Nice!

[Dr ZEE]

... history on any topic is the most important ....

Nah... I would say - fundamental science is.
...or is it too old school nowdays?

... hypothetical ...

... nothing "hypothetical" ...

Sure it's not hypothetical.
...., that's why nobody even cares to know what DAW/computer system Steve has.



***********

... Well, that's not really the point I was trying to make...

uh, I didn't realise that you were trying to make a point

.. it is a GREAT audio slave...

Maybe so, just like any GREAT digital sampler.
Based on that manual excerpt, no full chase lock in Cubase-4. But then again, it's a manual, so you never really know

 

[miroslav]

Sure it's not hypothetical.
...., that's why nobody even cares to know what DAW/computer system Steve has.

Steve left after the first page.
But it's not about *his* computer/DAW...it's about the process in general.

 

[Beck]

Refer to the attachment below for what my Cubase Studio 4 manual states specific to synchronization and how it handles different types of devices and protocols, and then proceed.




I'll be interested to read what, if anything, Tim provides in response. I'm not sure if all of what is stated in the above 2-page excerpt agrees with what Tim has been saying and vice-versa, but I hope it does.

The thing that jumped out to me is how Steinberg decided to handle the integrity of the synchronized system components, juxtaposed with the audio, and that is presented on the second page. They set the priority on the audio in a sense and basically, if I'm reading it correctly, the system components remain locked but the audio freewheels. I think it is safe to assume that, generally speaking, the audio sync issues will come out in the wash. In other words, over the course of a reel of tape, the sync between the analog audio and the digital audio will shift this way and that (secondary to the innate fluctuations of the mechanical transport and the digital audio "freewheeling"), but because the system remains in sync (the timecode reference coming off tape as master and the digital clock reference of the DAW chasing), there shouldn't be overall drift from start to finish between the analog and digital audio. The digital clock reference will be able to stay in tight sync with the analog transport (depending on the quality/make/model of the transport), and depending on HOW timecode was striped to the tape (ideally in some fashion that is referenced to the digital clock during striping, which I believe Tim spoke of earlier), things should be suitably consistent between from start to finish, and any momentary fluctuations resulting from the mechanical transport's normal anomalies will be give and take throughout and we should end up satisfactorily spot-on at any given point in the timeline even though the *audio* isn't "locked". Its pseudo-locked because (if we did it right), the timecode reference striped to the tape was generated from a device referenced to the same clock reference as the digital workstation at the time of striping, AND during recording/reproducing.

Don't know if I have this right, and Tim I'd love you to pick it apart if I don't (if you can stomach having to continue corresponding with us "children" ) just so I can thoroughly understand.

The invitation to "pick it apart" is sincere and genuine and expressed with respect, Tim. The bit about "children" is a bit of a poke...no, its just a plain poke. Tim you've made this kind of statement in the past, and I (correctly or incorrectly) sense frustration and sometimes I have to wonder why you continue to hang out with the "children" if it is that much of a consternation. You've been a tremendous resource here, and other venues I'm sure, but the "I'm blessing you little people with my presence" tone, which is how it comes across to me, is a poor representation of the kind of person of you that *does* provide a great deal of resourceful and helpful information. I just think its unfortunate.

Anyway, if I've got it reasonably right with how Cubase operates as a slave to a mechanical transport, it would work quite well with the exception of certain material. For example, if this system arrangement had been used during mixdown of a combination analog and digital tracks on something like Steve Reich's Different Trains with Pat Metheny, I think there may be audible problems...phase distortion because of some of the tracks' rapid and repetitious layered parts. So that is one situation I can think of where the mechanical transport as master could pose a problem. The easy solution in that case would be to set the digital "transport" (or rather the digital transport's clock reference) as master. BUT...there would have to be some pretty weird circumstances, or rather poor planning on the part of engineering or production, to cause the project to be at mixdown with a hybrid of digital and analog tracks. Realistically this wouldn't happen and naturally ISN'T how it happened. I'm just citing an example where IF that was the case, ATR as master could be bad if the system configuration and behavior was like it is (in my present understanding) with Cubase 4.

It would be interesting for me to scope and compare the output of a tone on my MRL test tape from my BR-20T when it is locked to its internal reference vs under control of the MicroLynx. I should be able to see differences in frequency and amplitude (secondary to speed fluctuations) and more importantly artifacts of cogging if present.

The explanation in the Cubase manual is a good one generally speaking... but the "problems" one anticipates are still open to interpretation. That is, a reader may look at the explanation and see some potential hazards that aren't cautioned against or implied in the explanation. Basically one can still jump to conclusions based on preconceived notions. One important point the Cubase manual addresses better than some is that SMPTE/MTC interacts with a DAW differently than a word-clock link in that the former does not control the converter clock... something I've been saying in this thread and many others before this. All one has to do is rig a frequency counter to the clock output of the digital audio interface under control to see that MTC does not cause any variation at that level. So I hope we've settled that here at least. Another good thing the Cubase manual does is separate the sections on digital sync and time sync... because they are different concerns and like comparing apples and oranges. It's a bad idea for programers to organize the different external control options in the same menu when designing the GUI.

But how is the information to be put into practice? Does it mean word-clock is more accurate than a MIDI level control? Yes! Does it mean word-clock is better than MTC for synchronizing devices in wall-clock-time for recording? No! Nor does it imply that it's better to slave the analog deck in a hybrid system.

As for your comment on syncing different formats at the mixdown stage (if I understand correctly) being poor planning. I don't agree with that because that is actually textbook how its done in a hybrid system. I do it that way... analog deck as master, two Echo Layla 24's and outboard MIDI sound modules/drum machines all running in perfect sync while mixing down to half-track analog. I have 44 main mixer channels available for this mixdown phase plus some additional submixers if needed. Conceptually its no different than synchronizing two analog multitracks to increase total track count. There will be some phase issues no matter what you use as master... but only when you try to record a stereo source with one track on one machine and one track on the other. There are procedures to follow when syncing any two devices. No two machines ever fully become one machine through syncing... and again that halds true whether the analog deck is master or slave in a hybrid system. The best you can do outside of the hybrid model would be two or more identical Audio interfaces, such as two Echo Laylas like I'm using, with one as master of course. Because when you're using different brands and models with different converters it opens up a whole 'nother can of worms. A third party master clock can do more harm. Here's an article is SOS you may find interesting that touches on that topic:

http://www.soundonsound.com/sos/jun10/articles/masterclocks.htm

 

[sweetbeats]

As for your comment on syncing different formats at the mixdown stage (if I understand correctly) being poor planning. I don't agree with that because that is actually textbook how its done in a hybrid system.

Right. I get that. My reference to "poor planning" was only specific to say Steve Reich's "Different Trains". I don't know if you are familiar with it, but I was just using that as an isolated example of what might be an unhealthy scenario if they had tracked some of Pat Metheny's layered/duplicated guitar parts on a digital non-mechanical transport-based medium (that freewheels the digital audio like my Cubase 4) and some on an analog mechanical transport-based medium and then had to use those two mediums in mixdown with the analog transport as master...I was thinking (note: not KNOWING...I'm conjecturing here as a matter of probing to hone my understanding) this scenario and mixdwon arrangement may reveal audible phase issues more prominently than as is typical *because* of the type of material. I haven't listened to this collection of work in years but I recall one piece in particular that has (this is from memory) long phrases of maybe 12-16 duplicate 16th note staccato guitar tracks at maybe 72bpm...maybe even more tracks. I'd have to bug my brother to borrow his CD and read the liner notes again. Anyway, I was thinking what if half those tracks were digital and half were analog and now we're pairing an analog transport as master and a chase-locked digital "transport" with freewheeling audio...and thinking there would be phase issues. But...duh...no...

There *WOULD* be if all the tracks were truly identical duplicates, or if one SET of tracks on one medium were tracked at the same time as the other (i.e. bussed to both mediums simultaneously with each take), but if all the tracks were individual takes (which they were) there wouldn't be audible phase distortion.

So forget it...just came to that conclusion here as I'm typing.

Hm.

I can't think of a single scenario where, so long as the tape on the analog mechanical transport was striped by a code generator that references the same clock reference as the DAW system, there would be a problem with the analog transport as master and a DAW software package that "freewheels" the audio while chase-locking its clock to the master clock.

NOTE...there is a caviat to this with old transports like Steve's Ampex AG-440, or my Ampex MM-1000, because these are good old resistor-based fixed tension transports...down-home "grab the tape and drag it on through", and what happens is often the tape speed varies from start to finish of a reel of tape because as the tape unloads from the supply reel the holdback tension has greater torque force potential as the tape spool decreases in diameter. This is relatively of no concern as long as the tape recorded on the machine is reproduced on the same machine, but take it to a different machine or a constant-tension machine of the same track format and watch out...timecode nightmare.

I'd also like to say I'm more comfortable now with the idea of the analog mechanical transport as master, that is with this understanding that my Cubase software is locked but leaving the audio freewheeling. My audio isn't being messed with, but what of software that DOES lock the audio? Then would there not have to be samples being dropped or added as the master transport fluctuates?

And I'd still like to put a scope on the output of an MRL tape tone being reproduced while my BR-20T is clocked to its internal servo reference vs under control of an external reference. That would be a good education for me I think.

 

[miroslav]

I can't think of a single scenario where, so long as the tape on the analog mechanical transport was striped by a code generator that references the same clock reference as the DAW system, there would be a problem with the analog transport as master and a DAW software package that "freewheels" the audio while chase-locking its clock to the master clock.

NOTE...there is a caviat to this with old transports......

Exactly.

If you have a master sync reference that was used during the striping process...it's a different situation, and sync is more reliable...but if you don't have one, or choose not to use one, or your deck can't be referenced/controlled by external devices...then you lose the reliable sync.

I'd also like to say I'm more comfortable now with the idea of the analog mechanical transport as master, that is with this understanding that my Cubase software is locked but leaving the audio freewheeling.

Well that's the tradeoff....having good sync via chase-lock, or letting Cubase freewheel the audio, but then you have the potentianl for drift if/when doing multiple dumps from a deck to Cubase.
For a single pass dump...the audio freewheeling is not an issue.

From the Cubase Manual (which I'm sure you've read):

Synchronizing Cubase’s playback
Let’s assume now that we use external timecode synchronization with Cubase. For example, we might synchronize playback to a tape recorder.
Timecode coming from an analog tape recorder will always vary slightly in speed. Different timecode generators and different tape recorders will also supply timecode with slight differences in speed. In addition, the shuttling of tape mechanisms due to overdubs and re-recordings can cause the physical tape to wear and stretch, which affects the speed of the timecode.
If you use a synchronizer that generates word clock and set up Cubase to sync to incoming timecode, it will vary its overall playback speed to compensate for such fluctuations in the speed of the timecode – that’s the whole purpose of synchronization.

What happens with the digital audio?
The fact that Cubase’s playback is synchronized to the timecode does not affect the playback of the digital audio. It still relies on the perfectly stable, built-in clock in the audio hardware. As might be expected, problems will appear when the perfectly stable digital audio gets related to the slightly varying speed of a system synchronized to timecode.
The playback timing of each event will not be in total accordance with the tape or the MIDI playback, since the playback speed of the audio is determined by the digital audio hardware’s built-in clock.

 

[Dr ZEE]

......there is a caviat to this with old transports... ..
watch out...
.

Watch this Good Old Master at work as MASTER and try to think of what is the essence of synchronization:



when flute blows too soon or too late, don't blame the MASTER.

 

[Beck]

Right. I get that. My reference to "poor planning" was only specific to say Steve Reich's "Different Trains". I don't know if you are familiar with it, but I was just using that as an isolated example of what might be an unhealthy scenario if they had tracked some of Pat Metheny's layered/duplicated guitar parts on a digital non-mechanical transport-based medium (that freewheels the digital audio like my Cubase 4) and some on an analog mechanical transport-based medium and then had to use those two mediums in mixdown with the analog transport as master...I was thinking (note: not KNOWING...I'm conjecturing here as a matter of probing to hone my understanding) this scenario and mixdwon arrangement may reveal audible phase issues more prominently than as is typical *because* of the type of material.

Ah... ok, I see what you were saying.

One thing that still seems difficult for some to grasp is that although the DAC clock is not controlled by an MTC stream the audio still does lock. When you see references to freewheeling or flywheeling its no different than how that works with two analog decks synced together, but a DAW is compensating for speed at the software level after A/D conversion. "Freewheeling" as some are interpreting it would not be sync at all. MTC most certainly keeps the slave in sync with the master... that's what its for.

How I work is to record everything to tape first and then dump those seven tracks at a time to my DAW. (track 8 has the SMPTE stripe of course.) I'll do that up to two times if I'm using both Layla's. Then I'll fill up the tracks on my analog deck one last time, but I won't dump that to the DAW. For my purposes recording to analog before transferring those tracks to digital is better than straight to digital, but I still want certain analog tracks to go straight to analog mixdown without being converted. Anyway, to make a long story short, that's why I'm still running the analog deck with the DAW at mixdown. Even after two analog to digital dumps and then syncing seven more tracks to what I've dumped to DAW the system is perfectly in sync during the final mixdown. There's no more drift then there would be if I were syncing to analog machines. That is, there's no audible phase issues unless like I said before, if I made the mistake of stretching a stereo source across two tracks on two different machines.

 

[miroslav]

If the DAW compensated for speed fluctuations AFTER the conversion took place...that WOULD certainly change the audio, and every minor tape fluctuation would cause audio glitches/pitch changes.

This is why a DAW re-samples on-the-fly when set to chase-lock, and the audio comes out the way it should, because the DAW is compensating for at as it happens...and not AFTER the conversion is done, which wouldn't make sense.
It can't resample to fluctuating speeds without involving it's digital clocks at some level and making adjusments.

Most people may not care about the resampling, and I think some DAWs just do it, without ever revealing or memntioning it in their manuals....but if you don't want constant resampling, the only way to avoid it is to run the DAW as master. That's a simple as it gets.

This is pure common sense stuff, and if anyone thinks it works differently, then provide proof/examples....otherwise, it's just "words".



Oh...I also think that looking at a DAW as *hardware converters* and *application software* as seperate tings, only confuses the issue.
While the actual converter clock may be set to a specifc sampling rate and not change at all during a synchronization setup....the DAW software is still able to re-sample and adjust its clock in chase-lock mode in order to keep in sync with fluctuating SMPTE from a tape deck.
Saying the "clock" in the converter doesn't change, is a moot point....it's how the total DAW system reacts to fluctuating tape SMPTE that is key, and not just the converter.


PS
And isn't it cool how there's finally a *digital* topic being discussed in the Anlog forum without it turning into an analog VS digital flame war or an anti-digital hate-fest.
.

 

[Beck]

This is why a DAW re-samples on-the-fly when set to chase-lock, and the audio comes out the way it should, because the DAW is compensating for at as it happens...and not AFTER the conversion is done, which wouldn't make sense.
It can't resample to fluctuating speeds without involving it's digital clocks at some level and making adjustments.

Well I will concede that perhaps your use of the term, "re-smaple" is throwing me off because I wouldn't call it that and don't believe it is technically correct as far as what's happening in the DAW at the MIDI level. Maybe we could explore that a bit more, as it seems to be a sticking point.

PS
And isn't it cool how there's finally a *digital* topic being discussed in the Anlog forum without it turning into an analog VS digital flame war or an anti-digital hate-fest.
.

Yep, well this is one area I find a DAW to be very useful and better at doing something than analog is. That doesn't happen often, so enjoy it!

 

[miroslav]

Well I will concede that perhaps your use of the term, "re-sample" is throwing me off because I wouldn't call it that and don't believe it is technically correct as far as what's happening in the DAW at the MIDI level. Maybe we could explore that a bit more, as it seems to be a sticking point.

I've been taking the basic explanations given by some DAWs at face-value. So when they say "re-sampling"....I'm understanding that they are talking about what happens to the incoming digital audio in a DAW-slave scenario as it attempts to chase-lock to a tape deck...
...but I also will concede that there's no clear, detailed step-by-step description of the specific interaction that occurs internally between the converter and the DAW application AFA the synchronization goes.
None of the DAW apps ever get into the real technical nitty-gritty stuff about their programming, you know, the stuff that most people take for granted and don't much care about (unlike you and I, and a few others around here )....
...not to mention that the internal sync implementation may not even be the same from DAW to DAW, which throws another variable into the mix....
...so by all means, it's something that could use more exploring.

I still say that in the end, even with the possible small sync imperfections and audio manipulation that might be happening....it's still good enough for Rock & Roll....but I always like to know what's really happening.
I don't think it's a bad thing knowing the small details.

 

[SteveM]

I got lost waaaaay back there. I will say that the process is "sketchy" with the Ampex and there is some timing issues now and then. So I've been doing a little of everything. The multitrack I use is a Roland VS1880. Not even sure if that is considered a DAW or not. Carry on!

 

[Dr ZEE]

...Roland VS1880

- is as DAW as it gets

I got lost ...

A "technical" discussion of a "process in general" in a "General DAW" is an adventure in a fictional jungle, a perfect place to get lost

**********
p.s.
a bite of food for thoughts for an audio purist-theorist

1. In essense , synchronization is an event alignment procedure.
2. Naturally, A recorded audio ( or continuous audio signal that is being recorded) embeds only two events: The Start and The End.
3. Only the Start of an uninterrupted audio signal can be synchronized to a given event in time without deforming the signal.
4. In fact, synchronizing recorded audio signal ultimately means deforming it.

 

[Dr ZEE]

-
4. In fact, synchronizing recorded audio signal ultimately means .....

In fact, strictly technically speaking, there's no such thing as "synchronizing audio", but there's such thing as synchronizing machines that record and playback audio.