I would imagine that the last, and only the last, re-clocker matters. For example, when streaming from Qobuz, or anywhere else, who knows what networking mayhem your digital bits encountered along the way. But then you run it through your Melco transport, and whatever chaotic, jittery state your bits were in, is now an organized data stream to behold. If the above is true, then Dean in London's network streamer is ultimately delivering its designed grouping of bits to his DAC. Everything in his home, before his network, streamer is no different than every router, switch, firewall, etc, on the public internet that also preceded his last hop (his network streamer). Network hops within Dean's home are no different than network hops between Qobuz and Dean's cable modem. I suggest that whichever box has the best, most advanced re-clocking design should be feeding Dean's DAC. Based on Melco's reviews, my guess is that his Melco box will generate the least amount of jitter and the least amount of noise, resulting in the best overall sound quality.
Good question. Good answer up to a point when I think it fell apart. IMHO the correct answer is it makes no difference how many points of clocking there are as long as it remains bit perfect. You can stream info from Tokyo to New York through hundreds of network points and the signal will be bit perfect. The bits can be buffered for years on a CD or a hard drive. No problem. The last 'reclock' before the bits hit the DAC is the one you hear. At this point the information must be perfect (easy) and the timing must be perfect(not easy). Low phase noise is the key to natural sounding digital as the timing AND the proper info defines the sound we hear. There is a notion that all the clocks upstream also matter. I think in fact what matters here is galvanic isolation to reduce common noise and perhaps a switch to reduce network traffic within the streamer.
Paul, your next segment should be on the difference between a master clock and a word clock in digital audio. There is much confusion even among affaciandos.
So, this brings up another question. Lets say the signal gets reclocked 3 times, then the 4 th clocks introduces jitter, is it possible to say that clock 5, being extremely low in jitter compared to clock 4 can now "fix" these jitter errors?
It is usually necessary to have several clocks in a system since each interface has it's own bit rate. Streaming has multiple format and error correction bits added to the bit stream and needs a completely different clock rate from say USB. Its only when it gets to the DAC the real audio sample clock is generated hopefully from a data stream PLL.
Spend money on a good DAC that can accommodate a reasonable amount of jitter from the source, and the problem essentially goes away. The jitter doesn’t become “real” until it gets converted into analog audio.
As usual, Paul goes off on some made-up "audiophile" fallacy about clocks and clocking devices. There is absolutely no need to purchase fake clocking devices to solve jitter in digital audio. The only two clocks that are important are the clock on the equipment that records the digital signal and the final or "master" clock for the playback system. The rest is just snake oil.
Let’s assume we don’t know if any one of the many clocks change the signal (via jitter, etc.). It would be conservative to have as few clocks as possible to prevent damage.
Correct. Unless one of the clocks in the chain is so bad it messes things up. With digital, good enough and perfect are functionally equivalent. So as long as the clocks are good enough it'd fine. Real world, only the last would make an audible difference 99.9% of the time.
Dan Lavery once said that if your converter sounds better with an external clock , then you have a really bad clock ... Any truth to the idea that the best clock is the internal clock ?
well. it really shouldn't matter. if they're not crap anyway the crc for the signal should be same in and out. you could even just save the signal while at it and play it again a year later.
Reality is that the clock that matters is the one that controls the DAC's sample conversion. Ideally that clock runs PLL-free and precisely on its own based on the sample rate of data and it just consumes from enough buffered sample data to never run out of data. PLL is a necessary evil when you start having to deal with multiple clocks and the DAC needs to try to sync with the source clock e.g. in a CD transport. Best is to run a single master clock in the DAC and data is pulled on demand asynchronously from whatever source incl. streaming. Your streaming audio, video, web browsing, online gaming etc. does not benefit from any "re-clocking device" as the internet always is asynchronously moving data. It's plain impossible to precisely clock music data over TCP/IP. For digital audio it's rather simple. Your DAC is 100% at peak performance on the data side if it can control the clock precisely (DAC's own clock is master) and the data is coming without errors. Data errors is not a real concern nowadays except if you use a transport with scratched CDs. Those cursing streaming needs to check status quo of it. I hated streaming myself a few years ago because it was compressed and could not match my CDs. I've collected CDs since the 1980s and vinyl discs since the 1970s...but they are now stored in boxes. With lossless CD (and better quality) streaming, being audiophile has never been better. Having multiple re-clocking devices in series might actually cause issues with the data flow potentially causing intermittent data dropouts. Besides the fact it would be waste of money having no benefits, it could actually cause damages in ensuring the DAC is always having adequate buffered data.
As a network engineer and an audiophile, I would say something like: "As long as the 1´s an 0´s gets thru, i don´t care if it´s smoke signals or binary info typed in ink on a paper. It´s the source file you should be worried about" Please don´t kill me if i´m being wrong or anything. But ones and zeros are the same on both a cheap network device or an expensive digital audio...whatever. And can somebody please explain to me what kind of voodoo they implement in a network switch for claiming the rights to call it "audiophile"? Computers can do all kinds of bad stuff but as long as a network can deliver your precious data, I can´t see what more a switch can do.
It can add noise to the signal, which doesn't matter with spreadsheets and word docs, but it does with high-end audio and video. That is why every digital tv has a digital noise reduction setting and you can hear a difference in audio.
What I took from that is: That you can have as many digital re-clockers as you like, and the signal will not be damaged, unless the re-clocker damages it, but if you have a perfect re-clocker then all is ok. Logically, then, in the real world, we should treat it like analogue and have as few things in the chain as we can get away with.
LMAO. Audiophile Ethernet gear is the absolute biggsst scam in the history of audiophile scams. But people still go for it. Might start a company myself.
'biggest scam in the history of audiophile scams' is quite the claim, lol. But yeah, total nonsense. Any internet streaming service like Spotify or Tidal will be encrypting streams to prevent piracy. That will also prevent your switch from reading the audio data, if it tried... which I highly doubt it is.
@@Pete.across.the.street Easy. It's a scam coz modern network is a guaranteed bit perfect delivery on content. This is literally the fundamental premise of how data networks function. So there's no more or less but perfect delivery. It's perfect or it's discarded. And for the other argument of timing that's out too, again due to how networks function. Data is stuffed into a packet and fired onto the network and as long as it makes it to the other end before the application times out it's considered success. And latency is constantly changing depending on the load on the network. As for noise it's again pointless. There will either be low enough noise that the packet is perfectly being received or there will be too much noise and the packet will be discarded. So literally when using networks as a transport medium, you'll either have perfect transport or none at all. And any attempt to sell you anything special for that is the worst sort of snake oil and utter hogwash.
A Hunter S. Thompson related mix, salvaged from wartime after-hour dance recordings... MP3 Backup DJ 13 - Qlimax Remixes 2006 th-cam.com/video/NBTmAZupUi8/w-d-xo.htmlsi=0oyEoEqC0nUKSa6o&t=189 My 4am nickname, "l'ordi" ..the computer ..now f'in stop smashing it ya bunch of invading barbs, Axis FORCE-indulging scabs. When ppl cannot understand what I am saying to them, they call me "le fou" (the crazy fool) ..while I accomplish what they rather neglect.
Very good question ! I assumed there would be one master clock that ruled them all
And the digital darkness BIND THEM !
There can be. A master clock can make a large difference for your server/streamer and even some DACs.
You would be correct!
Yes, exactly right, you want only one master clock that doesn’t need PLL to play catch up games with other clocks and can manage DAC timing precisely.
I would imagine that the last, and only the last, re-clocker matters.
For example, when streaming from Qobuz, or anywhere else, who knows what networking mayhem your digital bits encountered along the way. But then you run it through your Melco transport, and whatever chaotic, jittery state your bits were in, is now an organized data stream to behold.
If the above is true, then Dean in London's network streamer is ultimately delivering its designed grouping of bits to his DAC.
Everything in his home, before his network, streamer is no different than every router, switch, firewall, etc, on the public internet that also preceded his last hop (his network streamer).
Network hops within Dean's home are no different than network hops between Qobuz and Dean's cable modem.
I suggest that whichever box has the best, most advanced re-clocking design should be feeding Dean's DAC. Based on Melco's reviews, my guess is that his Melco box will generate the least amount of jitter and the least amount of noise, resulting in the best overall sound quality.
Good question. Good answer up to a point when I think it fell apart. IMHO the correct answer is it makes no difference how many points of clocking there are as long as it remains bit perfect. You can stream info from Tokyo to New York through hundreds of network points and the signal will be bit perfect. The bits can be buffered for years on a CD or a hard drive. No problem.
The last 'reclock' before the bits hit the DAC is the one you hear. At this point the information must be perfect (easy) and the timing must be perfect(not easy). Low phase noise is the key to natural sounding digital as the timing AND the proper info defines the sound we hear.
There is a notion that all the clocks upstream also matter. I think in fact what matters here is galvanic isolation to reduce common noise and perhaps a switch to reduce network traffic within the streamer.
Paul, your next segment should be on the difference between a master clock and a word clock in digital audio. There is much confusion even among affaciandos.
An audiophile reclocking network switch.
I guess the sales guy was out of reclocking audiophile network cables....
Back in the late 70s, when I was a digital engineer we called it edge triggered. Mostly JK flip-flops.
So, this brings up another question. Lets say the signal gets reclocked 3 times, then the 4 th clocks introduces jitter, is it possible to say that clock 5, being extremely low in jitter compared to clock 4 can now "fix" these jitter errors?
I had a wireless speaker that had a bad clock in it. That was so bad the music would randomly wow and drift in pitch.
It is usually necessary to have several clocks in a system since each interface has it's own bit rate. Streaming has multiple format and error correction bits added to the bit stream and needs a completely different clock rate from say USB. Its only when it gets to the DAC the real audio sample clock is generated hopefully from a data stream PLL.
But why not just one? The final device or the device feeding it.
isnt that why some believe in external clocks? - reduces the chances of getting jitter etc.
The only real reason to restore clock(jitters reduction), it’s to properly re-synchronize the the signal. It’s even great for doing.
I guess that is necessary to have enough buffering in each of these components to ensure this statement that Paul is making
Spend money on a good DAC that can accommodate a reasonable amount of jitter from the source, and the problem essentially goes away. The jitter doesn’t become “real” until it gets converted into analog audio.
As usual, Paul goes off on some made-up "audiophile" fallacy about clocks and clocking devices.
There is absolutely no need to purchase fake clocking devices to solve jitter in digital audio.
The only two clocks that are important are the clock on the equipment that records the digital signal and the final or "master" clock for the playback system.
The rest is just snake oil.
In '97, I had a:
Theta TLC
TLC/Audio Alchemy Pro
TLC/Pro /Pro
TLC/Pro/Pro32
Pro/Pro32/DL
Pro32/Pro32/DL
All used with the Lambda/Ultralink 2.
But, but Paul... BUT, PAullll, you said there is no perfect equipment.
if you have a "very good DAC", but you need some external "reclocking" device to make it work properly, is it a "very good DAC"?
Let’s assume we don’t know if any one of the many clocks change the signal (via jitter, etc.). It would be conservative to have as few clocks as possible to prevent damage.
I have wind-up clocks. battery-operated clocks, and electric clocks.
They don't seem to mess up my timing.
Ensure they're well synchronised then...
Is jitter only affected by last clock before converting to analog?
Correct. Unless one of the clocks in the chain is so bad it messes things up. With digital, good enough and perfect are functionally equivalent. So as long as the clocks are good enough it'd fine. Real world, only the last would make an audible difference 99.9% of the time.
Dan Lavery once said that if your converter sounds better with an external clock , then you have a really bad clock ... Any truth to the idea that the best clock is the internal clock ?
well. it really shouldn't matter. if they're not crap anyway the crc for the signal should be same in and out. you could even just save the signal while at it and play it again a year later.
Reality is that the clock that matters is the one that controls the DAC's sample conversion. Ideally that clock runs PLL-free and precisely on its own based on the sample rate of data and it just consumes from enough buffered sample data to never run out of data. PLL is a necessary evil when you start having to deal with multiple clocks and the DAC needs to try to sync with the source clock e.g. in a CD transport. Best is to run a single master clock in the DAC and data is pulled on demand asynchronously from whatever source incl. streaming. Your streaming audio, video, web browsing, online gaming etc. does not benefit from any "re-clocking device" as the internet always is asynchronously moving data. It's plain impossible to precisely clock music data over TCP/IP. For digital audio it's rather simple. Your DAC is 100% at peak performance on the data side if it can control the clock precisely (DAC's own clock is master) and the data is coming without errors. Data errors is not a real concern nowadays except if you use a transport with scratched CDs. Those cursing streaming needs to check status quo of it. I hated streaming myself a few years ago because it was compressed and could not match my CDs. I've collected CDs since the 1980s and vinyl discs since the 1970s...but they are now stored in boxes. With lossless CD (and better quality) streaming, being audiophile has never been better.
Having multiple re-clocking devices in series might actually cause issues with the data flow potentially causing intermittent data dropouts. Besides the fact it would be waste of money having no benefits, it could actually cause damages in ensuring the DAC is always having adequate buffered data.
Reclocking in an "audiophile" ethernet switch is a total nonsense.
lol, yeah. Sadly, I believe those are a scam. Even if tried to, your streams are likely encrypted. Good luck with that...
Ethernet, when using TCP, guarantees that each packet (of bits) is delivered intact.
Never! That'd be like "too much fidelity"-although, no matter how much I explain this to my wife, she fails to understand. (" -.-)
OctaveRadio (any streaming service) reaches users over man 10s of devices (routers, switches, firewalls etc).
As a network engineer and an audiophile, I would say something like: "As long as the 1´s an 0´s gets thru, i don´t care if it´s smoke signals or binary info typed in ink on a paper. It´s the source file you should be worried about" Please don´t kill me if i´m being wrong or anything. But ones and zeros are the same on both a cheap network device or an expensive digital audio...whatever. And can somebody please explain to me what kind of voodoo they implement in a network switch for claiming the rights to call it "audiophile"? Computers can do all kinds of bad stuff but as long as a network can deliver your precious data, I can´t see what more a switch can do.
It can add noise to the signal, which doesn't matter with spreadsheets and word docs, but it does with high-end audio and video. That is why every digital tv has a digital noise reduction setting and you can hear a difference in audio.
Send the singal asynchronously to the DAC vis USB and let the DAC do the clocking. Problem solved.
What I took from that is: That you can have as many digital re-clockers as you like, and the signal will not be damaged, unless the re-clocker damages it, but if you have a perfect re-clocker then all is ok. Logically, then, in the real world, we should treat it like analogue and have as few things in the chain as we can get away with.
LMAO. Audiophile Ethernet gear is the absolute biggsst scam in the history of audiophile scams. But people still go for it. Might start a company myself.
How so?
@@Pete.across.the.street yes
@@KarlHamilton how is it a scam
'biggest scam in the history of audiophile scams' is quite the claim, lol. But yeah, total nonsense. Any internet streaming service like Spotify or Tidal will be encrypting streams to prevent piracy. That will also prevent your switch from reading the audio data, if it tried... which I highly doubt it is.
@@Pete.across.the.street Easy. It's a scam coz modern network is a guaranteed bit perfect delivery on content.
This is literally the fundamental premise of how data networks function.
So there's no more or less but perfect delivery. It's perfect or it's discarded.
And for the other argument of timing that's out too, again due to how networks function.
Data is stuffed into a packet and fired onto the network and as long as it makes it to the other end before the application times out it's considered success.
And latency is constantly changing depending on the load on the network.
As for noise it's again pointless.
There will either be low enough noise that the packet is perfectly being received or there will be too much noise and the packet will be discarded.
So literally when using networks as a transport medium, you'll either have perfect transport or none at all.
And any attempt to sell you anything special for that is the worst sort of snake oil and utter hogwash.
A Hunter S. Thompson related mix, salvaged from wartime after-hour dance recordings...
MP3 Backup DJ 13 - Qlimax Remixes 2006 th-cam.com/video/NBTmAZupUi8/w-d-xo.htmlsi=0oyEoEqC0nUKSa6o&t=189
My 4am nickname, "l'ordi" ..the computer ..now f'in stop smashing it ya bunch of invading barbs, Axis FORCE-indulging scabs.
When ppl cannot understand what I am saying to them, they call me "le fou" (the crazy fool) ..while I accomplish what they rather neglect.
You...again? Trying desperately to push traffic to your channel. Hunter Thompson "related mix" You have to be kidding, Laugh out loud... loudly!
I watch this channel and sometimes share what is new or pertinent from mine.