The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Anything wrong with sending clock through MADI optical? DSP Audio Systems
Old 18th December 2010
  #1
Lives for gear
 

Anything wrong with sending clock through MADI optical?

Hi Slutz,

I have an SSL Alphalink SX which is my clock source. I have been running optical MADI to my RME HDSPe card, and running a BNC word clock cable separately. The RME card has a "piggy-back" card for the WC BNCs and MIDI I/O, which attaches to the main card with a ribbon cable jumper but does not take an actual PCI slot. Still, it takes up the space of another card. If one doesn't need MIDI or WC, it can be removed and the main card works fine.

Well, I'm getting a new Mac next week and the video card is way thicker than the one in my current machine, so with all 3 PCIe slots filled, I won't have room in there for the piggy-back card like I do now. I wouldn't feel good about letting it hang free inside or outside of the box.

So the question is, will sending the clock down the optical cable be just as good using the BNC? Maintaining an accurate clock is very important to me.
Old 18th December 2010
  #2
Lives for gear
 

I used to just wrap mine in cellophane and leave it hanging outside the computer chassis. Worked absolutely fine.
Old 18th December 2010
  #3
Lives for gear
 

Ha! I was afraid somebody would say that!heh

Guess I wouldn't have a problem with that if I knew there was a "hole" to get it out there. I only saw the 8-core model's insides, but I'm assuming the 12-core is the same (which is what I purchased). It's pretty tight.

Got the RME MADI, the UAD Quad, and the TC Powercore, for reverb- I don't think I can live without any of those..

As expensive as it was, I don't think I could bring myself to drill a hole in the damn thing when it gets here.
Old 18th December 2010
  #4
Lives for gear
 
synthoid's Avatar
 

There are two separate uses for word clock; one of them requires very tight sync / low jitter, and the other one is not nearly so critical.

The first is when you synching two converters together and you want them to be absolutely phase-locked. For example, you are going to be tracking with 32 inputs configured as 2 16-channel AD converters, and you want perfect phase alignment of all the inputs. In this kind of case using BNC word clock is the best bet.

The other is when you are simply trying to get digital sync for an all-digital transfer. For example, you are sending ADAT or AES/EBU from one device to another; the master and slave in the transfer have to achieve sync in order to get a clean transfer of bits. The timing in this case is not nearly so critical. You either have sync / lock or you don't, and in the case of the RME HDSPe stuff, it is very good about telling you whether you do or don't.

So, what I'm saying is that since the HDSPe doesn't have converters (I've got that right, yes? It's all-digital, like the RayDAT I think), it doesn't need a word clock at all, really. It can slave perfectly to almost any kind of incoming digital signal. If you give it a word clock, it will verify that it is in fact synched with the actual digital carrier(s) entering the card.

Long story short, there is no harm dropping the WC connection to the HDSPe unless I'm missing something about your setup.

-synthoid
Old 18th December 2010
  #5
Lives for gear
 

Quote:
Originally Posted by synthoid View Post
So, what I'm saying is that since the HDSPe doesn't have converters (I've got that right, yes? It's all-digital, like the RayDAT I think), it doesn't need a word clock at all, really. It can slave perfectly to almost any kind of incoming digital signal. If you give it a word clock, it will verify that it is in fact synched with the actual digital carrier(s) entering the card.

Long story short, there is no harm dropping the WC connection to the HDSPe unless I'm missing something about your setup.

-synthoid
Hey thanks for all that synthoid! Yes you have it right about my setup. The RME card just gets MADI in and out of the computer. I think it does have a pair of analog outs (headphones?), but I don't use them. I wish RME had put that on the piggy-back card instead of the WC.

I will eventually expand the MADI setup and then I will no doubt have to worry about the WC again. I want to keep the RME card because it uses the incredibly powerful Totalmix software, which I never actually used until recently when my digital console died.

It also just now occurred to me that I can A/B the sound of the 2 clock inputs with the little RME setup utility in the dock! Think I'll do that..

-Not that I don't believe what you wrote, but as with anything someone tells me about sound, "trust, but verify". heh
Old 19th December 2010
  #6
Lives for gear
 
synthoid's Avatar
 

Quote:
Originally Posted by Killahurts View Post
Hey thanks for all that synthoid! Yes you have it right about my setup. The RME card just gets MADI in and out of the computer. I think it does have a pair of analog outs (headphones?), but I don't use them. I wish RME had put that on the piggy-back card instead of the WC.
nah. The WC really has no use in that card, except for the case where you are for example sending audio out via MADI but there is no incoming audio and therefore no word clock in the MADI signal for the computer to sync to. I.e., a one-way digital transfer out of the computer that must be synched with some outside word clock. That's a pretty unusual situation, especially with MADI.

When you "A/B", you'll still be hearing the DAC in your SSL or whatever DAC you're using to monitor, which if I understand correctly is being clocked internally, as it should be. So you're really only A/B-ing the digital transfer, i.e., verifying that the MADI word clock and the BNC WC are in sync to the point that they give you a glitch-free transfer. The HDSPe Settings app should also be able to tell you this (look for 'Sync' versus 'Lock' or 'No Lock'). Always good to verify! (But it's not worth losing a slot over. heh)

-synthoid
Old 19th December 2010
  #7
Lives for gear
 

Quote:
Originally Posted by synthoid View Post
nah. The WC really has no use in that card, except for the case where you are for example sending audio out via MADI but there is no incoming audio and therefore no word clock in the MADI signal for the computer to sync to. I.e., a one-way digital transfer out of the computer that must be synched with some outside word clock. That's a pretty unusual situation, especially with MADI.

When you "A/B", you'll still be hearing the DAC in your SSL or whatever DAC you're using to monitor, which if I understand correctly is being clocked internally, as it should be. So you're really only A/B-ing the digital transfer, i.e., verifying that the MADI word clock and the BNC WC are in sync to the point that they give you a glitch-free transfer. The HDSPe Settings app should also be able to tell you this (look for 'Sync' versus 'Lock' or 'No Lock'). Always good to verify! (But it's not worth losing a slot over. heh)

-synthoid
Yes indeed, you're right! Thanks for all the info.

Well, the Mac Pro came today instead of Tuesday, when it was scheduled for delivery.. don't know what it is about FedEx lately, they're really on the ball!

And there is indeed no room in there whatsoever for the piggyback card. I got a 12-core and it has that Radeon 1GB beast of a video card. Even one of the 3 available PCI slots is going to be tight!
Old 19th December 2010
  #8
Gear Nut
 

I would use the WC input...it is different than clock sent down an optical line...I would AB them and make your decision from there...the clock science in a previous post seems to have a bit of an odour...although both ways ahould be fine as the RME has a PLL...just AB it and you'll know
Old 19th December 2010
  #9
Lives for gear
 

Clock sent down MADI is VERY prone to jitter.
in note to the original post, no holes need to be drilled, the ribbon cable just slips through the pci slot hole
Old 19th December 2010
  #10
Lives for gear
 
synthoid's Avatar
 

Quote:
Originally Posted by DannyMac View Post
Clock sent down MADI is VERY prone to jitter.
Gearslutz should be called clockaholics anonymous, I swear.

MADI is well designed. It sends a bunch of channels of digital audio down a wire. It is absolutely not necessary or smart or anything to send a separate word clock cable alongside it to help clock the data out at the other end. If it were, then when you see network TV guys using long MADI runs from stages to their remote trucks, you'd see both a MADI cable and a separate coaxial wire with BNCs on the ends carrying the word clock. But you don't. For a reason: 'cause MADI is designed properly.

There are no converters being clocked at the receiver end of the MADI connection. Only digital data being clocked out of an absolutely ordinary digital transmission protocol. Like firewire. Like ethernet. Like SONET. Like USB. None of those things needs a separate word clock. Neither does MADI.

-synthoid
Old 19th December 2010
  #11
Lives for gear
 
synthoid's Avatar
 

Quote:
Originally Posted by TheKnow View Post
the clock science in a previous post seems to have a bit of an odour.
Would you like to elaborate, Mr Nick/Knack/Know? Welcome back, by the way. Next time you sign up, you should call yourself Freddy.

-synthoid
Old 19th December 2010
  #12
Lives for gear
 
yrplace's Avatar
 

Quote:
Originally Posted by Killahurts View Post
Hi Slutz,

I have an SSL Alphalink SX which is my clock source. I have been running optical MADI to my RME HDSPe card, and running a BNC word clock cable separately. The RME card has a "piggy-back" card for the WC BNCs and MIDI I/O, which attaches to the main card with a ribbon cable jumper but does not take an actual PCI slot. Still, it takes up the space of another card. If one doesn't need MIDI or WC, it can be removed and the main card works fine.

Well, I'm getting a new Mac next week and the video card is way thicker than the one in my current machine, so with all 3 PCIe slots filled, I won't have room in there for the piggy-back card like I do now. I wouldn't feel good about letting it hang free inside or outside of the box.

So the question is, will sending the clock down the optical cable be just as good using the BNC? Maintaining an accurate clock is very important to me.
Not exactly the same situation but on our remote truck we send wordclock via MADI down as much as 1500 feet of fiber to our remote preamp racks using a RME 648 to convert the clock to BNC which is then sent thru a distro to 6-8 Aphex preamps. Works like a charm, so I'd think using the optical clock input should work fine on your system.

Mark Linett
Old 19th December 2010
  #13
Gear Nut
 

Quote:
Originally Posted by DannyMac View Post
Clock sent down MADI is VERY prone to jitter.
in note to the original post, no holes need to be drilled, the ribbon cable just slips through the pci slot hole
Yes and there is a good reason WHY your RME card has a BNC Word Clock in/out and takes up space...but by all means, do a quick test yourself and see. That's the best way to actually KNOW.

Oh yeah BTW TV broadcasts aren't the finest way to judge anything audio...I haven't heard a TV braodcast that sounds amazing...ever...just sayin'. Good luck!
Old 19th December 2010
  #14
Gear Nut
 

Oh and as well...since the OP is using MADI, we can assume that there is multitrack info heading back out to the DA portion if the converter, which, where Sympthoid admits, Word Clock is important...it's important.
Old 19th December 2010
  #15
Lives for gear
 
NetworkAudio's Avatar
Quote:
Originally Posted by TheKnow View Post
I would use the WC input...it is different than clock sent down an optical line...I would AB them and make your decision from there...the clock science in a previous post seems to have a bit of an odour...although both ways ahould be fine as the RME has a PLL...just AB it and you'll know
It all depends, if the clock originates at the AD end of the madi stream, and you are click free, I would rather use the clock in the madi stream than a 100meter WC run. If you are dealing with say the deltalink, you are forced to send WC separate from the madi stream. When I clock our delta this way, I keep the Mutec Iclock in the AD rack and send the (75meter) wc to the delta.
Old 19th December 2010
  #16
Gear Nut
 

Quote:
Originally Posted by klaukholm View Post
It all depends, if the clock originates at the AD end of the madi stream, and you are click free, I would rather use the clock in the madi stream than a 100meter WC run. If you are dealing with say the deltalink, you are forced to send WC separate from the madi stream. When I clock our delta this way, I keep the Mutec Iclock in the AD rack and send the (75meter) wc to the delta.

Well yes I'd say anything over 5 meters and I wouldn't be running a BNC cable...we're assuming the OP is not doing location recording and that this is his/her project studio, in which case a BNC is the way to go....but for long runs, yeah I'm not even sure how long is OK for BNC anyway
Old 19th December 2010
  #17
Lives for gear
 

My WC cable is about 35 feet, so just over 5 meters (it goes to a machine room).

I've only had time today to play some songs on iTunes, but I was able to use the little RME setup app, and switch back and forth between the 2 clock inputs.

At first I thought there was no difference because it is so subtle, but I listened a little deeper and did hear a difference, but it was incredibly slight. The WC sounded a little deeper and more open. I could hear it best with things like reverb tails, etc. The MADI sounded just a little more 2-dimensional and pinched, but again, it was VERY small, splitting hairs really.

I would have to think though, since iTunes is only playing 2 tracks, that the difference might be more pronounced with multi-track material, so I'll test that tonight or tomorrow.

To be honest, this result surprised me. I thought they would be exactly the same.
Old 20th December 2010
  #18
Gear Nut
 

Quote:
Originally Posted by Killahurts View Post
My WC cable is about 35 feet, so just over 5 meters (it goes to a machine room).

I've only had time today to play some songs on iTunes, but I was able to use the little RME setup app, and switch back and forth between the 2 clock inputs.

At first I thought there was no difference because it is so subtle, but I listened a little deeper and did hear a difference, but it was incredibly slight. The WC sounded a little deeper and more open. I could hear it best with things like reverb tails, etc. The MADI sounded just a little more 2-dimensional and pinched, but again, it was VERY small, splitting hairs really.

I would have to think though, since iTunes is only playing 2 tracks, that the difference might be more pronounced with multi-track material, so I'll test that tonight or tomorrow.

To be honest, this result surprised me. I thought they would be exactly the same.
that sounds about right to me...subtle, but not really for audio engineers.
Old 20th December 2010
  #19
Lives for gear
 
synthoid's Avatar
 

Quote:
Originally Posted by Killahurts View Post
To be honest, this result surprised me. I thought they would be exactly the same.
You are playing itunes with your SSL as word clock master (always). In one case you send the SSL WC to the RME in order for it to use it in transmitting the digital audio over your MADI link. In the other case, you let the RME derive the clock from the MADI link in order to send the outbound digital audio. Anyone with an understanding of how this equipment works knows that the digital audio going over MADI from the RME to the SSL is exactly, bit for bit, the same in the two cases. iTunes has not changed, and there is no mechanism anywhere in that signal chain for a single bit of audio to be changed, period, except in the case of an actual dropout (i.e., malfunction), which you would hear as a loud pop etc. The SSL, which is playing the audio out, is using exactly the same internal word clock in both cases.

What is your theory about what can be causing an audible difference? I suppose -- as always -- it is out the question for you to record this difference and let the rest of us hear it...

Mr. Know is back, shamelessly hard-selling word clocks into situations where they are grossly inappropriate. He knows better of course, but then again it was long ago established that he is a fraud and a shill. That's why he has been banned twice from gearslutz. As per Jules' request in the Moan Zone, I have reported Mr. Know's reappearance in clock threads, but so far no moderator intervention that I can see.

-synthoid
Old 20th December 2010
  #20
Gear Nut
 

Quote:
Originally Posted by synthoid View Post
You are playing itunes with your SSL as word clock master (always). In one case you send the SSL WC to the RME in order for it to use it in transmitting the digital audio over your MADI link. In the other case, you let the RME derive the clock from the MADI link in order to send the outbound digital audio. Anyone with an understanding of how this equipment works knows that the digital audio going over MADI from the RME to the SSL is exactly, bit for bit, the same in the two cases. iTunes has not changed, and there is no mechanism anywhere in that signal chain for a single bit of audio to be changed, period, except in the case of an actual dropout (i.e., malfunction), which you would hear as a loud pop etc. The SSL, which is playing the audio out, is using exactly the same internal word clock in both cases.

What is your theory about what can be causing an audible difference? I suppose -- as always -- it is out the question for you to record this difference and let the rest of us hear it...

Mr. Know is back, shamelessly hard-selling word clocks into situations where they are grossly inappropriate. He knows better of course, but then again it was long ago established that he is a fraud and a shill. That's why he has been banned twice from gearslutz. As per Jules' request in the Moan Zone, I have reported Mr. Know's reappearance in clock threads, but so far no moderator intervention that I can see.

-synthoid
This post is very clearly based in anger that the OP doesn't appear to adhere to your way of thinking.

Also...I have no idea what you are going on about with regards to "hard selling word clocks"....perhaps you could edit that stuff out..out of common courtesy...your call...however

To address your point...you are pretending that the card has no influence in the audio chain...so if that were the case, then using your thought process, then he should just leave his card clocked internally and the SSL clocked internally...

which makes no sense at all really....maybe I'm not understanding your science correctly.

and perhaps you could explain why jitter occurs, when as you say if there is error in digital transfer, it would only manifest itself as a loud pop or crackle?...again absurd to me, but I'm willing to hear the reasoning.
Old 20th December 2010
  #21
Lives for gear
 
synthoid's Avatar
 

Quote:
Originally Posted by TheKnow View Post
which makes no sense at all really....maybe I'm not understanding your science correctly.

and perhaps you could explain why jitter occurs, when as you say if there is error in digital transfer, it would only manifest itself as a loud pop or crackle?...again absurd to me, but I'm willing to hear the reasoning.
Oh, I guess you're right. What would I know after all? I'm only a computer engineer with a doctorate and professorship under my belt, a couple of decades of first-hand design experience, a dozen patents, etc. Whereas you're -- TheKnack. I mean, TheNick. I mean, you're TheKnow. Who can compete?

So let me bow out at long last Mr. Know: you win.

Bye.

-synthoid
Old 20th December 2010
  #22
Gear Nut
 

I really have no idea what that's about, but sure...as you wish...

Killa I'm interested in Knowing how your multitrack testing goes. Keep us posted as I'm sure there are others with similar questions
Old 20th December 2010
  #23
Lives for gear
 

Hey guys, no need for animosity here. This is all just esoteric sound voodoo, how serious can we really get about it? heh

I have to admit, it is possible that I was just imagining things, that's how small the difference (if any) was. I'm only a human. Hell, just clicking the mouse may have changed the way I heard the next thing.

For me, a multi-track test is the way to really know- and not just flipping back and forth, but actually working with one and the other for a few minutes, each. I will do that and report back after I get the new computer worked in..

For the record, I'm listening on a very well-tuned set of ADAM S3As, in a very well-tuned control room.. but so what?
Old 20th December 2010
  #24
Lives for gear
 
NetworkAudio's Avatar
Once the data is aquired the system is much more jitter tolerant.
Otherwise we would be completely screwed when playing back from a hard drive (massive jitter)
With a properly designed clock on the DA side, this is solved.
Old 20th December 2010
  #25
Gear Nut
 

Quote:
Originally Posted by Killahurts View Post
Hey guys, no need for animosity here. This is all just esoteric sound voodoo, how serious can we really get about it? heh

I have to admit, it is possible that I was just imagining things, that's how small the difference (if any) was. I'm only a human. Hell, just clicking the mouse may have changed the way I heard the next thing.

For me, a multi-track test is the way to really know- and not just flipping back and forth, but actually working with one and the other for a few minutes, each. I will do that and report back after I get the new computer worked in..

For the record, I'm listening on a very well-tuned set of ADAM S3As, in a very well-tuned control room.. but so what?
It'll be good to know if you hear a difference in that circumstance as that is the way most of us work.

For the record, I would have absolutely no problem using the clock coming down the optical line...and have many times. I personally hear a difference when clocked via BNC and when I can I set up that way.
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
DeyBwah / So much gear, so little time
6
Yannick / Remote Possibilities in Acoustic Music and Location Recording
24
GigaBoy / So much gear, so little time
0

Forum Jump
Forum Jump