API A2D superclock questions

[narco]

hi slutz

I am demoing an API A2D in my studio this week.

At first when I tried to clock it to my Mytek clock (from one of their 8 channel 96 units) i ran into weird issues, the "lock" LED lit, but the AES level dropped away, and the spdif went completely silent.

And so I looked at the instruction manual (of course i didn't do this first) and found out that the clock input only accepts "superclock", now I'm not a clocking expert so I have a few questions:

1> how do I use it in my studio that is otherwise all synced to a mytek clock?

2> Surely the API dosen't have to be the master (in a superclock-less studio)?

3> if I plug the spdif outputs of the A2D into a spdif in on my interface it doesn't sync automtically to my mytek does it? I mean you can only get a clock off an input not an output?

4> Using the spdif outputs without clocking will cause issues, will it not?

Thanks for you thoughts

Narco

[narco]

hmmm in another thread an API guys says this:

The choice of the word Superclock was probably not the correct one in this case. It was not intended to refer to any manufacturer's specific device or technique, but rather to the use of a stable external clock reference that produces the standardized clock frequencies. (kind of like saying Kleenex) Any or all of the available external clocks should work just fine with the A2D - in fact WE use the Rosetta 800 here to check all the outputs of all A2D's as they go through test.

but another guy has what seems to be the same issue as me

The A2D will not work with my Yamaha 01V96, but when I disconnected the mixer the A2D worked....could be a clocking problem, just don't know, never happened before with any other device and I need the 01V96 for playback.

both quoted from thread: http://gearslutz.com/board/showthread.php?t=71638

It seems to work fine with no clock connected, as soon as you connect an external (word) clock, it says it is locked but it stops working..

The meters stay the same but there is waaaaay less level in the AES signal, and no SPDIF signal. It all comes back the moment I unplug the wordclock..

Everything else in my studio that is clocked works fine, including when using the exact same plug that I am plugging into the API

anyone?
narco

[kurt]

"A coaxial BNC accepts external "Super Clock" which will illuminate the front panel EXT SYNC LED when locked."

This is weird. Super Clock was invented by Digidesign & is WC X 256.
(Word Clock x 256). I’m on a Mix system & all devices have Super Clock I/O. Don’t know if HD has that same clocking. My system uses Rosendahl Nanosyncs that has both WC & Super Clock outputs. Check if Mytek provides some switch changing WC to Super clock as Nanosyncs do.
Strange that API doesn’t use straight WC I/O?

[Kiwiburger]

Yup - that's the basic problem with the A2D. Bad design call. You'll just have to sync everything else to it.

I'm thinking about buying one anyway. I've got a Lynx AES16 with SRC. I got the SRC option hoping (cross fingers) that it might be able to resolve problems like this. I figure if it can do on-they=fly sample rate conversion between totally different devices running off their own internal clocks, it might be able to deal with rogue boxes like this.

But who the hell knows? I'm not sure I want to be the guinea pig.

[API Sez...]

Replies:

1> how do I use it in my studio that is otherwise all synced to a mytek
clock?

I am not sure what the Mytek requires, so I can't answer this. But maybe the following answers will shed some light on the question.

2> Surely the API dosen't have to be the master (in a superclock-less
studio)?

In it's current state, yes (sort of). The Converter in the A2D runs from a crystal cut to run at exactly the right rate to produce the selected sample rate +/- 0.01 %. It is very stable. It uses two crystals to run at either 24.576 MHz or 22.5792 MHz to produce 48 k or 44.1 k related rates. You can use an external clock source at those rates to lock it to an external clock, but clock frequencies below 20 MHz will not work.

3> if I plug the spdif outputs of the A2D into a spdif in on my
interface it doesn't sync automtically to my mytek does it? I mean you can only get a clock off an input not an output?

No, the way SPDIF works, the receiving unit MUST sync to the incoming data stream, in which a data clock is embedded. So in a sense, the sending unit is always master. Even if they run off the same clock source, this is still the case.

4> Using the spdif outputs without clocking will cause issues, will it
not?

The only issue it could possibly cause is a slight shift in phase between the two clocks. This should not cause any issues with the audio signal, as the receiver must sync to the incoming data stream to receive data, and should never miss any data.
The only time you should care about the phase shift between clocks is if you have two digitizers running concurrently. Then you might notice the phase shift. But that is why the multi-unit link port is built into the A2D unit to sync multiple A2Ds.

While we don't see it as a bad design call but an attempt to maintain the best possible A/D conversion, the lack of standards as the sampling rates have steadily been rising is part of the reason for the troubles between different manufacturer's units. Basic word clock, when upsampled the the highest conversion rates used today, can introduce an unacceptable (to us) level of jitter, which is why we went with a higher base clock rate.

There is a plan to introduce an option on the A2D which can sync to an external word clock. This does somewhat compromise the integrity of the digitizing clock, but for whatever reason, some folks want to do that.
Stay tuned.

Ted
A2D Designer

[kurt]

So, if I understand this right. One could send Superclock from A2D (Master) to Mytek in ext. sync mode & continue from there as usually. Or?

PS.
(Is it 256 WC or not?)

[API Sez...]

>So, if I understand this right. One could send Superclock from A2D
(Master) >to Mytek in ext. sync mode & continue from there as usually.
Or?

Not really. The high frequency sampling clock (Super Clock) from the A2D is only present at the link outputs, and in a format not usable by the Mytek device.
If the Mytek can slave to the SPI or AES/EBU output, then you can make the A2D master the entire downstream connection.
This is always done with the digital audio data stream, but not necessarily with peripheral clocks, etc. For that you would want to get details on how the Mytek device operates when receiving data from an unrelated clock source.

>PS.
>(Is it 256 WC or not?)

No.
Well, sometimes.
When operating at 44.1 or 48 k, the clock is at 512xWC, at 88.2 or 96 k, the clock is at 256xWC, and at 176.4 or 192 k, the clock is at 128xWC. The sampling clock is always fixed at 22.5792 or 24.576 MHz.

Ted
A2D Designer

[narco]

Thanks alot Ted

In the mean time I have figured out a work around.

I take the second lot of digital outputs from the API and put them into my D/A and sync the D/A to the incoming digital signal.

I then used the (interpreted) word clock signal from my D/A to clock my A/D

It seems to be working. I imagine there is more chance for jitter happening this way but it sounds ok so far.

I think I am going to pass on buying the API (even thoough I LOVE the sound of it) untill there is a version that syncs to wordclock, in which case I will be first in line to buy it.

narco

[kurt]

Hello, if you please, I’m too interested in the matter..

So, if I understand this right. One could send Super clock from A2D
(Master) >to Mytek in ext. sync mode & continue from there as usually.
Or?

Not really. The high frequency sampling clock (Super Clock) from the A2D is only present at the link outputs, and in a format not usable by the Mytek device.

So Super Clock has nothing to do with Ddesign Superclock?

If the Mytek can slave to the SPI or AES/EBU output, then you can make the A2D master the entire downstream connection.

But not the other way?

This is always done with the digital audio data stream, but not necessarily with peripheral clocks, etc. For that you would want to get details on how the Mytek device operates when receiving data from an unrelated clock source.

But not sending?

>PS.
>Is this WC x 256 or not?

No.
Well, sometimes.
When operating at 44.1 or 48 k, the clock is at 512xWC, at 88.2 or 96 k, the clock is at 256xWC, and at 176.4 or 192 k, the clock is at 128xWC. The sampling clock is always fixed at 22.5792 or 24.576 MHz.

Ted
A2D Designer

Can you please explain the reason for this & how A2D is connected & communicating with other digital devices?

Thank you very much.
/kurt

[narco]

I'll try to answer these (although I didn't design the unit, unlike ted!)

So Super Clock has nothing to do with Ddesign Superclock?

it is the same thing

But not the other way?

the API has no option of syncing to the digital ins (it has no digital ins), only superclock ins

But not sending?

I'm not sure what this question is refering to. If the mytek is sending a digital signal to an unrelated clock source it has no way of knowing anything about that clock.

Can you please explain the reason for this & how A2D is connected & communicating with other digital devices?

the clocking can only be connected with superclock connections (they are on the same connecters as normal wordclock, only the signal is different (as explained in teds post))

if you do not have superclock in your studio you HAVE to use the API as the master clock source, and have all other digital devices derive their clocks form the digital outputs of the API, as I ended up doing

I would like to add, that whilst it is admirable of API to aim for the lowest jitter possible, I can't help but notice that pretty much all the major manufacturers of digital converters find ways to make the jitter of normal wordclock acceptable, and don't use superclock.

narco

[max cooper]

Originally Posted by narco

if you do not have superclock in your studio you HAVE to use the API as the master clock source, and have all other digital devices derive their clocks form the digital outputs of the API, as I ended up doing

I wonder if we can get any insight as to why this was done this way. I'm certainly curious, but I only know enough about the subject to be dangerous.

[kurt]

You could try to find old: Digidesign Smpte Slave Driver. (SSD)
Can be had used for 50-75$.

It can take WC in external mode & spit Superclock.

[narco]

Originally Posted by kurt

You could try to find old: Digidesign Smpte Slave Driver. (SSD)
Can be had used for 50-75$.

It can take WC in external mode & spit Superclock.

serious? I'll look into that! And you can use it in a non pro tools system?

narco

[Kiwiburger]

Ted - thanks for those excellent replies. Please accept my apologies for stating that this superclock decision was a "bad design call".

Now that I understand the reason for this better, i'm happier to have the low jitter. I've ordered my first one from Mercenary - here's hoping they can supply. If it works out, I'd like to get some more and link them up as designed.

[kurt]

Originally Posted by narco

serious? I'll look into that! And you can use it in a non pro tools system?

narco

I’m quite sure.

Happy New Year.

[narco]

hey kiwiburger

order one from oceania in New Zealand. its cheaper (assuming you're a kiwi)

actually on second thoughts don't, cos I think I'm going to buy their demo one..

narco

[andychamp]

I'm currently using a MOTU896HD as my main/only interface and considering an API A2D to take care of inputs 9 and 10 via the AES.

Right now, there are no other digital devices in my setup, so it would'nt matter which unit is the master.

My main concern is the synching between both units:
- will the synch signal embedded in the AES be enough to synch both machines? (A2D master), or...
- will the A2D synch off the 896HD's clock via the BNC connector? (896 master), or...
- do I need a 3rd-party clock source to provide a master clock in 2 different "dialects", 1 each for the A2D and the 896HD?

[narco]

- will the synch signal embedded in the AES be enough to synch both machines? (A2D master), or...

yes, assuming the motu can sync to an incoming source

- will the A2D synch off the 896HD's clock via the BNC connector? (896 master), or...

-
no, it will sync through the actual digital signal, in the digital audio cable

do I need a 3rd-party clock source to provide a master clock in 2 different "dialects", 1 each for the A2D and the 896HD?

you don't "need" this, although if you want to use a different master clock from the API you will, the digidesign sync box is the only one that does it that I know of

Apparently API is working on an expansion card for the A2D that wil make it use normal wordclock, if you get this (when it is available) you will be able to use any clock source you have available, and sync with word clock using BNC. (The current BNC on the A2D uses superclock onloy and not normal wordclock)

narco

[rosstapha]

Originally Posted by API Sez...

While we don't see it as a bad design call but an attempt to maintain the best possible A/D conversion...

Well... I appreciate your concern for fidelity, but i wish you had more concern for usability.

Due to this design, i'm hardly ever able to use the A2D converters.
Instead, I am forced to use the analog outs through my other ADC.

Furthermore, if, as you say, your clock is so superiour , wouldn't it make more sense to offer a clock out option, rather then a clock in?
(when is that coming, anyway??)

I realise there would be some math involved - dividing your base frequency down to the 44.1-192 range, but wouldn't that be equivalent to the complexity of integrating a wordclock input?

p.s.
Don't get me wrong. Even without the digital outs I LOVE my A2D! Thanks!

[FossilTooth]

Wow. This kind of sucks. I feel like I've been duped.

I like API's on overheads. I guess this all means I can't reliably use the A2D on overheads and use my rosetta 800 for the rest of the drum channels with my 002R setup...

... or I could, but then I wouldn't really be able to take advantage of the extra 2 channels of A/D that I paid good money for.

I will have to second the notion that this was a really bad design call.

It doesn't really consider the needs of the end user. By definition, that's poor engineering!

I guess I could set the accept clock from the A2D, and then feed that clock to the rosetta via ADAT... if the rosetta accepts clock via ADAT (I'll have to double check).

The absence of word clock I/O on the A2D is just as bad as the absence of is almost as bad of an idea as the absence of wordclock I/O on the 002R. Damn. I just have to stop buying such awfully designed gear!

On the other hand, I expect such numbskulled design from PTLE, but not from API!

[ryanfrith]

Originally Posted by FossilTooth

Wow. This kind of sucks. I feel like I've been duped.

I like API's on overheads. I guess this all means I can't reliably use the A2D on overheads and use my rosetta 800 for the rest of the drum channels with my 002R setup...

... or I could, but then I wouldn't really be able to take advantage of the extra 2 channels of A/D that I paid good money for.

I will have to second the notion that this was a really bad design call.

It doesn't really consider the needs of the end user. By definition, that's poor engineering!

I guess I could set the accept clock from the A2D, and then feed that clock to the rosetta via ADAT... if the rosetta accepts clock via ADAT (I'll have to double check).

The absence of word clock I/O on the A2D is just as bad as the absence of is almost as bad of an idea as the absence of wordclock I/O on the 002R. Damn. I just have to stop buying such awfully designed gear!

On the other hand, I expect such numbskulled design from PTLE, but not from API!

If you decide you need to return your unit, and are past that point w/ whomever you purchased, or you have to pay some restocking fee, then seriously, contact API and let them know of your experience. They are a 100% satisfaction guaranteed company. I came to a happy resolution w/ them, but I decided to keep the unit for other compelling reasons.

[Kiwiburger]

I don't know how anyone could feel duped about this issue. It's the most discussed aspect of the A2D - hard to miss if you do the barest minimum of research.

Can the Rosetta not sync to the A2D via AES? Everyone says the Rosetta sounds better with an external clock anyway ...

[ryanfrith]

Originally Posted by Kiwiburger

I don't know how anyone could feel duped about this issue. It's the most discussed aspect of the A2D - hard to miss if you do the barest minimum of research.

Can the Rosetta not sync to the A2D via AES? Everyone says the Rosetta sounds better with an external clock anyway ...

In fairness, it was a little misleading if you read this quote from API:

API A2D

Question from "swan":

"forgive the noobness of this question...but i have a fireface and a rosetta200 synced with BNC wordclock. The A2D specifies that it accepts Superclock. I tried researching this, but i still do not know if the A2D will be able to slave to the Rosetta through normal BNC wordclock...If it can't, then i can't get the unit (please say it can)."

Answer from "API Sez..."

"The choice of the word Superclock was probably not the correct one in this case. It was not intended to refer to any manufacturer's specific device or technique, but rather to the use of a stable external clock reference that produces the standardized clock frequencies. (kind of like saying Kleenex) Any or all of the available external clocks should work just fine with the A2D - in fact WE use the Rosetta 800 here to check all the outputs of all A2D's as they go through test."

Seems like they're saying Rosetta will work fine and that any or all available external clocks should work just fine, which just isn't true. I felt the same way after my purchase..."duped"; although I've decided the keep the unit since it has so many great features.

[magibatalla]

I just purchased Prism Sound Orpheus. Long ago I planned to get an A2D in order to expand my AD channels up to 10. Reading this clock issues with the API really makes me worry. I don't think A2D clock is superior to that in the Orpheus, so I don't like the idea of having to slave the Orpheus, A2D being master...

Is API solving this issue soon? Is there a workaround? Should I consider other options (Neve DPD is just too expensive for my budget)? Any other high end 2 channel preamp with good conversion? Maybe an API 3124+ with external converison with ADAT out (Aphex 142)?

Budgetwise the A2D would be my preference, but I want to take the most out of my setup, leaving clocking duties to the Prism...

[davetron5000]

I considered buying one a while back, and after finding unclear info on what "Superclock" was and determining that, to sync my 002R, Digimax LT and the a2D would be difficult, if not impossible without additional gear, I decided not to buy it.

I was just not able to understand the decision to omit a standard word clock. Maybe I'm wrong, but I felt the A2D was targeted to a user exactly like me: PT LE user who wants a high-quality A/D converter and a high-quality preamp in one package; A "prosumer" for lack of a better word.

I don't know what the additional cost of putting a standard word clock sync option would've been, but it made the difference to me and I didn't buy it. I'd be willing to bet a lot of "prosumer" users would pick this device up if it were more compatible. And I seriously doubt the API name would be tarnished by using an "inferior" clocking method.

[Roc Mixwell]

Its not inferior at all,

You just need to find a way to distribute its clock source correctly to other digital gear.

There are ways, and I KNOW you guys don't want to hear this, but a Big Ben would solve this problem because you can use it as a SAMPLE CLOCK distributor, when patching the AES signal from the A2D into the Big Ben (setting to receive AES clock and then passing sample clock on BNC (or AES) into the next device to be synced.

[Unknown soldier]

Originally Posted by Roc Mixwell

Its not inferior at all,

You just need to find a way to distribute its clock source correctly to other digital gear.

There are ways, and I KNOW you guys don't want to hear this, but a Big Ben would solve this problem because you can use it as a SAMPLE CLOCK distributor, when patching the AES signal from the A2D into the Big Ben (setting to receive AES clock and then passing sample clock on BNC (or AES) into the next device to be synced.

In this case, the A2D would be the master clock and you could then distribute to spdif/AES through the Big ben. Would this yield any improvement in sound? Would the big ben "clean up" the API clock (supposedly it's pretty good already)? Seems like an expensive workaround - for the cost of an A2D and big ben you might as well get a high end AD converter....

[magibatalla]

Originally Posted by Unknown soldier

Seems like an expensive workaround - for the cost of an A2D and big ben you might as well get a high end AD converter....

Exactly... It would be much easier if API just fixed this. I would like to know wether we can expect this to be solved, and when.

I think I could just slave the Orpheus to the A2D, I don't use any other digital device so it'd be one way trip only, no need to distribute the clock signal. Anyway, I should compare the results of using one clock source or another, I paid a lot of money for the Orpheus and wouldn't like anything to degrade its