Is TDM Dying?

[EarHole]

I just upgraded to an Accel2 System and I have t say that I'm dissapointed.
It just seems that all the new plug ins are RTAS only and my DSP chips are barely getting
used. I have the waved bundle and plenty of TDM plug ins but it seems like somewhere in the
I pull up something that is RTAS only and I have to switch them all. I know you are supposed to be able to mix and match now but I get inconsistant results with that.

What's the deal? Are we moving to a more host based system despite it's shorcomings because
third party developers find it easier to develop for the RTAS path?

[jeff markham]

could you be more specific about which plugs? i think you said "waves", but
i'm not sure that's what your talking about.

if you are talking about Waves .. i assume you're using the TDM versions (which
cost about 2x the native versions).

also, there are order issues in the insert path about mixing and matching
TDM and RTAS plugs .. so maybe you could give an example of how you're doing
the inserts and then it might be a little more clear.

jeff

[Craig F]

from what I hear it is easier to code for RTAS than TDM and they only need to do it once
but start using ADC in 5.1 mixing and watch the chips fill up

[EarHole]

Hi thanks for the replies...

Well for example I was mixing a spot where the dialog track needed a little help with matching from
another location.I put a basic subtle Ren Comp TDM (more of just a gain control)and then the Firium EQ RTAS.The track would randomly not play any audio. I would have to disable and then reenable the
out put to get it to play again. It just seems that as I'm mixing sometimes I'll desire
the use of some insert that is only RTAS. Then I either have to go switch all the previous ones over
or deal with unexplained and random glitches. Yesterday I even had a case where the scrub didn't include the proccessing.

It just seems that all my Tried and True plugs are TDM, but more often lately I'm reaching for something that is newer and RTAS only.

I haven't done a 5.1 mix yet. That was the reason that I upgraded from my mix plus system however. I did a decent sized project in 5.1 and my system was crawling so I made the jump.

[Phil Jeffers]

could you be more specific about which plugs? i think you said "waves", but
i'm not sure that's what your talking about.

if you are talking about Waves .. i assume you're using the TDM versions (which
cost about 2x the native versions).

also, there are order issues in the insert path about mixing and matching
TDM and RTAS plugs .. so maybe you could give an example of how you're doing
the inserts and then it might be a little more clear.

The Native versions of Waves plug-ins are exactly the same as the TDM versions (that is, those that are available i.e. Renaissance Reverb, L2 etc.) the only difference is where they get their processing power from.

[airon]

I'm an avid(NO pun intended) scrubber and haven't tried RTAS plugins yet, because I'm stuck on a 5.1.3 TDM rig on an old G4(happy for what I do).

When scrubbing on a 6.x or 7.x system with a RTAS plugins, does it feel the same way ? Any glitches apart from the fumbled output you mentioned ?

[medmondson]

It's inevitable that TDM is, or soon will be dying, IMHO, in the sense that its Mixer and Inserts are architecturally limited to realtime processing. There was a time, of course, where this was its greatest strength (real time EQ? OMG!) But the world is going tapeless, fast. I work for a large television network, and we've recently initiated the first major phase in eliminating digital tape from our workflow. And we're not as far ahead of the curve as others.

Realtime laybacks are still, for now, acceptable and billable costs. But how long will it be before we start having clients, who mix down the street at a smaller Post House, telling us that so-and-so was able to output their 2 hour show to a WAV file, and have it posted to a server in 45 minutes? I'm guessing 2 years minimum, maybe 5 max. In that time it would serve Digidesign well to develop a solid means of outputting multichannel audio at >real speeds. Or at least acknowledge it.

Don't get me wrong- I cringe at the thought of using Logic or Nuendo, personally, for any type of high volume, intensive, deadline-oriented, professional workflow. And I don't expect Pro Tools to be on the bleeding edge like these other apps, where, if something doesn't quite work, it really doesn't affect someone's million+ dollar bottom line. I get that reliability takes time to develop. I get that single-user music production houses are willing to sacrifice some degree of reliability in the name of having the latest and coolest tools, and that Pro Tools can't (and shouldn't try to) compete at that level.

But... the tapeless world is coming, and TDM's realtime bottleneck appears just a tiny bit more restrictive each day.

[Jonathan Wales]

Well my basic answer is "of course not".

The thing is that TDM systems are still RTAS capable (which means you can have your cake and eat it if you want).

But mostly you must acknowledge that the simple strength of TDM is the very fact that the host processor is NOT needed for the basic audio processing tasks and thus is freed up for things like automation.

I am using HD6 systems and maxing those out, and there is now way at all that you could ever get close to the amount of mixing / EQ / dynamics / harmonizing / delay / reverbs (commonly 10 revibes on any given system at once).............. how much longer do I need to go.


Now to the ridiculous idea of non-realtime laybacks.

Which part am I missing. The point would be that in order for the material to be accepted to be aired, it would have to be played back for QC. No-one will air something that has never even been watched at the layback..... unless they are simply crazy.

Am I also missing something else. Isn't the picture on Digibeta, or D5 or HDCAM. How the hell do you get them to go faster than realtime??????


For feature work, Dolby insists on real time laybacks - at least in part because any screwup in an optical track becomes really expensive later on down the line.

The only laybacks that are currently being delivered are COPIES of mixes being put onto things like DVD-R or MO disks for delivery purposes, but those are COPIES of something that has already been done in real time and thus QC'd.

Yes, TV stations are doing playout from non-realtime media, and yes they might take fast delivery in certain specialized applications like live sports, and news and stuff, but do you reallly think that anyone will risk airing something longform (i.e. primary programming) that has never even been seen.

I don't think so for quite some time.

Oh, and by the way it would be better for the industry if we do keep those laybacks as billable don't you think..........

[philper]

Well my basic answer is "of course not".

The thing is that TDM systems are still RTAS capable (which means you can have your cake and eat it if you want).

But mostly you must acknowledge that the simple strength of TDM is the very fact that the host processor is NOT needed for the basic audio processing tasks and thus is freed up for things like automation.

I am using HD6 systems and maxing those out, and there is now way at all that you could ever get close to the amount of mixing / EQ / dynamics / harmonizing / delay / reverbs (commonly 10 revibes on any given system at once).............. how much longer do I need to go.


Now to the ridiculous idea of non-realtime laybacks.

Which part am I missing. The point would be that in order for the material to be accepted to be aired, it would have to be played back for QC. No-one will air something that has never even been watched at the layback..... unless they are simply crazy.

Am I also missing something else. Isn't the picture on Digibeta, or D5 or HDCAM. How the hell do you get them to go faster than realtime??????


For feature work, Dolby insists on real time laybacks - at least in part because any screwup in an optical track becomes really expensive later on down the line.

The only laybacks that are currently being delivered are COPIES of mixes being put onto things like DVD-R or MO disks for delivery purposes, but those are COPIES of something that has already been done in real time and thus QC'd.

Yes, TV stations are doing playout from non-realtime media, and yes they might take fast delivery in certain specialized applications like live sports, and news and stuff, but do you reallly think that anyone will risk airing something longform (i.e. primary programming) that has never even been seen.

I don't think so for quite some time.

Oh, and by the way it would be better for the industry if we do keep those laybacks as billable don't you think..........

The greatest value of TDM seems to me to be the guarantee of a certain track and plugin count no matter what the computer is doing. That is a powerful incentive in professional situations.

The non-real time layoff idea is not ridiculous. One does this only after carefully playing down the mix and being sure that all the "i"s are dotted, etc.. and all the plugs etc are functioning properly. I have done innumerable NRT dubs/bounces on systems other than ProTools with great success virtually all the time. It seems to me that once layback or layoff to VTRs and DA88 tapes are eliminated, and deliverables are entirely audio files, then NRT is going to be common and expected.

Is it really technically impossible to make TDM work in NRT? TDM III?

I don't like the idea of losing a billable line-item either, but I'm also sick of repairing my DA88s.

Philip Perkins

[Frank Kruse]

The non-real time layoff idea is not ridiculous. One does this only after carefully playing down the mix and being sure that all the "i"s are dotted, etc.. and all the plugs etc are functioning properly.
Philip Perkins

Don´t you print to a master-track WHILE mixing? When you are through the mix your track is done. Why would you need a non-realtime bounce?

frank.