Can anybody explain to me what the obsession is with people only wanting to use a USB connection from their streaming device to their DAC? This seems very prevalent in the PS audio forums, like coax or optical is just not up to transferring bits correctly so must have USB for anything to work correctly.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Why use USB from streamer to DAC?
- Thread starter kevin1969
- Start date
What’s your obsession with coax and optical? And why would you care what people at the PS Audio forum think?
On a serious note, USB is present on basically any modern device, and especially on PCs. Coax or Toslink you almost never see anymore these days. So it’s just very convenient to use USB.
And as @TonyJZX mentioned, USB also offers some modes of operation that the other two can’t offer.
Last edited:
It's 2023 and asynchronous USB is a much better interface for modern DACs. I use Roon so it's nice to be able to deploy cheap Raspberry Pi's with Ropieee as a Roon Bridge and USB takes care of that. I use optical from my TV for video streaming, and Coax from my blu-ray player (to HDMI extractor) for discs.
JLGF1
Active Member
- Joined
- Feb 23, 2023
- Messages
- 201
- Likes
- 144
Yep, like the WiiM streamers.
WiiMs are just not modern enough.
- Thread Starter
- #10
Maybe there's a reason people are using USB over coax or optical and I just don't know the reason why because I'm not an expert on the subject but thanks for your feedback.
- Thread Starter
- #11
Okay that makes sense. Thx!
JLGF1
Active Member
- Joined
- Feb 23, 2023
- Messages
- 201
- Likes
- 144
Funny, that doesn't stop them from streaming hi-res music across two floors and 3000 sq ft.
popej
Active Member
- Joined
- Jul 13, 2023
- Messages
- 281
- Likes
- 185
Most PC have decent analog output, probably with better SINAD than average amplifier. Why use external DAC at all?
As for Raspberry, I simply use HDMI for audio. Not the best solution but convenient.
Have you seen Amir's measurement of his Dell XPS desktop?
Dell XPS 8930 RealTek HD Audio Interface Review
This is a detailed measurement of the RealTek HD audio interface in my Dell XPS8930 Desktop system. The XPS 8930 retails for US $1,200 as of this writing so not a low-end desktop tower. Try as I might, I could not find the identity of the RealTek CODEC (ADC/DAC) used in it. Nothing shows up...
www.audiosciencereview.com
[Edit] And here is an IMD comparisons of a few other computers and a pc sound card.
Last edited:
JayGilb
Major Contributor
Almost every personal computer has an USB interface, the same cannot be said for coax or optical. I personally use optical, because I can't hear any difference of sampling > 48kbps.
AnalogSteph
Major Contributor
That's what you get when buying from penny-pinching OEMs, I guess. A 3.3 V analog supply will be fine. Yeah right. They clearly didn't care one iota about audio performance, it was just more thing to be ticked off. That's one thing I used to like about Fujitsu(-Siemens) PCs, they generally didn't use fancy codecs but had this attention to detail spent on them.
Last edited:
popej
Active Member
- Joined
- Jul 13, 2023
- Messages
- 281
- Likes
- 185
Looks like designed for a tube amp
Some years ago I was interested in performance of PC audio. Here some RMAA measurements:
Relatively inexpensive PCI card: Audigy 2 Vaule.
Rather low end chip used on motherboards: Realtek ALC892.
If we are comparing S/PDIF (or AES3) to USB UAC2 you might think that they should be the same since we should not expect to drop bits.
So let's assume that the digital data is identical from a data perspective. Let's also assume that whatever digital interface we are using is not adding noise to our DAC.
USB UAC2 has an important advantage over S/PDIF or AES3. The reason is jitter. The right data at the wrong time is the wrong data. This becomes increasing demanding as the data converter improves.
With S/PDIF the clock is recovered from the data stream. This means clock is generated at the server and will always deteriorate by the time it reaches the DAC DIR. The DIR recovers the clock. You have choices:
1. Use a PLL to provide jitter attenuation. The WM8805 was the typical best choice for this.
2. Use an ASRC to resync to a local master clock. If done with a good device, this will effectively make the jitter as good as the local clock.
3. Use a giant RAM buffer and reclock out with the local clock.
4. Do nothing to improve the jitter
FWIW: Option 4 is fairly horrible.
USB UAC2 uses the local clock as the master clock. This means the jitter is a function of the DAC where it should be much lower.
You can think of S/PDIF as a push method and UAC2 as a pull method.
Al
So let's assume that the digital data is identical from a data perspective. Let's also assume that whatever digital interface we are using is not adding noise to our DAC.
USB UAC2 has an important advantage over S/PDIF or AES3. The reason is jitter. The right data at the wrong time is the wrong data. This becomes increasing demanding as the data converter improves.
With S/PDIF the clock is recovered from the data stream. This means clock is generated at the server and will always deteriorate by the time it reaches the DAC DIR. The DIR recovers the clock. You have choices:
1. Use a PLL to provide jitter attenuation. The WM8805 was the typical best choice for this.
2. Use an ASRC to resync to a local master clock. If done with a good device, this will effectively make the jitter as good as the local clock.
3. Use a giant RAM buffer and reclock out with the local clock.
4. Do nothing to improve the jitter
FWIW: Option 4 is fairly horrible.
USB UAC2 uses the local clock as the master clock. This means the jitter is a function of the DAC where it should be much lower.
You can think of S/PDIF as a push method and UAC2 as a pull method.
Al
Last edited:
and it will handle multiple channels.
AnalogSteph
Major Contributor
It would have been better to do this one at 48 kHz since the resampling artifacts are obvious (high-frequency components in THD+N and visible periodic ripple on FR), but performance is still approaching what a CS4382 can do. In fact, the 0202 USB sports the 2-channel version of this DAC, the CS4392, so you can just look at a loopback for comparison:
Perfectly usable, with no periodic ripple and total dynamic range slightly beyond CD levels. Instantaneous dynamic range could be better, but I doubt this is going to be a problem very often if ever (it's still 80 dB - you try hearing white noise 80 dB below a real-life signal). And that's basically the bread and butter option you find on boards that really have to keep parts cost down these days. The common ALC897 is better.
Props for avoiding ground loop issues BTW.
Similar threads
