|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
24.6.199.104
In Reply to: Re: Could a USB DAc like the Crimson get input from Squeezebox? posted by laznos on December 31, 2006 at 00:25:37:
First some comments on terminology: all the USB audio modes are isochronous, this just means that the bandwidth for the pipe is determined in advance and reserved for that pipe on the bus. There are three modes: synchronous, adaptive and asynchronous. For both synchronous and adaptive the sender is in control of ultimate clock rate, the receiver has to "sync" to the transmitter. In the asynchronous mode the receiver is in control, telling the transmitter to speed up or slow down to match the receiver's rate. Almost all USB system's today use the adaptive mode.With the asynchronous mode its possible to have a very low jitter system by using a very low jitter clock right at the DAC. Unfortunately its not always true, there are a couple asynchronous implementations that do NOT produce very low jitter, it wasn't part of the design spec.
On S/PDIF and USB the jitter issue is complex. The common S/PDIF receiver chips produce a fairly low level of jitter, but I and others find the sonic impact of this jiter to be really anoying. The common adaptive USB interfaces produce more jitter, but it does not seem to be nearly as anoying as the S/PDIF jitter. Both the interfaces can be effected by noise, impedance mismatches etc on the physical interface (the cable), with S/PDIF seeming to be more sensitive to this than USB. Again this is because the DAC clock is derived from the data stream by some form of PLL, the more noise or jitter on the interface the more you will have on the output of the PLL.
The clock on both systems can be run through various "jitter reduction systems" such as special PLLs that are less sensitive to input noise and jitter, ASRCs etc. Some of these work very well and some do a good job of reducing jitter but change the data, its a very large space of tradeoffs.
The squeezebox is an interesting point in the spectrum, it has a low jitter clock directly controlling the DAC without any PLL or other clock circuit in the path, but its external and intrnal power supplies are not the best in the world. The result is a clock jitter that is about on par with the best its posible to get out of a VERY good S/PDIF or USB receiver fed through a very good VCXO based PLL. The analog side of the SB is not on par with the best output stages on "audiophile" DACs. The S/PDIF output of the SB is very good, but you still have to run it through a receiver in an external DAC, with all but a very few DACs the recovered clock jitter in the DAC is going to be significantly worse than what is in the SB. BUT the analog stage and power supply in the external DAC is probably going to be significantly better.
Then there is what I do with an SB3, hook it up synchronously with a clock from an external DAC. Have a very low jitter clock in the external DAC and feed it into the SB instead of its own clock. I then take the I2S signals fed to the SB DAC chip and send those to the external DAC and reclock them with the low jitter clock right before its DAC chips. This gives extremely low jitter and sounds incredible. To my knowledge only a few people have done this and nobody is doing this as a commercially available mod.
You CAN do this with a transporter, it will accept an external word clock which allows it to be synced to an external DAC. You still need to send the data over the S/PDIF link, but you can throw the recovered clock away and use a low jitter clock in the DAC to clock the DAC chips. This is almost as good, but a little bit of the S/PDIF nastyness still leaks through somehow. (I'm not sure what the mechanism is)
If you are not going to do what I did I think modying a SB to add a very low jitter clock and bypassing the whole analog stage with a much better analog stage gives very good results. Unfortunately that analog stage cannot fit in an SB3, so you need to build a separate box and build the SB3 into IT.
There are a lot of different options now and unfortunately not an absolute clearly better than anything else implementation. There are a lot of different compromises, some of which sound way better to me, but others prefer different compromises.
What I CAN say is that REALLY good sound does not come cheap, no matter what approach is taken there is no "magic bullet" that will produce really top notch sound for a couple hundred dollars, anything in that price range is going to be compromised in some form or another, the task is finding which compromises you can live with.
What I'm listening to right now cost well over $1k in parts and there are very few expensive parts in this, it could be much worse.
The best sounding digital devices seem to be those that have a good balance between doing the digital well with low jitter and also doing the analog side well, as well as the power supplies for both. Getting all that to work synergisticly is not easy, but its not black magic either.
Now after all that, here is the answer to the original question, yes you CAN hook up a USB DAC to slimserver network, I did it for about a year. I had a small fanless computer running linux that I ran softsqueeze on, it talked to the slimserver in the other room and sent the audio data over USB to the USB DAC. This actually does work quite well but is probably over kill if you are just using a single "output" system. I did it this way because I was using multiple SBs in the house run off one server. With the current system I'm using an actual SB3 interfaced to the DAC.
Its probably possible to modify a crimson to do the same thing I'm doing with the SB3, but I'm pretty sure that would void the warantee and Gordon probably wouldn't like you doing that either. Whether it would sound any better is an interesting question. Gordon is getting very good at the synergism I was talking about earlier, there is a good chance that any changes to the design will make it worse not better.
Well thats it for now.
Follow Ups:
John,Thanks for your comprehensive write-up.
Your work on connecting a DAC to the Squeezebox via I2S and connecting back via a low jitter clock makes a lot sense to me.
Conceptually I also like the idea of using Ethernet instead of USB. It seems a lot more flexible that my digital audio box is a real node on a standard network.
The challenge for us non-digital design engineers is implementing this in as painless a way as possible. I did take a digital design class in college but don't have the equipment (or time) to root around surface mounted circuit traces looking for I2S signals. I also know from my software engineering background that there are grounding issues, clock issues, noise issues, and probably a plethora of other issues when it comes to digital design.
Hopefully at some point this year there will be enough DIY products and forum information on wiring, etc. for me to be able to do this project.
For example, Guido Tent recently came out with a DIY CD player that uses a NOS PCM1704 chip. The first product is a full CD player but he intends to offer the various boards separately. So I could potentially use his I2S input DAC board and his analog output stage board. That would help solve the DAC end of the project.
The relatively inexpensive and good sounding Lite DAC50 or DAC60 also seem like excellent platforms for an IS2 Squeezebox/DAC. But I would need explicit instructions on how to wire into the Lite DAC I2S, install a low jitter clock, etc.
That leaves some messy work on the Squeezebox end, but perhaps with good instructions I could handle that end too.
I really like certain aspects of the Squeezebox. I think the user interface on the hardware is terrific. I like the network connectivity. Running a stock Squeezebox with an upgraded linear power supply through SPDIF into my Lite DAC50 just didn't cut the mustard, however.
Anyway, just thinking aloud here on how to proceed! You have whetted my appetite for what sounds like a very good solution.
OK. I haven't perused all of the threads on this topic, but how does the Airport Express (AE) compare to the SB? Can you achieve the same good results of the SB by using a USB modded AE? THANKS.
no such thing as a USB modded AE. nor for that matter a USB modded SBboth are network devices that live on ethernet or wifi (802.11) networks.
think of the AE as a receiver (radio receiver which is all that wifi is). one thing it can received is an audio signal which can be then sent along via a mini Toslink adapter built into the plug to a DAC.
in contrast the SB has a DAC built in as well as a remote control and display that allows you to access your music library from wherever you have an SB. In comparison, the AE is a one way device that will only play what it receives and provides no way to access the hard drive.
Xmas,The AE uses a USB controller for the printer port and the AUDIO port. Basically you can remove the AUDIO board which sports a TI/BB PCM2706 and solder on a be-headed USB cable and route it out the same place that the 1/8 jack use to be.
I did this ounce and it worked fine.
Better than a computer... NO WAY.
I did this ounce and it worked fine.Better than a computer... NO WAY.
Gordon,
This post is made possible by the generous support of people like you and our sponsors: