R/E/P Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: [1]   Go Down

Author Topic: Dan: Please Check My Homework  (Read 3982 times)

maarvold

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 853
Dan: Please Check My Homework
« on: October 13, 2006, 12:12:06 pm »

Dan,

I have a composer friend who is (in my opinion) needlessly suffering as he tries to sort out all the opinions and information about convertors and clocking.  I wrote this condensed version.  Would you please correct, or augment, anything that needs it?  

Thank you.  


1. Any reasonably well-designed internal clock will provide higher integrity (lower jitter) clocking than any external clock.  Period.  End of book.  

2.  When using multiple digital devices they must all be clocked from the same master clock source to work together without pops, clicks and other anomalies in the digital domain.  

3. Clock integrity comes into play at the A-to-D conversion--where data and resolution can be lost (and never recovered)--and at the D-to-A conversion--where it can change the sonic character and resolution of the playback (can you say "wide and deep soundstage"?).  

4. When data is moving between devices in the digital domain, clock integrity does not matter nearly as much AS LONG AS ALL DEVICES ARE PROPERLY RESOLVING TO THE SAME MASTER CLOCK AND CABLED/TERMINATED PROPERLY.  

5. In a 'perfect world' one should record slaved to the [reasonably well-designed] internal clock of the A-to-D and play back slaved to the [reasonably well-designed] internal clock of the D-to-A.  

6. For many people creating and living in a 'perfect world' is too much of a pain in the ass.  Enter: the high quality external clock.  The advantage is that you set everything up one time (cabling, termination, master clock source on each device), everything is happy, you never have to think about clock ever again and it sounds good (although not quite as good as if you follow #5).  You can also have people bring over other digital devices (DA-88 or a laptop/GigaStudio/Hammerfall card, for example), resolve them to your clock and transfer digital data back and forth perfectly all day long without worrying.
Logged
Michael Aarvold
Audio Engineer

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: Dan: Please Check My Homework
« Reply #1 on: October 13, 2006, 02:27:11 pm »

maarvold wrote on Fri, 13 October 2006 17:12

Dan,

I have a composer friend who is (in my opinion) needlessly suffering as he tries to sort out all the opinions and information about convertors and clocking.  I wrote this condensed version.  Would you please correct, or augment, anything that needs it?  

Thank you.  


1. Any reasonably well-designed internal clock will provide higher integrity (lower jitter) clocking than any external clock.  Period.  End of book.  

2.  When using multiple digital devices they must all be clocked from the same master clock source to work together without pops, clicks and other anomalies in the digital domain.  

3. Clock integrity comes into play at the A-to-D conversion--where data and resolution can be lost (and never recovered)--and at the D-to-A conversion--where it can change the sonic character and resolution of the playback (can you say "wide and deep soundstage"?).  

4. When data is moving between devices in the digital domain, clock integrity does not matter nearly as much AS LONG AS ALL DEVICES ARE PROPERLY RESOLVING TO THE SAME MASTER CLOCK AND CABLED/TERMINATED PROPERLY.  

5. In a 'perfect world' one should record slaved to the [reasonably well-designed] internal clock of the A-to-D and play back slaved to the [reasonably well-designed] internal clock of the D-to-A.  

6. For many people creating and living in a 'perfect world' is too much of a pain in the ass.  Enter: the high quality external clock.  The advantage is that you set everything up one time (cabling, termination, master clock source on each device), everything is happy, you never have to think about clock ever again and it sounds good (although not quite as good as if you follow #5).  You can also have people bring over other digital devices (DA-88 or a laptop/GigaStudio/Hammerfall card, for example), resolve them to your clock and transfer digital data back and forth perfectly all day long without worrying.


You said:
Any reasonably well-designed internal clock will provide higher integrity (lower jitter) clocking than any external clock. Period. End of book.

I say:
That is correct.

You said:
When using multiple digital devices they must all be clocked from the same master clock source to work together without pops, clicks and other anomalies in the digital domain.

I say:
As a rule, yes, though one can “realign” with use of sample rate conversion devices.

You said:
Clock integrity comes into play at the A-to-D conversion--where data and resolution can be lost (and never recovered)--and at the D-to-A conversion--where it can change the sonic character and resolution of the playback (can you say "wide and deep soundstage"?).

I say:
Yes. At the AD the change is there forever. At the DA the sonic change depends on the DA, so better DA will fix problems (not so with AD, where the sonic alteration can not be fixed after the fact). Sample rate conversion also gets effected by jitter, though the sonic art effects are different then those of an AD and DA.

You said:
4. When data is moving between devices in the digital domain, clock integrity does not matter nearly as much AS LONG AS ALL DEVICES ARE PROPERLY RESOLVING TO THE SAME MASTER CLOCK AND CABLED/TERMINATED PROPERLY.

I say:
Yes. Moving data in the digital domain can tolerate a lot more jitter then required for conversion clocks, and one can tolerate jitter up to a certain point with NO deterioration at all. The idea is simple – move 0’s and 1’s from one place to the other, and as long as all the 0’s stay 0’s, and 1’s stay 1’s you have a perfect transfer. We do such stuff a lot (the internet, reading CD’s and so on). At some point when the jitter is huge, the data transfer will cause problems, but that is more then 100 times the jitter required for great AD sound.

You said:
In a 'perfect world' one should record slaved to the [reasonably well-designed] internal clock of the A-to-D and play back slaved to the [reasonably well-designed] internal clock of the D-to-A.

I say:
There are times when you need to use external clocks (such as when you use a lot of gear, many AD’s not all in one chassis and so on). As a rule, using external clock is more jitter, but the overriding requirement to sync many devices may require use of external clock. I would say: “use internal AD clock WHENEVER POSSIBLE, and when using external clocking, go for gear that has very good internal VCXO circuitry. With good hardware, the degradation is not great.
For DA: I would not send a DA an external clock, because by the time it gets there, it will be too jittery, so just find a good DA with good clock implementation. In principle, a DA that is the clock source (internal fixed crystal) will be best, but few devices have such and interface.

You said:
6. For many people creating and living in a 'perfect world' is too much of a pain in the ass. Enter: the high quality external clock. The advantage is that you set everything up one time (cabling, termination, master clock source on each device), everything is happy, you never have to think about clock ever again and it sounds good (although not quite as good as if you follow #5). You can also have people bring over other digital devices (DA-88 or a laptop/GigaStudio/Hammerfall card, for example), resolve them to your clock and transfer digital data back and forth perfectly all day long without worrying.

I say:
Maybe so, but I see a lot of problems when people try to set things to external. Much of the hardware can receive data directly without a clock cable at all. Say you have an AD operating in internal clock mode, and you wish to send the data to a hard drive. That does not call for a clock connection. The clock rate is already there, on the same digital audio cable that carries the music (such as AES or SPDIF). The line digital receiver IC locks to the preamble, it separates the L data from the R data and also from the information data bits… it does not need an additional clock. Often, including such additional clock connection creates more problems then it solves (time difference between preamble and word clock can cause problems).
So often you can clock the additional devices to the most critical clock (the AD internal clock, when possible), and the destination can receive the digital data from various devices WITHOUT a need for a clock cable.    

Regards
Dan Lavry
http://www.lavryengineering.com
Logged

maarvold

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 853
Re: Dan: Please Check My Homework
« Reply #2 on: October 13, 2006, 02:33:27 pm »

This is EXACTLY what I was hoping for.  Thank you very much.  
Logged
Michael Aarvold
Audio Engineer
Pages: [1]   Go Up