher the sampling rate, the better the sound quality? I just can't make sure...

And if the USB Audio can provide the " best " digital sound quality, then why added other ports? Ain't that just costing more on the manufacturing end?
Hope you guys can give some insights on this, or share some of your experiences, because this is the first time I saw an AES/EBU port on a audio streamer...

Best,

 

[Frank Sol]

Which OUT sounds better with YOUR setup ?

Dunno... You tell us

Why different OUTs ?

To better match individual systems

 

[cpt_paranoia]

I purchased the Munich M1T by Silent Angel,

A Raspberry Pi 4, in a box, with added audio interface circuitry.

Essentially, audio interfaces fall into two categories:

1. source-clocked interfaces
2. destination-clocked interfaces

SPDIF & AES are source-clocked interfaces
I2S and USB can be source-clocked or destination-clocked.

There are two types of degradation that can be carried into a DAC by the digital interface:

1. Clock noise (jitter)
2. Electrical noise

The best way to eliminate clock noise is to provide a clock, as close to the DAC chip as possible, and isolated from electrical noise.

In a source-clocked interface, the clock must be carried in to the DAC, over an interface. This can pick up electrical and timing noise, from cables and the various interface components (e.g. optical transceivers). This noise can couple directly into the DAC output, as sample jitter, unless measures are taken to clean up the clock within the DAC. For SPDIF and AES, this clean up can be done by the clock recovery circuit that is required to extract the clock from the single, biphase encoded signal used by SPDIF/AES.

In a destination-clocked interface, the DAC requests samples from the source, using a two-way interface, with handshaking. This sends blocks of samples to the DAC, and these are put into a FIFO buffer, and read from the FIFO by the DAC chip as required. A control mechanism monitors the FIFO content, and sends block requests to the source. The control mechanism must prevent FIFO overflow or underflow. The DAC clock, local to the DAC chip, can be as clean as possible, and is used to fetch samples from the FIFO. This is the philosophy used by an integrated CD player; the CD read mechanism is slaved to the DAC clock, and the FIFO is provided as part of the CIRC error correction & recovery.

I2S can use a source clock, provided by the streamer, or a destination clock, provided by the DAC. Not all I2S sources will accept a destination clock, though.

USB can push data into the DAC (source-clocked 'synchronous USB'), or have data pulled from from the source (destination-clocked 'asynchronous USB'). Not all DAC USB interfaces can support asynchronous mode operation.

From a technical point-of-view, a destination-clocked USB is the interface I would prefer to use, as it offers the potential for the highest quality audio, by using a local DAC clock and a FIFO buffer to isolate USB timing from the DAC timing. The 'causation barriers' I mentioned in your other recent thread on 'audiophile switches' is relevant here: the FIFO is a causation barrier.

As always, implementation is also important; electrical noise can be carried in by coax SPDIF, I2S, USB, Ethernet, etc. It is the job of the digital interface to prevent that electrical noise reaching the DAC analogue output. That can be done by good design. Equally, a bad design will allow electrical noise to reach the DAC output. No interface type is inherently better; it depends on implementation. But a destination-clocked interface has the potential to be the best.

 

[cpt_paranoia]

Why different OUTs ?

To better match individual systems

Yup. To provide versatile support for different equipment.

 

[cpt_paranoia]

A Raspberry Pi 4, in a box, with added audio interface circuitry.

Given that it claims to support Roon and AmazonHD, I wonder what OS they are running on it. I can't think of an open source audio OS/app that supports AmazonHD. Maybe they've written their own AmazonHD plugin for LMS or r1Audio. Which would be interesting, as I'm not sure Amazon have released the API. As it's not in the 'super audiophile' £15k streamer league, I doubt they have written their own OS/app entirely, so I'm sure it will be based on one of the usual suspects.

Also interesting that the DAC o/p version is only another £200; that's not a big delta, given the £700 baseline for the purely digital streamer version.

I confess that I am suspicious of any product from a company that sells an 'audiophile switch' (which appears to be nothing more than a conventional switch and an unnecessary TXCO) for £400... Hope it works out for you. Given the non-availability of RPi 4s at the moment, and the corresponding price hikes being seen, that £700 price tag doesn't seem so unreasonable...