In that vein... New study questions the packet switching theory of internet communications. 🙂
That would be nice - could probably be made very small. Then a companion box with a couple of compact class D amps....
How would my ears prove that the data was in fact in packets and took different routes (actually it rarely does as routers stop that nonsense).
You will know, if you just trust your ears. If the packets stay on a route with proper cables, there will be much more air in the sound, improved musicality and depth, and the timing will be much more engaging.
Well that will make my job easier, knowing what I am laying out does not exist and is just a figment of my imagination really helps... Maybe acid ought to make a comeback, I could then have a trip that involves some reality🙂
I did think of the class D amps in the same box as well, SMD where possible, multi-layer, short signal paths etc.
I did think of the class D amps in the same box as well, SMD where possible, multi-layer, short signal paths etc.
Ah confusion on my part. I use wireshark at work to fault find why things aren't working across 100 racks of equipment. That does have the added complication of MPLS which for home use gets terminated in the broadband box.
I may start a thread on the synergy of using different routers and how they all sound different to my ears...
Of course I am currently limited with the choice of cable and cannot understand why some Audio cable manufacturer has not done a silver wire based CAT5/6 cable
😕
Right, I am really disappointed with the choice as well. I tried the Denon high purity copper, cable but as soon as I visited a couple of Internet sites, it lost its purity.
What we really need is audiophile grade transcontinental fibre optic cable. Imagine how much money we could make from a pure audiophile sea cable from the US to Asia....
I'm sold, what does worry me though is what would happen if I plugged my Audioquest Ethernet cables in the wrong way round as it clearly states they are directional....
I did wonder, this will explain how my teenage daughters became possessed (like the Exorcist on steroids, but no pea green soup) I put the Ethernet cable from the music server the wrong way round to her Logitech Touch....
You should, as there are many professionnals doing perfect measurements but are still making bad sounding dac devices. Ah MPLS : there is a new possible QoS : a flag named "good sounding DAC" !
Thank you to twist the neck to audio myths, this thread sounds interrestings and very informative. We have many things to learn here for most of us. At least I learn a lot here.
So staying on the streaming Inside a PC you mean :
Jitter is created only from the output device ? I.e.: buffer of the USB hardware interface ? (managed by the chips on main mother board + RAM ?). Are there any other source of jitter in this processus ? changement of cunsumption with some others tasks (processus) which can affect the transport of datas from the buffer to the outside ? To rephrase it
: is the buffer is sensible to its outside environment : layout, PS, etc ?
How looks like the phase noise of a clock before and after the output buffer ? (Bad question because bad understanding ?)
Does a driver can influence things with USB as well (Linux vs W.) ?
So ok for the HDD/SSD/Linux/W for the cpu to the RAM : the same. If things are heared it is not here. (CPU not affected by time and tasks whic can affect the jitter to the output buffer.
Why are the SD cards reader have lower jitter than PC ? Is it the same but the little sd card reader is less Noisy so subjective sound quality perception is better ?
@ Marce,
Hello,
as you asked about the source about the SB Touch and the smps PS and the jitter, it is here (Art member) : The SB3/SPDIF output thread.......
With MCLK, does a 2 cm link with a matched impedance cable between a PCB and two vias near the dac chip can affect the jitter ? (plugs, soldering of the plugs ? continuous ground breaked despite a coax at 50 ohms or a very short distance : stacked 2cm ?
Ah, if you had more time for you and your hobby I should be very pleased to hear your version of the Subbu you begann one year ago with your own pcb drawing !
Yes the all in one device is hard to come. If you look at e.g. the work of Soren and the spidf input : you have still to manage the layout to put a pulse transformer, add resistor, a cap, etc... I surmise than made by some people like me with a poor experience those diy solderings and little layout, despite the rest of the good design can be a source of problem (well there is a fifo but no isolation here...)
But I don't know if it's a point of view but I find the pro and enthusiasts here to show more and more good digital projects. Maybe I'm wrong and we stay in the audio myths but with more sophistacated presentation ?
Well, didacted information is very welcome 🙂
Last edited:
It is entirely possible that a totally transparent system isn't actually considered "good sounding".
The opposite of the evil bit? 🙂
Jitter, by definition, is caused by variance in the output audio clock. If it is generated only by a local oscillator, it can theoretically (but to a very minor degree) be affected by power supply variation and noise. If it is generated from the input data clock, there are more sources of disturbance.
A buffer is a buffer. The timing of what goes out has nothing to do with the way it came in. Only relevant thing is the stability of output clock.
In most systems no, as all the driver does is put another bunch of data into the buffer when the I/O controller tells the buffer is starting to get empty. Actual output is done by separate controller.
Yes.
Do you have any measurements showing that to be the case?
Spdif is different from USB. Spdif doesn't have any way to synchronize source and DAC, so DAC pretty much has to follow timing from spdif source.
If it measures fine and sounds like crap, perhaps the wrong thing is measured. It is common practice to measure what is easily measurable instead of what is relevant.
EXAMPLE : DAC outputs a differential voltage equal to its Vref (or AVCC) multiplied by the digital sample value. If the sample value is zero, then the output is zero*, and you can be proud of a very low noise floor. Now if you do the same measurement with a constant digital value equal to full scale, then the DAC outputs its VRef directly. Then maybe you notice the VRef has a huge amount of noise or ripple in it, which would make the SFDR and other characteristics pretty bad, or raise the noise floor when doing full scale sine test. Talking SNR when your noise isn't additive is ... maybe a bit misleading.
* = it depends on the I/V CMRR also
For audio, jitter only matters at the converter where it gets embedded in the analog signal. So it matters only on the relevant pin of the converter chip.
As for digital transmission, if you have enough jitter to cause bit errors, well... you gonna have a bad day. But at audio bit rates, that's difficult to pull off.
Same for I2S... as long as it stays in digital form, and the clock-to-data skew and jitter are smaller than say, a quarter bit, then there is no loss of information.
Doesn't matter in asynchronous mode. And... how did you measure that ?
If properly done, should be ok. If you got common impedance coupling in the GND of your coax because your clock module has its own mains-powered psu, maybe you get a few ps extra jitter @ 50 Hz. You also get a nice antenna for free though...
Jitter.....
Understanding and Characterizing Timing Jitter - National Instruments
As to my own version of the Subbu DAC, I have been too busy doing PCBs non stop to play with a layout for myself, also the repercussions from showing it have made me put it on hold.... One day, though I would like to expand it to have a clad D output and some volume control etc. so I have a all in one unit (powered by an SMPS wallwart or Laptop supply
)
That threads a bit strange, there does seem to be a lot of mystique surrounding what is basically a simple digital interface (its not RF I do RF layout its a different game altogether)... A while ago I did some quite extensive simulations of SPDIF interface with different cable lengths, severe impedance mismatches, different rise times very hard to really mess up the signal (even though in some cases it looked rather bad as a waveform) and the waveform soon cleared up with some simple termination...
Again I do wonder some worry endlessly about streaming a few bits from source to destination, I bought a cheapish PC, big disk, plenty of RAM, it just sits there with no one logged on running Squeezebox Server hardwired to my wireless the squeezeboxes run wirelessly most of the time. I don't seem to have many problems 😀
Anyway back to the mundane for me tomorrow, a switch/router to lay out then some SMPS's for, well lets just say some very interesting analogue and digital ciruitry
Understanding and Characterizing Timing Jitter - National Instruments
As to my own version of the Subbu DAC, I have been too busy doing PCBs non stop to play with a layout for myself, also the repercussions from showing it have made me put it on hold.... One day, though I would like to expand it to have a clad D output and some volume control etc. so I have a all in one unit (powered by an SMPS wallwart or Laptop supply
)That threads a bit strange, there does seem to be a lot of mystique surrounding what is basically a simple digital interface (its not RF I do RF layout its a different game altogether)... A while ago I did some quite extensive simulations of SPDIF interface with different cable lengths, severe impedance mismatches, different rise times very hard to really mess up the signal (even though in some cases it looked rather bad as a waveform) and the waveform soon cleared up with some simple termination...
Again I do wonder some worry endlessly about streaming a few bits from source to destination, I bought a cheapish PC, big disk, plenty of RAM, it just sits there with no one logged on running Squeezebox Server hardwired to my wireless the squeezeboxes run wirelessly most of the time. I don't seem to have many problems 😀
Anyway back to the mundane for me tomorrow, a switch/router to lay out then some SMPS's for, well lets just say some very interesting analogue and digital ciruitry
- Status
- Not open for further replies.