Clocks inside streamers

Why is there a master clock inside a streamer and how important is it?

[@onepieceatatime]

Even a broken DAC is right a few thousand times a second.

[@brianatkinson6600]

Excellent channel! I love the fact that these guys use some antique test equipment in their collection. Use it if you can get away with it. I've been dealing in and using old test for many decades.... this new stuff doesn't stack up in build quality. The old Tektronix and HP equipment are pieces of art.

[@TS-ex4dl]

The final DAC clock should be of the best quality.

[@njrumenos]

I thought the streamer clock would become the master clock when using I2S

[@Pete.across.the.street]

If you feed the final clock crap it's still going to be crap. You still want the first clocks to be decent

[@ThinkingBetter]

@ NO...the streamer is the data source and should not master the clock. The DAC must do this or you have multi clock sync and PLL issues. Problem is I2S is using synchronous data transfer. Ideally the data is transferred asynchronously to allow the DAC to fully and precisely time the tick of each sample being transferred to analog.

[@edmaster3147]

The clock is only a small part of the device, and not that significant in my beliefs. A bit of jitter and time smear is just fine, most of the time the devices have far more pressing issues. Plugging in your power cord the other way round in order to pole the transformer differently could make DAC sound better than a slightly better clock. Just stay away from high sample rates and bit depth to make the clock less critical.

[@Pete.across.the.street]

@edmaster3147 you could use $20 speakers to make it not matter as well

[@user-od9iz9cv1w]

I appreciate Paul's history of CDP evolving into the streamer/DAC. A CDP was a data source, the CD device, a clock and a DAC. We then introduced separate DACs leaving the CD spinner and clock separate. DAC's started to have clocks which is smart. The closer the clock is to the DAC the easier it is to keep the signal clean. Now transports have evolved to be network connected computers with the possibility of local storage or network accessed data. Still the clock is key to the sound. If the clock is perfect in the DAC, no other clock is necessary. End of story. But if the clock can be bypassed, then putting a better clock in the streamer might improve your sound if you can connect it without messing up the signal. For me this has just made it a perfect environment for DIY. When gaming required better PC's than general purpose home computers provided, we all built out own gaming rigs. It was not tough using the same lego parts DEL and HP used. It is the same with a streamer/DAC. You can build or buy all the parts and upgrade them independently. Or you can let great companies like Paul's do it and just enjoy it.

[@kennixox262]

Remember when Teac's Esoteric line had that three box CD player with an "atomic" clock? It was a separate transport, a DAC and stand-alone clock.

[@edmaster3147]

Might PS Audio try setting up parallel multiple clocks in order to find a constant in the errors? Last year I've seen an implemantation of such a oscilatiing circuit adding over a 'hinge' which was quite entertaining.

[@markuslehmann5049]

Many thanks for answering my question - I really like your channel!

[@DrTune]

Did he answer your question?

[@DjRjSolarStar]

Oh my god, you own the same brand of linear adjustable lab power supply as me! I have the smaller version, it's been indispensable over the years for testing amplifiers and custom electronics circuits.

[@rambler3080]

Notice how he didn’t answer the question. He knows that having a steamer clock the signal only to have the DAC buffer (previous clocking is stripped at that point) and then reclock the file doesn’t make any sense.

[@pierpaoloazzalin9505]

I've always had the feeling he didn't really answered the question

[@RandySmith-iz1ml]

Great explanation Paul, thanks! That leads me to the question of the NAS, if one is ripping CD's to a NAS and then listening to the music from the NAS, does it matter what quality the CD player or transport is when we originally rip the CD to the NAS? Thanks so much!

[@kevinmills5293]

If Paul is citing jitter as the cause of poor sound quality, then no. Jitter is not a factor when ripping a CD. As long as there is no data corruption the ripped data will be perfect.

[@Bassotronics]

Just use a quatum variable speed plutonium atomic clock. 😁

[@ThinkingBetter]

Ideally the DAC runs the actual timing precisely and everything else upstream is just moving data asynchronously with adequate buffering in the DAC to avoid dropouts. Since a player needs to be in sync with the actual audio, the DAC buffer can’t be too large or you will, for example, experience pressing pause will cause a latency. One clock master is all you need and all talk about multiple clocks/reclockers just obfuscate what truly is ideal architecture. Jitter becomes a non issue with such architecture and in reality it’s also simpler. Problem is that old hardware interfaces were made for source dictated clock from either CD, DSD, Blu-Ray or broadcast video where you end up with at least two clocks. For modern Internet streaming, there is no such issue and hardware interface to DAC should be asynchronous.

[@user-od9iz9cv1w]

Right on. The best place for a perfect clock is as close as possible to the DAC. OTOH, if you own a DAC with a mediocre clock, you can potentially upgrade it's sound by using a streamer with a better clock iff the clock in the DAC has the option to be bypassed. Still not ideal.

[@ThinkingBetter]

@@user-od9iz9cv1w Actually only the DAC's clocking of samples matter for audio fidelity while all other clocks are about ensuring the data is available for the DAC without dropouts. There is no actual good reason in "re-creating" the digital audio problems of multiple clocks and PLL in a modern online streaming solution. A "streamer" should output the digital samples asynchronously to allow the DAC to have full PLL-free control of the clock.

[@edmaster3147]

putting in a small resistor on a clock line to a DAC chip could be quite helpful.

[@Douglas_Blake]

Simple .... a streamer is a computer. It connects to a server via TCP/IP and essentially downloads the music file. The file is coming in many times faster than the playback speed, so it is buffered in memory. The ACK/NAK protocol is used to arbetrate downloading chunks of the file to avoid buffer overflows. Now this computer has a sound chip on it, a DAC, that runs at the sample rate, not the main buss rate, so it needs a clock to time it's data consumption. Often the sample clock is derived from the main clock so they are more or less synchronized. When the buffer gets below a certain point, the DAC logic signals the main computer to download another chunk of file. Running in a circle until the playback is finished. In the end it's not much different than USB Audio from a PC.

[@rosswarren436]

Most streamers these days still have a DAC in them, either a "good" one at their price point or one just for "convenience", knowing that the buyer will likely be using a much better external DAC. So yeah, I can see that streamers that have those built-in DACs still need a decent clocking. The engineer guy who is asking the question does bring up a valid point. At some price point for streamers, why not simply do away with that "convenience" DAC and clocking and merely be a true transport sending the data to an external DAC with its own very good buffer and clocking. It might not be adding much to the cost of these devices, but why pay for something we aren't planning on using?

[@TheDanEdwards]

This confusion all arises due to TOSlink hell. TOSlink is just a thing that should not have been; it was designed (as Paul note) that the CD spinner sent out the data with the timing. Today everything upstream from the actual DAC (the device that does the conversion, not a box labeled "DAC") should not matter as far as clocking.

[@BobbyTwoTimes]

What just happened? Uncle Paul is the greatest

[@hwpno3]

All of this talk is about reducing or eliminating jitter.. True? If so, what does jitter sound like if it gets through your system? Thinking back over 30 years ago when I had very budget equipment with better than budget speakers I don't recall anything objectionable when listening to CD's. Which I loved the sound of. Even today I stream directly to a Mid end receiver and it sounds fantastic. I have a $150.00 Sony Blu-ray player that plays everything and sounds fantastic, especially SACD and DSD disks. All the years I've never heard anything objectionable. So what does jitter sound like?

[@kenwuesq]

Thank you for the comment. I agree. There are a couple examples on the internet, but none from manufacturers or audiophile publications. Please give samples so we know what to listen for. I am genuinely interested. Is it wobbly pitch? Pop? Crackle? Complete cut off in sound? Hiss? These are descriptive words that will help.

[@DrTune]

​@@kenwuesqit doesn't sound like anything, when audiophiles went digital vendors had to come up with some analog-like bogeyman that they could charge you more and more money for allegedly better and better performance. Less jitter! 10x less jitter! Oven controlled crystals mounted on rubber shock absorbers! It's absolute nonsense. Sure it's possible to drive a dac with an artificially terrible sample clock and if you make it bad enough you can measure and _maybe_ hear extra harmonics (the essentially noise in the dac clock can modulate the audio signal) but it's a complete non-issue for even the most bargain basement Amazon cd player or USB streamer digital output. The snake vendors realized you couldn't be convinced that one digital number from their expensive wood grained device was different to the same number coming from a $29.95 plastic box so instead the jitter boogyman was invented, so they can say "sure, it's the same number, but ours has much lower jitter" and because jitter is a analog value and regular people aren't able to readily hear or measure it, it's ideal for separating fools from large amounts of money. See also; audiophile network cables, etc

[@kenwuesq]

@@DrTune Thank you and I agree about bogeyman. Even when manufactures talk about “noise,” they’re not specific. My baby crying is noise. Traffic is noise. What should we pay attention to so we can identify and fix? Hiss? Hum? Pop? No sound? Skipping music?

[@phillipmorris9847]

you cant hear it until you unhear it, then you wont want to hear it again.

[@hwpno3]

Interesting! Thanks for the comments

[@thinkIndependent2024]

Yep Paul's is RIGHT!!!! Multiple Clocks with Re-clocking WORKS !! I'm moving on to A.I.projects but a Smart DDC is needed in Audio I invented such an improvement but have no time to get it out of the door to Public .

[@ThinkingBetter]

@@thinkIndependent2024 No, it’s totally nonsense. For internet audio streaming you simply need a precise DAC clock that runs the sample conversion and then all the music data is moving asynchronously with buffering ensuring there are no dropouts. Jitter becomes a topic on internet music streaming when the music data is not managed fully asynchronously from data center storage (HDD or SSD) vis the internet, music streamer and to your DAC. Re-clocking is a stupid snake oil scam and just showing a company’s lack of digital audio architecture and engineering competency.

[@thinkIndependent2024]

​@@ThinkingBetter 1 lump or 4 lumps of Sugar in your Tea? Sure you can build a DAC that can do noise shaping and filtering Top Notch. Understand this in the Audiophile audience just like everything else most just listen to the Music others know exactly what's missing. Lucky for me dur8 the Pandemic I was stuck at home with hundreds of thousands of test equipment and time to study exactly how audio works from start to end ... Yes!!! Music benefits in refining the last step before the entering the DAC...before the pandemic gave me time to Analyze my answer would have been similar to yours even with my 40 year's of electronics and computers. Last thing electronics are built to a price point Example : Microsoft could have built everything into there software..look at all the things must buy... Boeing could have just built attack angle mitigation into the 737 Max!!! But the didn't please consider the facts above...

[@NoEgg4u]

Marcus, in Cologne Germany, A DAC is a digital to analog converter. That process does not include a clock, which is necessary to reduce jitter. The clocking operations are performed by the transport box. A CD player is a combo box, where it is both a DAC and a transport (the reading of the bits from the CD is the transport section). The streamer that you inquired about is a transport. It receives data from some source (the internet, a computer on your network, its own internal storage drive), and re-clocks it to minimize jitter, and hands it off to the DAC for it to create analog voltage..

[@razisn]

All DACs include clocks. For (asynchronous) USB connection it is the receiver's (DAC's) clock which is used. For S/PDIF, DAC clocks can be used for clock comparison and correction or for buffering and re-clocking. Btw the clock's job is to enable the DtoA conversion process, it is not just 'to reduce jitter'...

[@SteveWille]

@@razisnTotally… I’m not sure how Paul managed to talk about this topic without mentioning this fundamental difference in the various digital connection types.

[@NoEgg4u]

@@razisn It is my understanding that the DAC's clock syncs to the transport's clock.

[@razisn]

@@NoEgg4u what I said above

[@edmaster3147]

How does a DAC chip time? Perhaps a word clockline (lets say a high on every word) and a bit clockline, on every bit. So a divided by 24. And then some magic happens, lets take an 8 time clock to allow oversampling. Chips kinda like being told what time it is and what to do, so to make them happy timekeeping is a bit of a management issue.

[@csimet]

Interesting subject & info, but I just can't help myself.... to notice the obvious electrical panel cover sitting on the floor to his right, meaning the building power panel is not covered and what looks like wires sticking out. OSHA would not be happy. 😉

[@uribar-ner5055]

Does lifting up the clock affect the sound?😅

[@user-od9iz9cv1w]

A slow move like lifting a device is insignificant. But... Clocks are microphonic. Vibration does impact jitter. Clocks are also generally temperature sensitive. The best clock will perform best when implemented in a temperature controlled environment in a mechanically isolated environment to free it from all vibration including the sound of the music. Lifting the oscillator off a PCB and putting it in a blob of silicone will have an immediate and significant change to the sound.

[@steve8553]

Isn't it true that when you connect a DAC to a Streamer using a S/PDIF COAX the master is the streamer's clock (without a reclocker). If USB, there is no clock in the transmission protocol, so the DAC is the master?

[@ThinkingBetter]

@@steve8553 With asynchronous audio transfer via USB you can allow the DAC to master the clock without messy multiple clocks trying to work together through a PLL. Jitter absolutely does not need to be a topic for a modern internet streaming solution but it becomes a topic when using older synchronous data interfaces like SPDIF, Toslink or I2S. This video doesn’t convince me PS audio gets it right about asynchronous audio to the DAC.

[@steve8553]

@@ThinkingBetter Thanks. I have very nice DAC connected via COAX to a mediocre streamer. It's nice. But I tested connecting to my MacBook via USB. I was shocked how much better USB is. I was told USB was inferior, but that's not a universal truth is it?

[@edmaster3147]

@@steve8553 might be the streamer at fault?

[@steve8553]

@ Yes, exactly. With S/PDIF (coax) all things being equal, your result is only as good as the sender's clock. But USB is asynchronous, so the receiver's clock rules. (So I've learned and experienced.)

[@ThinkingBetter]

@@steve8553 USB is way superior with higher bandwidth and a proper protocol supporting asynchronous data transfer.

[@hifisoftware]

I think it's also.important to mention that all decent DACs will buffer, which means that they will use their own clock. And that means that it does not matter if the original source is jittery or not, or what it is. As long as bytes come in, then they will play perfectly. Otherwise some folks might think that expensive CD players somehow send "better" bytes than an absolutely cheap CD player. Or some might think that some bit perfect sources will sound better than others - which of course is not true. But of course different digital sources might have the same performance but mastered differently. And how it was mastered is where the difference really is.

[@m_sherzod]

kzbin.info/www/bejne/eIrSg3eFm7qGfassi=3dInPxHIKJTetGq5 So this is actually a viable upgrade for the Wiim Ultra you mean?

[@janwillemkuilenburg7561]

I guess the simple answer would be that the digital stream requires a clock to reduce jitter / smearing that is caused whenever the stream passes a device or an interconnect?

[@user-od9iz9cv1w]

I'd argue no. There is no need for the data to be anything other than bit perfect until the moment it is presented to a DAC for conversion. Then the timing is equally important to the data. The last clock to specify the timing that the DAC uses is the one that affects the sound. If the DAC clock cannot be bypassed, the streamer clock really doesn't matter. Likely it is the noise isolation in the streamer that improves the sound. If the DAC clock can be bypassed, then a clock with lower jitter in the streamer will improve the sound if the connection can maintain the clock signal.

[@Pete.across.the.street]

​@user-od9iz9cv1w isn't it only bit perfect until the ethernet connection though?

[@RoderikvanReekum]

Who uses streamers anyway? Nothing can beat vinyl records and a turntable. Do not listen to anyone who thinks digital audio is good, it has to be fully analog.

[@Patrick-vl2gx]

Who needs cars and planes anyway? Horses and buggys are the only way to travel.

[@RoderikvanReekum]

​​@@Patrick-vl2gxThat is different, vinyl is still TOP quality audio, digital audio does not get close to it. Same as you see more of the beautiful landscape when you use Horse and buggys!

[@spenceralridge4958]

That is a pretty narrow-minded statement. And also wildly inaccurate.

[@snoochpounder]

Oh yes nothing better than going for a run with a vinyl player strapped around your neck.

[@spenceralridge4958]

@ Well, of course. We all know that you cannot get any decent sound unless comes from a low speed spinning disk of low grade and easily scratched petroleum by-product that is read by a very thin metal sliver that oscillates inside of a tiny coil, all the while perched on a platform that must be heavy enough to prevent any accidental movement. Yep, gotta run with all of that because nothing, absolutely nothing, can beat that Rube Goldberg machine that was invented over a century ago. Cannot be improved upon in any way.