Then I wonder why every discussion on digital music reproduction via usb - both hardware and software is related to jitter and its reduction?
Cheers,
Sid
CD vs FLAC. Which is better?
- Thread starter Rupam
- Start date
Then I wonder why every discussion on digital music reproduction via usb - both hardware and software is related to jitter and its reduction?
Cheers,
Sid
i think you are getting me wrong.
i never said jitter is not an issue for audio.
The genesis of this whole discussion is
A Rs1000 cable sounds inferior to a Rs 20,000 cable. which i tend to disagree.
as good quality Rs 1000 cable, is not expected to cause any jitter. That too 100% of the time and hence cause a bad quality audio listening to a whole song.
Comparing cables at affordable level like few thousands, i have not seen any difference.
Now someone can come and say try a 10K/20k cable, which i wont even try.
But if someone with loads of $$$$ can spend such exhorbitant amount for a USB cable and also have ears to notice the difference, if at all it is there.
I have no issues, spend your money , as you may be right and i may be wrong
i will end my debate here
sound_cycle
Well-Known Member
Hi,
While I agree with you mostly, I disagree on this point.
Jitter is an issue ONLY when there is conversion from Digital to analog or the other way around.
Other things cables etc AFAIK (if you'll permit the cheap pun) have not even a "bit role" to play.
ciao
gr
sound_cycle
Well-Known Member
And while ATM I don't have an explanation or understanding as to why, I am undergoing that discovery.
My PC is connected to DAC/ amp via optical S/PDIF.
Now the sound is quite different depending on whether I draw power through a surge strip (through which other devices PS3, STB, LC TV, DVD player - all two pinned by the way) or from by itself a grounded power socket. I have read (hfv and pers comm from fm) that a linear power supply and surge strip or not make a difference and I would definitely like to try a linear PSU, currently I am using the wall wart.
It took me over two months of frown lines and a bit of serendipity to figure that out.
ciao
gr
Last edited:
reignofchaos
Well-Known Member
Bifrost is a decent dac for the money. Sounds quite good. However do not use optical. It is the most jitter prone interface out there.
There's guaranteed to be at least 250ps of jitter in an optical interface with potentially more depending upon implementation of the source. With the bifrost you can use coax or usb. Either is a much better interface.
I guess the point I was making is how would one know if a 10k/20k cable makes a difference or not without hearing it first? Is it based on theory that these cables make no difference? Then obviously Gordon Rankin is wrong (and AFAIK he knows a lot about usb, more than the average person). Secondly what is exorbitant? It is only relative. One could spend $500 on a usb cable in a system that costs $50k and it is only 1% of the total cost of the system.
Cheers,
Sid
Exactly. Jitter is an issue in the time domain when there is a clock involved !
USB data has no Clock..it appears only after the SPDIF has been generated/is being generated
A signal is Always Analogue even though we call it digital
It is digital only when it is "stored"
sound_cycle
Well-Known Member
Okay. what goes through the coax / toslink/ hdmi to your DAC is digital too.
Analog vs digital - digital can never sound as good, right ? Wrong...
ciao
gr
Last edited:
what goes through the coax / toslink/ hdmi to your DAC is digital too.
Analog vs digital - digital can never sound as good, right ? Wrong...
ciao
gr
the signal transmission is actually analogue. From what i could find out it is not 0s and 1s but a wave (not a square one)
eg
so the bleu signal is at the begin of the cable and the red at the end..and the gap is jitter
since the signal is Digital but its Propagation is in the analogue domain, it is subject to the same challenges as any other signal in the analogue domain
The above is using the bi phase mark code hence the smaller 1-0 transition (2 times) is a 1 and the wider one (only once) signifies a 0.
The above is an exaggeration to represent..but this is why there are differences
this is a nice place to read up..http://www.epanorama.net/documents/audio/spdif.html
Last edited:
sound_cycle
Well-Known Member
From the article you quoted
Every bit of the original data is represented as two logical states, which, together, form a cell. The length of a cel ('time-slot') is equal to the length of a databit. The logical level at the start of a bit is always inverted to the level at the end of the previous bit. The level at the end of a bit is equal (a 0 transmitted) or inverted (a 1 transmitted) to the start of that bit.
The signal may be transmitted electrically but it is made up of 1 and 0 ie bits
How does it work with an optical cable ?
The idea of the circuit is that the digital data is fed through a series current limiting resistor to the LED (typically LEDs only require 5-30mA). A digital 0 is 0V and the LED is off. A digital 1 is +5V and the LED is on. That's it!
The data stream is digital. Further processing is digital/ conversion to analog. Try out this thought experiment feed that output to an amp, can you expect to hear Mahler's fifth ? No. The waveform is therefore not an analogue music signal . QED ?
S/PDIF transmits a digital signal
http://www.hardwaresecrets.com/article/everything-you-need-to-know-about-the-spdif-connection/82
ciao
gr
Last edited:
We are not talking the same thing
All I am saying is that wave propagation in a cable is still analogue in nature and subject to the same challenges a analogue signal has and hence the cause of jitter .
The signal is digital..no doubt about it.
sound_cycle
Well-Known Member
The transmission of the digital data is electrical or optical but is 1 and 0 and not analogue. It is not an analogue signal. It is not treated as an analogue signal (To take your interpretation, i am typing this on an analog computer and transmitting it over an analog internet because somewhere an electrical/ optical state is used to represent my post)
And No the transmission does not induce any jitter
The cable is not the cause of jitter
ciao
gr
So what happens if we use a SPDIF cable which is not 75 ohms but say 25 ohms ? We still get music in the end and it is NOT broken music. What kind of errors going on in the transmission ?
By the way, I have tried many 75 ohms (claimed) digital cable in my system and they all sound different, so something must be going on in the "analogue" domain to affect the digital data
.reignofchaos
Well-Known Member
So what happens if we use a SPDIF cable which is not 75 ohms but say 25 ohms ? We still get music in the end and it is NOT broken music. What kind of errors going on in the transmission ?
By the way, I have tried many 75 ohms (claimed) digital cable in my system and they all sound different, so something must be going on in the "analogue" domain to affect the digital data
Lots of reflections between the sender and receiver which will blur out the edges - that is what my personal experience is when using a less than optimal cable.
Last edited:
reignofchaos
Well-Known Member
The cable can cause reflections between the sender and receiver if it is not the correct length and impedance when using spdif. These reflections cause all of sorts of problems. If you think that all spdif outputs are the same, maybe you should look at this page
CD_transport_DIY
See the oscilloscope traces? They can range from perfect square waves to absolute garbage - especially when using toslink.
Yeah but thats an analog explanation
. What happens to the digital data ? The 1s and 0s I mean. They are lost, garbled or just intact ?
@GR, my understanding too of electrical signal that represents a bit over a transmission medium, is an analog waveform. At its simplest, the receiving end looks for a particular level to interpret whether the received bit is a One or a Zero. Of course practical implementations are much more complicated. For example, a transition from a low to high signal level may represent a One, and vice versa. Also, special situations like continuous Ones or continuous Zeros are handled in special ways to make the transmission valid. For example, in most digital communication channels an "All Ones" incoming stream indicates a (critical) alarm that there is a break from the sending end.
The square wave representation of Ones and Zeros is, at best, a didactic representation, and not true representation of the actual signal.
Also someone mentioned Reed-Solomon coding. It's typically used where forward error correction is required like in satellite communication channels where a request for retransmit would take too much time (240 msec round trip) and is therefore not acceptable for real time applications (think TV broadcast where about 8 frames would be lost in 240 msec, or a voice conversation where 240 msec would leave an uncomfortable gap). Of course the downside of any FEC scheme is the bits that have to be sacrificed for the mechanism to work - typically only three bits out of every four sent is actual payload and one bit is consumed in FEC. Of course one can implement even more strict ratios like 1/2, or less demanding like 7/8.
Back to topic.....
The square wave representation of Ones and Zeros is, at best, a didactic representation, and not true representation of the actual signal.
Also someone mentioned Reed-Solomon coding. It's typically used where forward error correction is required like in satellite communication channels where a request for retransmit would take too much time (240 msec round trip) and is therefore not acceptable for real time applications (think TV broadcast where about 8 frames would be lost in 240 msec, or a voice conversation where 240 msec would leave an uncomfortable gap). Of course the downside of any FEC scheme is the bits that have to be sacrificed for the mechanism to work - typically only three bits out of every four sent is actual payload and one bit is consumed in FEC. Of course one can implement even more strict ratios like 1/2, or less demanding like 7/8.
Back to topic.....
sound_cycle
Well-Known Member
So what happens if we use a SPDIF cable which is not 75 ohms but say 25 ohms ? We still get music in the end and it is NOT broken music. What kind of errors going on in the transmission ?
By the way, I have tried many 75 ohms (claimed) digital cable in my system and they all sound different, so something must be going on in the "analogue" domain to affect the digital data
Disclaimer : I do not claim to have enough knowledge and understanding to answer the question you pose. Let me try, apologies for any lack of understanding that might cause any misrepresentation
The transmission of data over spdif is digital data.
Digital audio is created by converting an analog musical signal by sampling it with a given fixed frequency Fc (the 44.1, 48, 96 or 192 kHz number). Each sample level is converted into a 1 or 0 and stored ("recorded')
Now you cannot hear the music from that sequence of 1 or 0. Transmitting that sequence of 1 and 0 over a cable does not make it analog.
When it is time to listen to Mahler's fifth the samples have to be extracted from our stored digital media, fed to a DAC which has to convert/ recover the analog signal.
Now that recovery has to be at the same Fc that was used to sample the analog signal, else things will go bad.
That is where a clock comes in.
So the digital audio protocol and cable, has to transfer the sound - as digitized sound samples let me emphasize - and also information that is important for the pace of the conversion, ie the clock. S/PDIF (Sony/Philips DIgital InterFace) does this trick
Not all clocks have the ultimate precision and accuracy. So the clocks that provided the Fc while recording and recovery will not agree. That disagreement will cause an imprecise reconstruction of the sampled analog signal.
That biphase mark code thingy arj posted on, works something like this. Each bit is transmitted as two semi-bits of equal duration. Some technical magic happens with this arrangement that facilitates figuring out and extract a sampling rate clock.
The recovered clock is as a reference for a Phase locked loop. It does some black magic and the two clocks, are now on the same average frequency (ie they are locked). And now we are looking at the precision of the clock used in the recovery.
And just to save some yada yada yada, the data dependent jitter goes out of the picture.
Therefore the cable does not cause any jitter.
The waveform deformation that in arj's post is a form of jitter, then there could be line induced jitter (not that it affects optical transmission) and several other kinds too. the electronic wizardry we use attempts to tackle them.
The point we need to be really concerned about is what is happening at the DAC end of things.
The cable is probably not where you would want to throw your money at. I will spend on a DAC
If you want to insist that it is analogue, and base further discussion on that assumption, okay I concede your point. Maybe you should elaborate further on analog processing of the analog data that is transmitted (note: no quotes or rolling eyes). Another question to ask is what about streaming services like Tidal ? If transmission over cables caused things to go horribly and uncorrectably bad that should be impossible, right ?
edit : @joshua, I agree with what you have posted. Except for this critical bit "is an analog waveform" As to the Reed-Solomon coding, yes it is used where forward error correction is required like in satellite communication (and ofc we do). Posted a link somewhere above.
ciao
gr
Last edited:
@Dr Bass, from transmission line theory, we know that matching the characteristic impedance of a transmission medium/channel (which an SPDIF cable is) to the source or the load maximises the energy transferred, and also avoids reflections. If one uses a 25 ohm cable, it creates an impedance mismatch with the 75 ohm source. It also is a mismatch to the 75 ohm load/destination. The cable will reflect a part of the energy sent by the source, back to the source. So a new waveform is created and will exist and interact with the original waveform. These waveforms will interact and inter modulate each other, thus giving rise to see third waveform. Whatever signal that manages to reach the destination is no longer the original, and to make things worse it will undergo further distortion.
Since everyone on this thread is talking about jitter.
How does it sound like?
You know, like I know how low and high bit rate sounds like, how high and low noise floor sounds like, how high and low sampling rate sounds like etc.
But how does this low jitter vs high jitter sound like?
How does it sound like?
You know, like I know how low and high bit rate sounds like, how high and low noise floor sounds like, how high and low sampling rate sounds like etc.
But how does this low jitter vs high jitter sound like?
Wharfedale Linton Heritage Speakers in Walnut finish at a Special Offer Price. BUY now before the price increase.
Similar threads
