Question about using USB for data input and out to DAC

[Mike B.]

I read somewhere that it was preferable to use stored data in via ethernet from a NAS over direct in using a USB port. Anyone have any insight into if this is true or false?

[aKnyght]

two issues that come to mind are:

transport of data over ethernet is error checked + corrected -- over USB it is not for most streaming applications. not sure how material a concern this is, especially for audiophile systems where electrical noise is kept low.

ethernet can be used to transfer data over long distances -- 100 meters for CAT6. USB is for short distances -- less than 2 meters.

for all you will ever want to know about USB -- and more: https://beyondlogic.org/usbnutshell/usb1.shtml

[Mike B.]

I set up a older router close enough that I could use a 25 foot CAT 6 cable I had cryoed at Jena. This router has a USB port and a program called ReadyShare. I had it set up with my usb data drive which sent the data to the NUC over that cable. It worked but there was a more significant delay and my remote control program on a tablet was very slow to respond with errors. So I switched back to the USB drive using a .75 meter Audioquest Carbon cable and it works well. Ports on the NUC computer are USB 3.

[Puma Cat]

I read somewhere that it was preferable to use stored data in via ethernet from a NAS over direct in using a USB port. Anyone have any insight into if this is true or false?

i'm struggling to understand the question.

"Preferable" in what way? I don't understand what the requirement(s) are that underpin this question.

So, lets see if we can formalize it a bit: what is the problem you're trying to solve or outcome you're trying to achieve?

[Mike B.]

I read that is was not a good idea to use USB ports to do both data in and out to the DAC at the same time. I think it was written by Steve N. of Empirical Audio. I tried to find the thread at another site with no success. I am trying to understand why that would be a bad idea? I assume it causes sonic degradation. My NUC has no additional ports like Firewire or Spdf.

[Puma Cat]

I read that is was not a good idea to use USB ports to do both data in and out to the DAC at the same time. I think it was written by Steve N. of Empirical Audio. I tried to find the thread at another site with no success. I am trying to understand why that would be a bad idea? I assume it causes sonic degradation. My NUC has no additional ports like Firewire or Spdf.

I still don't understand what you're specifically trying to do and what you think the problems you're having are.

[Mike B.]

I still don't understand what you're specifically trying to do and what you think the problems you're having are.

Then you can't be any help. Move along

[Randy Myers]

In my configuration I use USB for data in on my DAC. The signal coming from my server to the Recovery and then to the DAC. Output from my DAC is analog, ala, why it is called a DAC (digital to analog converter). The XLR outputs from my DAC go to my pre-amplifier.

I personally have no use of have a USB or other digital outputs on my DAC.

[Mike B.]

I was hoping there was a digital engineer on site. This is more of a specialist question for someone who designs or works with USB circuits in audio related applications. I will see if i can reach Steve by phone this morning. I will also try to contact Ted Smith. Thanks to those who posted.

[NekoAudio]

USB Audio uses a different variation of the protocol than USB Data.

USB Data has packet error detection and retransmission, just like Ethernet or your hard disk and provide delivery and data integrity guarantees. USB Audio, designed as a real-time data stream, does not bother with retransmission because it's considered better to have the correct playback for "now" instead of falling behind. Imagine, for example, if you were watching a live stream on your computer but over time errors in the audio transmission to your USB audio device caused the audio to become farther and farther out of sync. It's the same reason live video (e.g. remote news correspondents) does not retransmit and might freeze and then jump to the correct time while streaming video (e.g. Netflix) will pause and rebuffer if it needs to.

Because USB audio is designed to be real-time and does not retransmit in event of errors, if you are asking your audio device to do too much (or if your computer is overloaded) then you will have problems with the data delivery. That being said, there are many computer USB audio interfaces that are capable of doing a lot at the same time over a single USB connection, since USB 2.0 is much, much more than enough for stereo lossless audio.

In other words it shouldn't really matter as long as the product engineers did their job right.

With respect to your original question:

I read somewhere that it was preferable to use stored data in via ethernet from a NAS over direct in using a USB port. Anyone have any insight into if this is true or false?

From an engineering point of view this is false if you are comparing pure data transmission in both cases. For example loading a file from the NAS versus loading a file from a USB connected hard disk or flash drive.

However if you are considering Ethernet or USB data transmission versus USB audio transmission, it's no longer comparing apples to apples. And the above concerns with data loss apply to USB audio.

[Randy Myers]

By far, your best way of storing data for playback is internal within the computer/server. The question of transporting the data becomes moot since at the worse SATA does not have any issues and at the best M.2 drives connect directly to the motherboard similar to RAM. The only use of USB in this case is to transport between the server and the DAC. Definitely no issues doing this for short runs. Additionally, some USB re-clockers have also shown to have positive affects going between the server and the DAC.

[Mike B.]

Thank you NekoAudio. This is what I was hoping to read. I use the memory playback function provided by jriver 25. So, the music storage hard drive loads data at the start. Do you have any experience with the large number of USB add on devices to improve USB performance?

[NekoAudio]

Thank you NekoAudio. This is what I was hoping to read. I use the memory playback function provided by jriver 25. So, the music storage hard drive loads data at the start. Do you have any experience with the large number of USB add on devices to improve USB performance?

I'm not 100% sure and it's probably best to ask on the JRiver forum, but having JRiver load all the data into memory just means the full audio data is in RAM on your computer (where JRiver is running). That's different from having the full audio data in memory on your playback device. Old iPods would do this too, because it minimized hard disk spin time to save battery life.

If your computer is then connected to a DAC over USB for playback, you're using USB Audio for that communication.

In the majority of cases you shouldn't need any USB reclockers or isolators or similar devices. You could always try one, especially if your environment is noisy or if your computer isn't so great or if your DAC exhibits sensitivity to the USB connection/data, since there are some cases where it might help. But the first thing I'd recommend trying, out of anything, is to use a USB cable with ferrite rings to see if that makes any difference.

[aKnyght]

...In the majority of cases you shouldn't need any USB reclockers or isolators or similar devices. You could always try one, especially if your environment is noisy or if your computer isn't so great or if your DAC exhibits sensitivity to the USB connection/data, since there are some cases where it might help. But the first thing I'd recommend trying, out of anything, is to use a USB cable with ferrite rings to see if that makes any difference.

+1

a few months ago i experimented with a USB cable used with a professional, ultra-high-resolution film scanner which has ferrite rings at both ends... wow!! night and day results with a HUGE drop the noise floor. it is not coming out of the system.

[Audioseduction]

I'm not 100% sure and it's probably best to ask on the JRiver forum, but having JRiver load all the data into memory just means the full audio data is in RAM on your computer (where JRiver is running). That's different from having the full audio data in memory on your playback device. Old iPods would do this too, because it minimized hard disk spin time to save battery life.

If your computer is then connected to a DAC over USB for playback, you're using USB Audio for that communication.

In the majority of cases you shouldn't need any USB reclockers or isolators or similar devices. You could always try one, especially if your environment is noisy or if your computer isn't so great or if your DAC exhibits sensitivity to the USB connection/data, since there are some cases where it might help. But the first thing I'd recommend trying, out of anything, is to use a USB cable with ferrite rings to see if that makes any difference.

I too use JRiver as my playback software. I have it set to load the entire track into RAM first then start to play. Works GREAT!

[Puma Cat]

+1

a few months ago i experimented with a USB cable used with a professional, ultra-high-resolution film scanner which has ferrite rings at both ends... wow!! night and day results with a HUGE drop the noise floor. it is not coming out of the system.

I wouldn't be putting any ferrite cores on any cables used for audio. Ham radio, yes, but not audio (stereo).

[aKnyght]

I wouldn't be putting any ferrite cores on any cables used for audio. Ham radio, yes, but not audio (stereo).

YMMV ...but why not?

the SQ was nicely improved when i tried a usb cable with ferite rings - confirmed with a simple blind a/b test on a family member. there seems to be a well established noise reduction property to them.

https://en.m.wikipedia.org/wiki/Ferrite_bead

[Puma Cat]

YMMV ...but why not?

the SQ was nicely improved when i tried a usb cable with ferite rings - confirmed with a simple blind a/b test on a family member. there seems to be a well established noise reduction property to them.

https://en.m.wikipedia.org/wiki/Ferrite_bead

Anything along the lines of ferrite cores, inductors, chokes, etc., used for noise reduction impact dynamics, both macro and micro-dynamics.

Take at look at the very best USB cables on the market e.g. Audioquest, Shunyata, Nordost, Crystal Cable, Wireworld, etc. They are built to provide absolute best performance regardless of cost. Virtually none of them use a ferrite core on the cable.

[aKnyght]

Anything along the lines of ferrite cores, inductors, chokes, etc., used for noise reduction impact dynamics, both macro and micro-dynamics.

Take at look at the very best USB cables on the market e.g. Audioquest, Shunyata, Nordost, Crystal Cable, Wireworld, etc. They are built to provide absolute best performance regardless of cost. Virtually none of them use a ferrite core on the cable.

fair enough ...but, given my empirical experience with ferrite cores positively impacting dynamics, i would not dissuade anyone from giving them a try -- and, i think that is all that is being suggested here.

one can always resort to a $2k USB cable if unhappy with the results.

[Randy Myers]

In the majority of cases you shouldn't need any USB reclockers or isolators or similar devices.

I have used a Wyred 4 Sound Recovery unit since they were released. I actually grabbed one from EJ a couple of days before they started shipping their first batch ever. I would not say that it is 100% necessary, but I also say it has made a nice improvement. My computer and my USB cables are both very solid. In my view, no matter what level equipment you have the Recovery makes a nice improvement and for its price you simply cannot go wrong. By the way, I use AudioQuest Coffee USB cables. The only real down fall that I have seen is that two USB cables are required, therefore, with good cables this can get a bit pricey . I also recommend an iPower power supply for the unit.