That's funny, good to hear from you. I've temporarily given up on Nuendo Live because there appears to be no interest. Steinberg has improved their communication skills on this forum the past couple of years, most other genuine issues receive at least a "we're considering fixing that" or something similar. Silence on one of Nuendo Live's primary professional features - AAF export - indicates silence on the product as a whole. I appreciated Timo's initial response but it appears resources have not been allocated to Nuendo Live because recreation of the problem is easy. If they looked at it, determined it affects too few folks to add a selector for proper AAF export, then at least say so and allow us to write it off as not usable in our applications.
Since my last post we've gathered some other information further indicting Nuendo Live's timecode capabilities. The timestamp put in the file is different from reality based on three variables -
Audio buffer size
PreRecord amount (including zero amount)
LTC>MTC converter used
Re Create -
Record a signal chasing TC
Change buffer size (including any combination of restart)
The timestamps will be significantly different. Timestamps also vary if you change the Prerecord amount including zero. Note that the timestamps are never close to accurate - in our tests, based on the variables, they were off as much as 4 *frames* from reality.
The choice of synchronization device changes the timestamp much more than in Nuendo 5. In Nuendo 5 and Nuendo Live we've tested two different Motu models, Rosendahl, Lockstep, Smpte Reader, and RME TCO - all compared to ProTools HD running a SyncHD. We do a 10-pass test for each synchronizer determining accuracy, repeatability, and max/min/mean deviation. In Nuendo 5 the timestamps will vary slightly; the worst was perhaps 3ms off reality (project not in front of me at the moment). Max deviation differed between the units; some varied by as much as 1ms pass to pass, others varied by no more than a couple of samples.
In Nuendo Live however changing the synchronization device resulted in extreme changes in the timestamps; as much as 2 *frames* variation, repeatable, in addition to the already incorrect offset of the other two variables, buffer size and prerecord amount.
Summary of Nuendo Live's MTC timestamp capability - the timestamps are always incorrect by a few frames, and vary more than other applications.
One last problem that an associate has tested causing him to give up on Nuendo Live completely. He was having some issues with Nuendo Live and long recordings - sometimes the files would be shorter than other devices by up to 4 frames after a 3 hour recording. This was determined by importing Nuendo Live's files into another workstation along with files from the other recorders - a PTHD system, a JoeCo, and the wav files from the DDRs. The Nuendo Live files of course had incorrect timestamps, but if you lined up a hit at the top of the files, a hit three hours further may be earlier than the other files by up to 4 frames. This indicates Nuendo Live was dropping samples, and quite a few of them. Not predictable but happened occasionally using the same hardware as Nuendo 5, therefore not a hardware issue. Obviously all clocking variables have been eliminated (we're good at that).
It's unfortunate the apparent lack of interest on Steinberg's part to a product that folks pay money for. The TC synchronization is simply broken, it is not the same as Nuendo 5. As a beta tester for the product I'm disappointed to see it this way, even the graphics issues I dealt with during the testing have only been partly resolved - Nuendo 5 has significantly improved graphics performance.
Too much information, hope your Nuendo world is going well.