|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
97.89.28.78
Besides the bit rate capabilities of the sound card, does it make any difference which card is used if your going digital out to a DAC?
Follow Ups:
What are the sound characteristics of jitter? How can high or low jitter be translated to sound characteristics? I am sure I have heard gear with very low jitter and with high jitter before but never knew what to attribute the sound too. So how can I listen and determine, ahhh this setup needs something for jitter...
I have also read that jitter is not an issue these days, hence the absence of anti-jitter boxes of yester-year....any truth to that?
I would guess that 99% of setups do not have anything to specifically address jitter, from the way it sounds everyone who listens to digital music [should] have a device to handle jitter.
I use an optical connection from my soundcard to 'galvanically isolate' my audio gear. I do not worry about jitter from the sound card because the optical signal feeds a Theta TLC which is used to reclock the signal and reduce jitter. A digital coaxial connection is made from the TLC to the Channel Islands VDA2 DAC. I use JRiver and an EMU 1212M with ASIO.
IMO, it is futile to worry about jitter from a PC sound card. As far as I can ascertain, they are all pretty much the same with respect to jitter on the digital out.
In general I find jitter to cause a loss in "inner detail" which usually relates to a "flatter" sound. It looses the "liveness", it becomes boring. Sometimes I find jitter also effects bass significantly. This is strange because I would expect high frequencies to be more susceptible, but I frequently hear a significant improvement in bass articulation when decreasing jitter.
Its hard to listen to a piece of gear and say "thats got high jitter" just by listening, because many other things can cause similar sonic effects. By building my own gear I've been able to do quite a few tests where I can hold everything equal except change jitter and can definately hear major improvements in sound by lowering jitter. But frequently similar changes can be had (for example) by upgrading the power supply of the preamp.
Another issue is all jitter is not the same. The spectrum has a lot to do with it. I have one receiver with 200 ps of jitter that sounds significantly better than another with 50ps, BUT the spectrum is radically different between them. Thus just picking the lowest published "jitter number" will not gurantee the best sound. Unfortunately you still have to listen!
I also have large misgivings about any system going through (the never to be sufficiently damned) S/PDIF interface. This is so frought with peril that its almost impossible to get a really good clock out the other end. This makes me very suspicious of "jitter reduction boxes" that output S/PDIF, they STILL have to go over the interface. Yes if the "transport" is really bad they can help, but its never going to get really good.
There are a few DACs that actually do an extremely good job of cleaning up the clock from a S/PDIF interface, but they are very few and far between. These are the only ones I would consider using with S/PDIF. And even then the jitter from the recovered clock has a tendancy to make itself heard through mechanisms such as ground noise and supply contamination, it takes a lot of work to keep it out of the final result.
John S.
Sounds like Jitter is something we all have to live with so some degree. I dont know how jitter is introduced with a computer based system when the material source comes from RAM. Even still, it sounds like jitter can be embedded in the source material from the start being digital in nature....what do you think? Possible....
Computer-based systems are not immune from jitter more than any other digital system.
Jitter starts at the master clock, either in the PCI card, USB converter, WiFi converter or the computer clock plus software. Then it gets worse as it is divided, buffered and converted to S/PDIF (unless it stays I2S). Then, any cabling and buffers at the other end as well as poor power delivery at teh D/A chip also add jitter. One must address ALL of these to have a truly low jitter digital system.
Steve N.
Wow, it seems like jitter is impossible to get rid of....
Not really. I dont listen to jitter, ever. You just need to make informed choices. S/PDIF cable for instance should be either 6" long or 1.5m long.
Steve N.
S/PDIF cable should be either 6inches or 4.5ft???
Now I'm really scratching my head.....
Read this paper:
http://www.positive-feedback.com/Issue14/spdif.htm
.
Here's a link to a Stereophile Test CD whose description contains "hear what jitter sounds like".
Thanks....but somehow paying $9.95 to hear what something sounds like that I [dont] want to hear seems strange.
Besides, if I have it I am listening to it for free...:-)
I am collaborating with highdeftapetransfers.com to generate a test track of about 20 seconds to demonstrate jitter. The Track is a transfer to .wav of track 10 of Nora Jones "come away with me":
1) first transfer is the original commercial CD
2) second transfer is a CD-R rewritten by me on a low-jitter CD writer
Both tracks will be played back on a CD player and then redigitized using their A/D converter (professional).
The tracks wil be small enough to send by email.
Steve N.
Thanks that makes me laugh.
In my limited experience, yes, and many here with rather more than me also emphatically say it does.
I had to ditch an old Midiman Dio2448 because I couldn't get completely drop-out free SPDIF out, despite the fact that the bloody thing was expressly designed and sold for the purpose, hence D(igital) I(nput) O(utput).
I replaced it with a Delta 410 which solved the drop-out issues but, everything else being equal (which it was), definitely sounds different, and at least to begin with I felt in some ways possibly not as good as the Dio.
It's a questtion of jitter,(ie. clock-precision), and not just how much, but it's frequency-distribution, frequency-correlation and all that.
You might want the DAC to be galvanically isolated from the computer when using coax or aes/ebu. Some soundcards and/or DACs are transformer-coupled. Also active devices like jitter boxes may provide isolation.
audio gear that is galvanically isolated from other pieces of gear in the chain.
Is galvanic corrosion really an issue?
see link for definition of Galvanic Isolation.
Ahhhh, okay thanks!
Unfortunately for the consumer, most DACs don't have any jitter rejection in the audio band, so anything you can do t lower the source jitter will help. Most soundcards don't have dedicated clocks either, instead relying on phase lock loop generated clocks.
I guess no such thing as a perfect anything, but the Lavry is used in professional studios and should have some jitter rejection.
Professional gear has not impressed me yet. I have modded a few pieces. Nothing magic there that I can find, in fact there seems to be a lack of general knowledge of what are good parts (mostly capacitors) to use.
The Lavry does have some nice features, and the Crystallock was a nice feature until it got removed due to technical problems I believe.
Is that actually true. All my sound cards seem to have a crystal oszillator on them. What is that being used for if not for clocking the output?
I think the question is whether that clock has low jitter and for the mass market cards the answer is probably no.
USB sound cards need to use PLLs for clocking unless they implement async mode like the latest Wavelenght DACs from Gordon.
Cheers
Thomas
One for every possible sample rate? Just one for 48K? Some do have dedicated oscillators for the main sample rates, but many don't.
The Wavelength DACs still use PLL circuits, in fact there was a thread here announcing that he just updated the firmware to control the PLL dynamics for lower jitter. But that's really a different topic. I'm not sure how the conversion clocks are implemented in his designs.
Gordons latest Crimson update uses a non-PLL asynchronous USB protocol.
Things are changing so fast in this area, you have to read everyday to keep up.
Steve N.
He was using adaptive endpoints and updated to asynchronous endpoints. Can you run the TAS1020 without the internal PLL circuits? I think you still need to sync to the incoming data, but going to async endpoints you can use an external oscillator in place of the adaptive clock generator for the codec, which I think is what he does in some of his designs. I apologize for the misinformation since he was actually talking about the PLL control in his adaptive designs, not the new one.In any case, this is kind of off topic. What about the question of soundcards? Don't they often use PLL generated clocks to control the SPDIF output stream?
Yes, you can run the TAS1020 without using the PLL.
As for sound cards on the PCI bus, I believe they do it both ways, with or without PLL. Depends on the card. Some even have word-clock input so they could be driven by a reclocker or DAC with word-clock output.
Still sounds like we're talking about different things, so I'll just leave it here since it doesn't pertain to this thread topic. In the previous post I was referring to the PLL for the USB bus, noting that the adaptive clock generator for the codec would be out of the circuit for asynchronous isochronous operation used in the new Wavelength design. But I think you still need to sync to the incoming audio data on the USB port since it isochronous. See below from the datasheet, bold highlight mine. Whether these derived clocks have any affect through interaction on the external precision clock driving the codec would probably be somewhat dependent on circuit layout, but I'm sure the Wavelength products are designed very well in that regard based on the glowing customer feedback.
2.1.2 Clock Generator and Sequencer Logic
Utilizing the 48-MHz output from the PLL, the clock generator logic generates all internal clock signals, except for the codec port interface master clock (MCLK) and serial clock (CSCLK) signals. The TAS1020A internal clocks include the 48-MHz clock, a 24-MHz clock, and a 12-MHz clock. A 12 MHz USB clock is also generated. The USB clock is the same as the internal 12-MHz clock when the TAS1020A is transmitting data, but is derived from the data when the TAS1020A is receiving data. To derive the USB clock when receiving USB data, the TAS1020A utilizes an internal digital PLL (DPLL) driven from the 48-MHz clock.
The PLL is still used to save the data into the FIFO of the TAS1020, but the output clock is completely independent and non-PLL. Pointers are used to gate the source stream rate. The output rate is fixed.
Ask Gordon for more detail.
I thought there must be some kind of feedforward mechanism for the DAC to make the source adjust its sample rate to match the fixed clock in the DAC, but haven't looked into it enough yet. Thanks for the info, I guess it's all in the endpoints and the initial synchronization. Interesting stuff. I obviously should've done some more research before posting :-)
Right, I am sure a computer has many PLL in all sorts of places ....
The material point is that the conversion clock is not derived from a PLL but directly from the oscillator similar to what happens on most sound cards.
You also probably know that you do not necessarily need a crystal per frequency but you can start with a common multiple and divide it down. However, many cards have two. One for frequencies that are a multiple of 8000hz and one for frequencies that are multiple of 22.100Hz.
I just picked one from the shelf and the two crystals are 22.5792Mhz and 49.1520Mhz. This specific card does only support audio sample rates of 44.1Khz and up.
Cheers
Thomas
The jitter from the S/PDIF output make a big difference. Those cards with low-jitter on-board clocks and good implementations of S/PDIF (impedance-matching and fast edge-rates) will have lower jitter.
The S/PDIF cable you use will also matter. The quality and length is important. This can also add jitter.
Steve N.
If you want to use ASIO, you will need a souncard that supports ASIO.
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: