Sorry, but the irony was not recognizable for me because it sounded as if you were using this device yourself.

"I have transferred a lot of data between computers through networks..."

Yes, without such an Device needed at all i ment and i used all kind of Networks since beginning with an Intel I386 processor around 1992 with an BNC Network And never lost Data this way.

And you has CRC profing or anything like this on Datastreams so if there is an error the Data get send again or you get an error Massage at any Point.
Music Data get buffered and reclocked on Streamers or USB DACS they run Asyncrone so anyway. 

Also CamillaDSP uses an Buffer you can change the Number.

As Example with an Buffer off 1024 i got silent kliking and poping noises in between, after i changet it to 2048 this noises was gone that was because i used 96kHz Samplerate and 1024 is defoult for 48kHz Samplingrate.

The RME Card Driver has such Buffer on Hammerfall DSP Settings Page. But not sure at wich point i need to change it.

We have onlinebanking long time already think what happen if a bit would flip while trasfering Data thrue an Network

The MSB probably LOOL

"I have transferred a lot of data between computers through networks..."

Frunse wrote:

Here something for Farytale lovers

https://nuprimeaudio.com/product/sw-9x/

I have transfered a lot of Data between Computer thrue Networks but never lost an Bit lol

Also my Switches uses most Time LWL in between so they are isolated by nature.

But at the End i found an "EINHORN" an Switch with Clock input LOOOOOL

Network switches don't have a word clock input in the first place, that concept doesn't even apply here. The TCXO mentioned is just the switch's own internal timing reference for its Ethernet PHY, completely unrelated to any audio clock domain.

Frunse wrote:

I have transfered a lot of Data between Computer thrue Networks but never lost an Bit lol

Sorry, but this is nothing special. There are mechanisms in the underlying internet transport protocols that prevent data loss.
Either the protocol handles it itself (TCP), or when using UDP, this is the task of the higher protocol layers.

According to the product description of the network switch:
"The SW-9X is a high-end 8-port gigabit Ethernet Network Switch with built-in low noise power supply, AC filter, and ground shielding. It improves the network performance by reducing signal noise. By reducing circuit and power supply noise, and using a customized TCXO clock, SW-9X delivers accurate network signal that ensures stable music data transmission."

Well, a network switch with audiophile-quality claims.
It sounds like they adopted the vocabulary of audiophile customers and tried to translate this to the network business
to entice audiophiles who need LAN switches into buying these products, convincing them it's the best they can do.

An 8-gigabit switch is nowadays absolutely no rocket science anymore. The best that you can do is to buy a standard, manageable 8-port Gigabit switch with the feature set that you need and to save some money for other things.

Sorry. but with €1500 this device is overpriced and the marketing claims don't hold up technically.

Frunse wrote:

.....Think about DANTE now ........
What comes next...PC with external Clock input for CPU

AVB and Dante have their own mechanisms to ensure proper transfer of audio streams. AVB focuses on Layer 2, Dante supports Layer 3 (routing). Dante's clock synchronization is handled via PTP (IEEE 1588) directly between the audio endpoints, not by the switches in between. A switch's internal oscillator quality is irrelevant to that process, regardless of marketing claims to the contrary.

I ment it ironcly i hope it was not misunderstandable.

Also every coper Ethernet Adpter has an Transformer like isolator on Out so every Ethernet device is isolated like ASE Ebu on professional Devices i guess.

That was the first Time i saw an Switch with an external Clock input ....
In german we cal it EINHORN, normaly not seen in Real world.

And this one has an gold plated connector so it sounds even better

Here something for Farytale lovers

https://nuprimeaudio.com/product/sw-9x/

I have transfered a lot of Data between Computer thrue Networks but never lost an Bit lol

Also my Switches uses most Time LWL in between so they are isolated by nature.

But at the End i found an "EINHORN" an Switch with Clock input LOOOOOL

Network switches don't have a word clock input in the first place, that concept doesn't even apply here. The TCXO mentioned is just the switch's own internal timing reference for its Ethernet PHY, completely unrelated to any audio clock domain.

Sorry, but this is nothing special. There are mechanisms in the underlying internet transport protocols that prevent data loss.
Either the protocol handles it itself (TCP), or when using UDP, this is the task of the higher protocol layers.

According to the product description of the network switch:
"The SW-9X is a high-end 8-port gigabit Ethernet Network Switch with built-in low noise power supply, AC filter, and ground shielding. It improves the network performance by reducing signal noise. By reducing circuit and power supply noise, and using a customized TCXO clock, SW-9X delivers accurate network signal that ensures stable music data transmission."

Well, a network switch with audiophile-quality claims.
It sounds like they adopted the vocabulary of audiophile customers and tried to translate this to the network business
to entice audiophiles who need LAN switches into buying these products, convincing them it's the best they can do.

An 8-gigabit switch is nowadays absolutely no rocket science anymore. The best that you can do is to buy a standard, manageable 8-port Gigabit switch with the feature set that you need and to save some money for other things.

Sorry. but with €1500 this device is overpriced and the marketing claims don't hold up technically.

AVB and Dante have their own mechanisms to ensure proper transfer of audio streams. AVB focuses on Layer 2, Dante supports Layer 3 (routing). Dante's clock synchronization is handled via PTP (IEEE 1588) directly between the audio endpoints, not by the switches in between. A switch's internal oscillator quality is irrelevant to that process, regardless of marketing claims to the contrary.

https://nuprimeaudio.com/product/sw-9x/

I have transfered a lot of Data between Computer thrue Networks but never lost an Bit lol

Also my Switches uses most Time LWL in between so they are isolated by nature.

But at the End i found an "EINHORN" an Switch with Clock input LOOOOOL

.....Think about DANTE now ........

What comes next...PC with external Clock input for CPU

https://www.diyaudio.com/community/thre … st-8023326

And here you can find some Picture from my Speaker and Work in Progress with Waveguide:

https://www.diyaudio.com/community/thre … 25.428467/

I use it with an Scanspeak 18W/4434G00 in a 2 Way Design down to 60Hz Bassreflex (5 Speaker build from my Design thrue an Woodworker).

Also Designed an Waveguide (3D printed) for it and tested it already that gives me the possebility to use it from 1800Hz 96dB/octave FIR filtered on CamillaDSP with low Distotion. The Waveguide i will CNC cut than in to the Baffel and place the Chassi from behind every where Round eges for smooth response. At the Moment it works at 3kHz Crossover point without Waveguide.

For Bass i use right now 2 10" Bassreflex Subwoofer and buoght already 4x 12" Chassis for 2 Ripole on an Axel Ridthaler aproved design But that work in Progress.

So i will design 2 new Frontspeaker with it and 2 18W/8434G00 Side by Side for even bigger Waveguide, I use HYPEX OEM Amp's NC122 2x125Watt 4Ohm and NC250 1x 250Watt 4 Ohm for Subs
So i have than an 7.2 Setup for my new Apartment

My actual DSP Project runs here:

https://www.diyaudio.com/community/thre … up.440807/

English is my third language so i am not every time able to understand every Joke people made with native English capabillity.
And also not able to descripe all well like on Motherlanguage.

I would like to test Planar Chassis like this:

https://www.soundimports.eu/de/grs-pt50 … gKZzfD_BwE

But never did right now.

Piega has also something like Planar chassis not directly Ribbon like this:

https://www.hifisound.de/de/Lautspreche … oener.html

Its every Time the same

And this happen much offten than i wan't

Yes, the same thing in these situations is you.

You are fooled by marketing wordsalad and get disappointed by user forum because others disagree to the marketing.

Magazins about tech stuff often include promotion articles rather than journalistic material.

Building loudspeakers is a great art work.
My favorite is Piega from here just at the other side of the Lake Zürich in Horgen.
What is your opinion about ribbon tweeters ?

Now its time to calm down, best wishes and kind regards

Discussion split from the other thread and moved, it's not related to the original topic.

Frunse wrote:

Please stop talking if you can't provide prove off what you talking about.

Please refrain from telling other forum members what to do or referring to them as "trolls". Thank you.

If at all, jitter is an issue during conversion. The HDSPe AES does not have any converters.

Sorry but if i do an serius technical Question i don't like to heare something about farytales or anything like this.

Its every Time the same, if i didn't like an question than why answere to it, its worthless to read for me.

And this happen much offten than i wan't

So i am 56 YEars old and build speaker since iam 12 and Audio is my Hobby not my Business.

I have not the Time to bother around with such kind off people that gives no sirious Answere.

If it's not my Busines or i can't give a prove or technical description why i need to answere to an question with jokes?

That is not Helping at all and cost only time and Space here ont that server LOL,

But it seems People has much Time for making Jokes.......I can't lough about.

Nice Week Gentleman!

@Frunse:

Regarding your core question - if I understood it corrrectly -
Black Lion clock on the Word Clock input of MUTEC MC-4 ahead of the HDSPe AES).

The HDSPe AES uses RME's SteadyClock technology, rated at >50 dB jitter suppression at 2.4 kHz (see section 26.3 of the technical specs). That's the same published figure as RME's reference converter, e.g. the ADI-2 Pro FS R BE.
RME's own measurements show this isn't just a theoretical figure: they tested it with deliberately degraded input signals, artificial jitter up to 500 ns, far beyond anything a real-world clock source would ever produce, and verified the results of jitter compression. See this very informative RME video to SteadyClock FS: https://www.youtube.com/watch?v=Ti0aHW-zYcs

We don't actually know the self-jitter figures of the HDSPe AES's own circuitry or of the Black Lion clock in any comparable, apples-to-apples way; isolated picosecond specs from different manufacturers, measured under different conditions, aren't meaningfully comparable anyway.

What we do have is RME's own jitter suppression and THD+N data, which gives good reason to trust that this part of the chain is handled very well already. Given that even RME's jump from the original ADI-2 Pro to the FS version, a roughly 20 dB improvement in jitter suppression, is generally agreed not to have produced an audible difference, it's hard to make the case that an external clock would move the needle further upstream of that.

So feeding a very low-jitter external clock into the HDSPe AES's Word Clock input is unlikely to make any measurable or audible difference at the end of the chain, since the card already handles that job to a standard that's apparently already beyond what's perceptible.

For the MUTEC MC-4, it depends on what you actually need it for:

• As a reclocker ahead of the HDSPe AES: same reasoning applies, the card downstream (HDSPe AES) already does that job.

• As a format converter, if your sources use different digital formats (PC, TV, streaming box) and sample rates. That's a genuinely useful role. Its internal clock is sufficient for that purpose.

• As a sample rate converter, if you want to unify everything to one sample rate: that's a fair reason to use it. But worth noting that upsampling is purely digital number-crunching, interpolating new sample values from existing ones. There's no waveform yet at this stage, so clock quality doesn't affect the calculated result. The MC-4's internal clock is entirely sufficient; an external clock (-> Black Lion) adds nothing here. And whatever comes out of that step gets reclocked again by the HDSPe AES afterward anyway.

Worth keeping in mind for the DAC stage: if you were ever to use RME's own reference converters (e.g. ADI-2 Pro FS R BE or the new successors), their internal femtosecond clock is also factored into the actual conversion itself, giving a slight further edge there.
That said, the jump from the standard ADI-2 Pro to the FS version, despite the measurable jitter improvement, wasn't reported as an audible difference by users, so even that edge appears to be more of a measurable refinement than something you'd actually hear.

Bottom line: The picosecond self jitter "war of numbers" is useless without measuring compareable results.
RME's SteadyClock FS technology already delivers excellent jitter suppression, the kind that's measurable but no longer audible. My recommendation in this scenario: I'd trust that and avoid overcomplicating the setup. The MC-4 makes sense if you want to consolidate everything to AES and a single sample rate, that's its real value here. The Black Lion Word Clock delivers IMHO no real/audible value.

The resulting setup

different audio sources
   |
MC-4 (roles: SRC, format converter, clock master)
   |
HDSPe AES (SteadyClock, jitter suppression, ideal clock slave due to SteadyClock technology)
   |
Connected D/A converter (clock slave)
Side note: ideal would be RME reference converter with SteadyClock FS/EX, utilizing the internal clock for D/A conversion additionally)

Thank you thats an answer i can count on!
So thats what i want to know.
The MC-4 is still usefull for me at the Momennt, to provide 96kHz in to RME Card so it don't need to bother about different Sources like 48 kHz or 192kHz to feed in.

I read here that the newest Windows Driver support different Samplerates on Driver side already but i use Linux and on Linux i didn't have official supported Driver.

I am also not sure how CamillaDSP on windows 10 (ASIO) could handle it, need to test it by my self, no problem.

CamillaDSP was already running on that RME Card with Windows 10 and the official Driver, not the newes here mentiond.

But at that point i used an PC with HDMI Out for testing and it was set to 96kHz fixed Output smaplerate, i used only Wav files for testing. That worked well already.

But for my Bluerayplayer , TV i need to use HDMI Out with different Samplerates comming from Source.

So now i know that i don't need to buy any Clocks on that Setup but need to figure it out how to handle the different Samplerates in between the best way.

CamillaDSP offers something for that in Software but only on Linux aviable, but thats also makes sometimes Problems i read if you updates Softwareversions and so on, so an external option i would prefer, switch on and forget style.

If i get an Surroundprocessor (Hyperion DPR-16 or Nuprime HX16-AES) in the future may be it handle it already by his own not sure yet.

And yes i would ike to use RME DAC's but the most multichannel Devices have not an Multichannel AES input or you have to pay for also for a lot of Inputs that i don't need

I was looking into RME M-32 DA Pro II-D but it has only DANTE or ASE67 and Madi inputs such a Device with AES in and i am on youre Side
And for me i didn't want to add an AES to DANTE converter i would than go for an HDSPe AoX D but i didn't

HAPI mk3 cost near double that and has only 16 Channel out but it has AES inputs.

And i will stay with AES it has 192kHz cability and didn't shrink the Channel Counts on high Samplerates used vor Ages and i can thrust on.

I need to feed Wires to the Speaker and it makes no differences for me if its an Ethernet Cable or an AES or XLR Cable

The upcomming Surroundprocessor has also AES and DANTE capabillitys for 16 Channels Out.

Please refrain from telling other forum members what to do or referring to them as "trolls". Thank you.

If at all, jitter is an issue during conversion. The HDSPe AES does not have any converters.

how i could use the WORDCLOCK in and Out

In and Out at the same time is not possible.

If I understand you right, the interface as master sending Wordclock out to the Mutec and then the improved wordclock back into the interface ?
This is not possible.
The interface is either master clock-mode internal, or slave clock-mode external BNC

If the Mutec can be master clock then the interface can be slave sync to it, no problem, but no improvement as the interface use Steady clock technology.

Btw Jitter was a thing 30 years ago.
2002 RME introduced Steady clock and the hole Jitter problematic was history.

Regarding your core question - if I understood it corrrectly -
Black Lion clock on the Word Clock input of MUTEC MC-4 ahead of the HDSPe AES.

The HDSPe AES uses RME's SteadyClock technology, specified at >50 dB jitter suppression at 2.4 kHz (see section 26.3 of the technical specs). That's the same published figure as RME's reference converter, e.g. the ADI-2 Pro FS R BE.

RME's own measurements show this isn't just a theoretical figure: they tested it with input signals carrying deliberately injected jitter of up to 500 ns, far beyond anything a real-world clock source would ever produce, and demonstrated the resulting jitter compression.
See this RME video on SteadyClock FS:
https://www.youtube.com/watch?v=Ti0aHW-zYcs

We don't actually know the self-jitter figures of the HDSPe AES's own circuitry or of the Black Lion clock in any comparable, apples-to-apples way; isolated picosecond specs from different manufacturers, measured under different conditions, are not directly comparable.

What we do have is RME's own jitter suppression and THD+N data, which gives good reason to trust that this part of the chain is handled very well already.
Given that the jump from the original ADI-2 Pro to the FS version, roughly a 20 dB improvement in jitter suppression, is widely reported by users as not producing an audible difference, it's hard to make the case that an external clock (here Black Lion) would move the needle further upstream of that.

So feeding a very low-jitter external clock into the HDSPe AES's Word Clock input is unlikely to make any measurable or audible difference at the end of the chain, since the card already handles that task to a standard that's apparently already beyond the threshold of audibility.

For the MUTEC MC-4, it depends on what you actually need it for:

• As a reclocker ahead of the HDSPe AES: same reasoning applies, the card downstream (HDSPe AES) already does that job.

• As a format converter, if your sources use different digital formats (PC, TV, streaming box) and sample rates. That's a genuinely useful role. Its internal clock is sufficient for that purpose.

• As a sample rate converter, if you want to unify everything to one sample rate: that's a fair reason to use it. But worth noting that upsampling is purely digital number-crunching, interpolating new sample values from existing ones. There's no waveform yet at this stage, so clock quality doesn't affect the calculated result. The MC-4's internal clock is entirely sufficient; an external clock (-> Black Lion) adds nothing here. And whatever comes out of that step gets reclocked again by the HDSPe AES afterward anyway.

Whilst jitter suppression in the digital domain is certainly important if any AD or DA converters are connected to these devices, it is most crucial at the point where AD and DA conversion takes place. Here, the HDSPe AES in front of the DAC stage in your setup delivers excellent performance in terms of jitter suppression, and the use of RME Reference converters with SteadyClock FS/EX would go one step further by incorporating the internal clock.

That said, the jump from the standard ADI-2 Pro to the FS version, despite the measurable jitter improvement, was not reported as audible by users, so even that edge appears to be more of a measurable refinement than something you'd actually hear.

Bottom line: Comparing picosecond self-jitter specs across manufacturers is not meaningful without a common, comparable measurement basis. RME's SteadyClock FS technology delivers jitter suppression that is measurable but, by available reports, not audible. My recommendation: trust that and avoid overcomplicating the setup. The MC-4 makes sense if you want to consolidate everything to AES and a single sample rate, that's its real value here. The Black Lion Word Clock adds IMHO no audible value in this chain.

The resulting setup

different audio sources
   |
MC-4 (roles: SRC, format converter, clock master)
   |
HDSPe AES (SteadyClock, jitter suppression, ideal clock slave due to SteadyClock technology)
   |
Connected D/A converter (clock slave).
Side note: ideal would be an RME reference converter with SteadyClock FS/EX, utilizing the internal clock for D/A conversion additionally.