Does this sound right to you? (avchd codec and vegas)

Rippie

Rippie

Veteran
Ive been editing an avchd project in vegas on a
duo core 3.0 ghz processor and 4 gigs of ram in xp.

For the most part until recently it would run the clips (1920x1080) at full frame rate
and at cut points drop a few frames for a few moments then go back up to 29.97.

These frame drops were driving me crazy never knowing how an edit exactly looked because it always slowed down so I decided to build a smokin new comp.


Just put it together a few days ago..

i7 quad core 2.66ghz (overclocked to 4.2)
6 gigs of 1600 ram
Evga x58 MB
evga 260 216core video
64 bit vista

Granted my overclocking might not be nearly perfect yet, but
its incredibly disapointing that in vegas the avchd files are STILL
barely staying at full frame rate.

I was optimistically hoping id be at full frame rate AND get very good frame rates
with magic bullet on. Not even close.

Does this sound right that this power hungry format is still struggling with
this pretty dang high end comp?

Or am i missing something?

Thanks

​
 
Magic Bullet is going to be a real-time killer and a render hog in any situation.

I'd watch the overclocking. I've never been a fan of it.
 
yea the magic bullet thing is just an extra bonus i was hoping for.

I just want to get the basic vegas/avchd thing working as well as im going to get.
 
Slowdowns at cut points will happen on even the speediest of machines.

AVCHD, like HDV, uses GOP encoding, which is a "Group of Pictures." One full reference frame is stored, and then for several frames, all it stores are the changes from the reference frame. Then it stores another full frame and starts another group.

What happens is that if you are lucky enough to cut on a full frame, Vegas can handle it easily. More likely, you are cutting on one of the frames that is just changes, which means Vegas also has to go back through several other changed frames to get to the reference frame, reading them all and figuring out what the frame you cut on should look like.

Precise technical explanation here:
http://en.wikipedia.org/wiki/Group_of_pictures

All of this work requires a lot of processor power, and also disk access. I have yet to see a machine fast enough to always do it smoothly, but one option is to transcode the footage to another (much larger file size) format and edit that, or transcode to standard DV for editing, and use VASST Gearshift to switch back to the original HD footage when you render out.

It's already better than it used to be, though... in the early days of HDV, you could only cut on the full frames, which meant it could end up cutting up to 8 frames away from where you wanted!
 
stragways thanks for the explaination.

This is very useful info, and I wasn't aware of it.

Wish I knew before i just dropped $2grand in computer parts ;)
 
You must log in or register to reply here.
Back
Top