-
1 frame variation in capture from HDV…comments?
Hey again
Posted this once before, but i think it was confusing
using an HDV tape with visual time code recorded in, and capturing the same in and out points on the tape multiple times via firewire, i found a 1 frame variation.
I Made the tape from a timeline on fcp. the tape tc str arts at zero and runs to 30:00. its NDF. i then logged and in and out from the tape regardless of the window (i know the window would never match the tape so i did not bother to try) I then re-logged the same clip 4 more times, thereby making 5 separate clips with identical time code. i then batch captured all 5. i repeated this several times.
3 out of 5 captures showed the first frame of visual tc to be 8:15, 2 of the captures showed it to be 8:14. therefore i can ssume that the interpretation of time code from the mpeg2 stream on the hdv tape is subject to a 1 frame error
is this what is to be expected? is capturing from HDV not frame accurate? are there not soem kind of agorithm to keep it accurate, like the drop frame algorithm
if so i will modify my workflow accordingly, i just want to know what others have found, or any comments you may have. i invite discussion
SONY HVR m10U
FCP 5.0.4
OS 10.3.4
dual 2.7 g5bee eph
Bob Flood
Greer & Associates, Inc. -
2 Replies
-
Bob,
Keep in mind, with HDV there is no such thing as an actual “frame.” Because of its 15 frame GOP structure, HDV is not (at this time at least) 100% frame accurate. HDV records metadata that is translated into 15 frames of video, and each time this translation occurs there are minor variations. I believe I read in the technical specs that HDV has average accuracy of plus or minus one to three frames.
DRW
-
DRW
hey guy!
OK! i thought something like that was the case as, like you say, HDV is not really a series of frames as its a data stream.
The reason for my testing is that i have read some posts, and i think walter talks about it as well, that people are capturing and then re capturing
vis a vis the ol edit workflow (lots at lo rez then final at hi rez)I cant see how that would work based on this test. the current mind set would be to capture all yer gonna use at full rez , and then make lo rez proxies to edit with then relinking when all is said and done the downside of that is if you are busy like we get sometimes and you only have one system soemone has to stay late! (we gotr spoiled by having 2 edits!)
bee eph
Bob Flood
Greer & Associates, Inc.
Reply to this Discussion! Login or Sign Up