Its not the music data embedded in the packet its loosing one of the surrounding bits that causes the problems, the bits that control the data packet, they are just as likely to be corrupted with a bad interface and these bits will be noticeable as control of the contained data can be lost, I believe...
edit:
having read the rest of the thread and more its going to have to be a pretty bad cable or interface for it to cause drop-outs.
Last edited:
The USB interface has been measured for many years... The packet noise that JS found was on the internet already as was the rest....
An awful lot of all this is to create a market to sell into, like the USB hub craze...
its generalised statements such as this that bug me (from the Regen marketing), implying that every other USB interface is flawed apart from theirs...
Of course if you want to design your own USB audiophile hub, then you can start with this module...
https://www.verical.com/pd/microchip-technology-development-kits-and-tools-evbusb2412-758822
The schematics and user manual are on line so not hard to put together a USB hub is it....
http://download.siliconexpert.com/pdfs/2010/7/7/3/57/53/571/smsc_/auto/evb2412user.pdf
For what it’s worth, the bit-perfect test is explained on our website:
EC designs - XTOS Info
click on “Bit perfect test explained”
In order to check if a CD player offers bit perfect playback:
Download the bit-perfect test CD from our website and burn to CD. It contains a 0dB sawtooth test signal, so make sure the power amps are switched-off or set analogue volume control to minimum.
The 44.1/16 bit-perfect test takes approx 1.5 seconds to complete (44,100 samples / second can be checked).
With 24 bit test files it takes much longer. Wit say 192/24 we can check 192,000 samples per second. There are total of 16,777,216 samples to be checked. This takes 87 seconds.
Put the CD in the CD player to be tested.
Use suitable converter to obtain a Toslink interface (Coax to Toslink for example).
Connect the CD player Toslink output to the Toslink -input- of a mac or soundcard
Use a suitable application to route the Toslink input to the USB output.
Connect the USB output to the XTOS. XTOS Toslink output is not connected.
Play the test CD and if playback is bit-perfect the white LED comes on.
CD-player containing bit-perfect test CD -> (S/PDIF to Toslink converter) -> Toslink input (mac or soundcard) -> USB out -> XTOS.
This way we can test both, S/PDIF and Toslink sources.
For USB sources we simply connect the XTOS to the USB output.
This way, we can test all S/PDIF, Toslink, and USB audio sources.
Bit-perfect test was primarily intended to proove if the audio data (on the file) was tampered with -before- it is sent to the USB interface.
ah you've fixed the link. Explained is a littl exaggerating tho.
Let me try and rephrase it. Bitperfect test was developed by the bitperfect software company to sell their software?
Let's call it what it is- shilling his products.
If there's missing bits, this shows up in conventional distortion tests.
Ok ok - my take was that you claim *playback* bit-perfectness when it is only up to your box thingy. A lot can happen beyond that point...
//
There is a lot of nonsens in this discussion. But to answer the original question. The issue that you have is probally the power over usb, that can't deliver the juice the audiophile grade analog components of your convertor need, that's why it sounds better with the external psu. Also a laptop psu is in no way audiophile, so the dac is powered with a cheap build low quality psu on audiophile level...
The signal in usb connections is digital, not analog. If the signal is disturbed by anything, the receiving electronics will not accept it and go in error. The integrity of the audio is not damaged by the transfer like what is possible with analog connections as filters in the receiving part (the dac) take out any distortion when converting the signal to analog after the transfer trough usb. If those filters can't handle it, they won't pass nothing. It's ony in the analog part of the signal and during conversion that there is a change of signal possible, not in the digital part.
The signal in usb connections is digital, not analog. If the signal is disturbed by anything, the receiving electronics will not accept it and go in error. The integrity of the audio is not damaged by the transfer like what is possible with analog connections as filters in the receiving part (the dac) take out any distortion when converting the signal to analog after the transfer trough usb. If those filters can't handle it, they won't pass nothing. It's ony in the analog part of the signal and during conversion that there is a change of signal possible, not in the digital part.
No, you can have non-bit perfect playback without -any- dropouts.
When you use the digital volume control on the computer, most sample values transmitted by the USB interface no longer match the sample data contained in the audio file. The bit-perfect test will then fail.
When you enable the sound enhancer or equalizer on the computer similar thing happens. The spectrum of the audio signal is changed in favour of say classical music or vocals. No drop-outs, transmitted sample values no longer match the sample values contained in the audio file, bit-perfect test fails.
When the computer hardware, OS or DAC drivers have been updated there is a small chance for errors and related non-bit perfect playback.
The bit-perfect test feature costs us one LED, one resistor and some code. It was added as a free extra feature in order to rule out bit-perfect related sound quality issues. It saves us a lot of time answering emails and phone calls related to this problem.
XMOS produces a similar converter:
XK-USB-AUDIO-HPA XMOS | Programmers, Development Systems | DigiKey
If there would be no audible degrading when tampering with the sample data from the original audio file we wouldn’t have bothered.
The problem is that such errors often go unnoticed at first, the sound quality degrades marginally and we quickly get used to it. The bit-perfect test rings the alarm bell when such things happen so we know something is wrong and fix it.
Some bit-perfect playback issues could be measured by using suitable measuring equipment but not all audiophiles have such equipment or know how to use it.
The XTOS on the other hand offers an easy test method that can be performed during listening, simply by playing the bit-perfect test. The bit-perfect test is always active and runs on a separate CPU core of the XMOS, so it is instantly available at all times. So if you suspect something is wrong when listening to your audio set you can quickly run the bit-perfect test to make sure the source is not causing it.
One could argue that a lot of things can go wrong after the signal leaves the XTOS. We also checked that by using a second XTOS to check the first. Then both white LEDs should come on. In practice, communication errors with both, USB and S/PDIF are -very- rare when using interlinks that meet specifications.
When data corruption occurs due to interlink related issues, it usually translates to audible drop-outs.
Last edited:
What!!!!!
Firstly there is a limit to the power USB cab supply, in the USB spec, only a NORON would design a USB DAC (supplied via USB) that drew more current than can be supplied by a USB port...
The second paragraph is confusing me somewhat...
It is the transport of data through the USB cable I and this discussion is primarily addressing, not deliberately changing the data by using processing of some sort.... The data delivered from one USB Phy to another USB phy (or one transceiver to another), but we are talking physical level data transfer and changes to the sound caused by cables mainly.....
What do you mean by good error correction ? There's no error correction in usb audio protocols at least (missing data will not be resent).
He has an €875 euro magic preamp to sell. EC designs - Mosaic VC. I have to say an interesting solution, even if not actually a problem to solve. But having volume control from the PC whilst the 'bitperfect' light stays lit is good marketing message...
Don't use volume controls myself, I have found they all degrade sound to a greater or immense degree, I have a set of tailored earmuffs that have calibrated dB losses depending on the thickness of the knit and the type of wool used, so if I want a lot of attenuation I use an angora sheep blend knitted with a UK size 0, when I want to go louder I have a pure silk set of earmuffs hand crocheted by some old ladies in Gozo. Subjectively removing the volume control from my system has been an eye opener, blacker silences, less screechy highs, warm ears etc...
Ummm. Lot's of USB products draw the maximum allowed power especially during power up. In the industrial world USB cables with larger gauge power wires are available. These losses can limit you to very short cables.
I have seen countless problems using "certified" cables to power devices that meet all the requirements. Additional problems are due to power losses in the internal pigtails feeding front panel USB connectors, losses in the USB PTC fuses, crappy or failing capacitiors on the motherboard or power supply ...etc.
Yes there are lots of badly engineered products out there. However a DAC does not need to suffer from any of those, and many don't.
I am not sure what side you are on here. the fact that you have an issue when a capacitor fails is exactly that, a failure. Not sure how that relates to the paranoia of the audiophile?
I am not sure what side you are on here. the fact that you have an issue when a capacitor fails is exactly that, a failure. Not sure how that relates to the paranoia of the audiophile?
I am just supplying a different take on the subject. For USB 1/2, power problems are much more common than data integrity issues. Another issue is ground loops and conduced noise. If your DAC feeds an external device like an amplifer you will have "noise" issues.
The brutal ripple currents that capacitors have to endure in the multiple regulators in a PC will degrade any electrolytic capacitor. This will cause ripple on the power feed. This ripple plus the DC losses can cause issues. The older the PC or the use of cheap capacitors will guarantee excess ripple. You won't know your capacitors are failing until you start have issues like BSOD or resets.
Do you actually mean 'will have noise issues' or just 'may have noise issues'. If what you say is true then a cheap USB hub will solve that, no magic audiophilia required.
But, other than gross cases no one seems to have captured subtle changes in sound quality due to anything that you mention. It works, or it sounds horrid. no veils to lift or sheen to polish. Happy to be proved wrong, but so far nothing.
But, other than gross cases no one seems to have captured subtle changes in sound quality due to anything that you mention. It works, or it sounds horrid. no veils to lift or sheen to polish. Happy to be proved wrong, but so far nothing.
I'm saying with all the high frequency and AC current slugs traveling inside a PC, provide an alternate "path" and some of that junk will leak via that new path. A USB device has a direct connection to PC power and ground. Anything you connect to it that is not floating can provide a path and introduce a ground loop. In many cases this junk is inaudible but is measurable. I have had issues with AC hum as well as other audible junk especially if you crank up the volume.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.