[Paul Dickin]

HI
An ex-Avid colleague and I edited two archive footage shows using FCP. My project for the two-hour timeline came to about 10MB.
His came to 194MB before the Mac gave up and failed to complete it!!!

What he was doing was creating a timeline of cut source footage, about 20-40mins, then opening a new project and dragging that sequence into the new project, and closing the first project.
So his new project had no clips in the Browser, just a sequence.
After he had done this about 15 times he had created his 194MB masterpiece, with no source clips anywhere in sight in his final project – just sequences that referenced earlier unopened projects. So in effect every project he had ever worked in had to be invisibly opened by his final project for it to know anything about all his nested/nested/nested… timeline clips.

I don’t know if any of this applies in your case, but I finished the show by creating a new project using the timecode data of his clips – which was down to 10MB again 😉

(I hadn’t realised that Avid clip management was so different to FCP’s.)

[Paul Dickin]

Hi
This post about altivec optimisation is interesting 😉
http://www.simdtech.org/altivec/archive/msg?list_name=altivec&monthdir=200506&msg=msg00037.html

[Paul Dickin]

Hi
If tone is too savage I use the razor tool, and either replace the section with some matching background noise, or just reverse the razored section, which is usually sufficient a disguise.

[Paul Dickin]

Hi
A lot of support for SCSI seemed to vanish at the announcement of OS X 10.3.3 – 10.3.2 was better. Otherwise it might have to be Jaguar 10.2+, which used to run OK on my dual 450 with an Apple-supplied Adaptec U2W SCSI card.

[Paul Dickin]

Hi
When G4s came along and replaced G3 Macs, Altivec-aware software got a speed boost of (very roughly) at least double speed.

So if you take away Altivec in the move to Intel CPUs you get 200% CPU boost with 50% application efficiency…
Result, not much 😉

Now I’m not the sort of person to know whether SSE3 or whatever succeeds it can be made to replace Altivec. Nor whether its a trivial recode or a massive one.

But my guess is that Apple will get out an Intel-optimised version of iTunes that runs OK a LONG time before it gets FCP up to a speed-boost much beyond where the Quad-core G5 is already taking it.

Like June 2007… 🙁

[Paul Dickin]

Hi
This software extracts the audio( and much more):
http://www.apple.com/downloads/macosx/video/mpegstreamclip.html

[Paul Dickin]

Hi
Most likely there was a timecode or video glitch in the captured footage – something that may be invisible to the eye, but which throws FCP…
Try bootting from the OS X install disk 1 with the C key held down at startup. Then ignore the installer screen, but go up to the menus and find Disk Utility, and run Repair Permissions on the start-up drive. (Do it again if there are any amount of repairs to do).
Then try rebooting as usual…

[Paul Dickin]

HI
I bought one with the stock graphics card a couple of weeks ago (ex-demo from big store with 1 year warranty) for $1340 (or rather

[Paul Dickin]

Hi
I had this when trying to run a corrupted captured file.
QuickTime Player couldn’t run the file off the hard drive either.

However running Repair Disk in Disk Utility repaired that bad file, and FCP opened it fine after that.

Probably Disk Warrior would be an even better utility to run to repair your drive’s files.

[Paul Dickin]

[Matte] “The tracks on the tape can nearly OVERLAP….”

Hi
That’s how it works in SP mode, the tracks are adjacent (with DVCAM there is a significant gap).
With LP mode the written data tracks overlap, so each edge of the recorded tracks has the previous or next track on the same area of tape.
With a digital recording the replayed digits are either readable or they’re not, so the theory of LP goes that 25% of the track is garbage, then 50% of the track is a good recording, then the other 25% on the other side is garbage again.
But the 2 garbage areas are randomly different, so the 25%+25% of garbage that passes under the read head just doesn’t quite match (by a teensy bit) the 50% of good recording down the centre of the track, and the output of the read head is theoretically a few percent more readable digits than the noise.

Sorry this isn’t a scientifically rigorous (nor precisely accurate) description, but it makes it more clear why SP is not a mode to ever ever use…