|
|
Could this be what's causing the stutter?
There is something just don't make sense to me. Something just don't add up.
Is it supposed to be this way???
Take a look below: a general example of DVD-RB for encode/trim/rebuild of cells.
(CELL:1) on the AVS file, it trims only 3615 frames instead of 3616 ?
Does 0 count as one frame? only if you add 0 as one frame to 3615 then = 3616 ?
But if 0 = one frame, then wouldn't it be 3617 by adding 0 to 3616 to encode on the REBUILDER.ECL?
(CELL:2) it trims from frame 3616?? doesn't it belong to (CELL:1) for encoding??
If frame 3616 doesn't count as part of (CELL:2), then there is only 4331 frames trimmed in (CELL:2) instead of 4332 ?
On the REBUILDER.INF Sectors overlap?
(CELL:1) First_Sector=0 Last_Sector=41846
(CELL:2) First_Sector=41846 Last_Sector=91674
The first sector on (CELL:2) shouldn't it be 41847 instead of 41846 ?
***************************
***************************
(CELL:1)
from: REBUILDER.ECL
frame_first=0
frame_last=3616
encode_first=0
encode_last=3616
--------------------
from: AVS
trim(0,3615)
--------------------
from: REBUILDER.INF
First_Sector=0
Last_Sector=41846
***************************
***************************
(CELL:2)
from: REBUILDER.ECL
frame_first=0
frame_last=4332
encode_first=0
encode_last=4332
---------------------
from: AVS
trim(3616,7947)
---------------------
from: REBUILDER.INF
First_Sector=41846
Last_Sector=91674
***************************
***************************
If it's supposed to be this way then never mind.
0 counts as a frame. So 0 - 99 would be 100 frames.
Also the numbers you see in the table are correct. The quot;lastquot; sector is actually the first sector of next cell. It's done that way purposely for internal processing.
Just a note all: All of the entries in REBUILDER.INF are there for DVD-RB internal use. Please do not modify them unless you are very, very familiar with what they do and how they work.
Are you actually experiencing any stutter? Because it is so rare as to be pretty much non-existent.
Yeah, I'm stuck with the movie Around The World In 80 Days (Jackie Chan). I've tried all versions even 0.75 but still get the same stutter same places.
This time instead of encoding the whole movie, I run the three cells where one of the stutters occur and found out a few things.
CELL:1(original 83MB) + CELL:2(original 99MB) + CELL:3(original 1MB) = stutter at CELL:1/joint/CELL:2
CELL:1(trimmed 26MB) + CELL:2(trimmed 27MB) = OK, (seamless same joint no stutter) only trimmed cells to smaller size.
---------------------------------------------------------------------------------------
Ok that's one thing, here is another from original size: take a look at (CELL:3) 74 frames???
(CELL:1)V01000000001003
from: REBUILDER.ECL
frame_first=0
frame_last=3616
encode_first=0
encode_last=3616 ---------gt; actual encoding from QuEcn and CCE = 3616 frames
(CELL:2)V01000100001004
from: REBUILDER.ECL
frame_first=0
frame_last=4332
encode_first=0
encode_last=4332 ---------gt; actual encoding from QuEcn and CCE = 4332 frames
(CELL:3)V01000200001005
from: REBUILDER.ECL
frame_first=0
frame_last=76
encode_first=0
encode_last=76 -----------gt; actual encoding from QuEcn and CCE = 74 frames ???????????
(CELL:1) + (CELL:2) + (CELL:3) = stutter at CELL:1/joint/CELL:2
--------------------------------------------------------------------------------------
I know I shouldn't be changing anything, but I guess it wont hurt from trying.
So, here is what I change:
Took one frame off or out at the stutter joint of cell 1 and 2
(CELL:1)from: REBUILDER.ECL changed both (frame_last, encode_last) from 3616 to 3615
(CELL:2)from: REBUILDER.ECL changed both (frame_first, encode_first) from 0 to 1
(CELL:3)no stutter here so didn't make any change
Final result: (CELL:1) + (CELL:2) + (CELL:3) = OK, a perfect seamless joint (no stutter)
Seamless joint hmmm... does that mean the two frames that I took out be duplicates?
Those are all correct. That's the way CCE notates them. I'll take a look at the code concerning the last one... how did you check to see how many were actually encoded?
Is this the NTSC version of this movie? If so I'll go buy a copy (yep... that's where all the donations go).
I took some screen capture images when CCE was actually encoding the cells, and it tells you the total number of frames for encoding.
I wish I know how to paste the images here so that you can have a look, but I don't know how.
And yes, the movie is NTSC version.
you should have tried the 2 removals seperate aswell. so 1 encode with just the first cell reduced by 1 and then another encoder with the first frame removed and see what happens then. do either of those stutter ?
I just got back from Best Buy.... with a copy of quot;Around the World in 80 Daysquot; NTSC, R1 -- I'll run it tonight and see what I get.
@mchipk
What decoder/version are you using?
(CELL:3)V01000200001005
from: REBUILDER.ECL
frame_first=0
frame_last=76
encode_first=0
encode_last=76 -----------gt; actual encoding from QuEcn and CCE = 74 frames ???????????
How many frames does the AVS say?
Ok. I found the cell you are talking about -- it is in VTS 09.
It is 76 frames. I loaded it into CCE and it was still 76 frames. I loaded the AVS into VirtualDub and it was 76 frames (and then checked the file itself and it was 76 frames).
I encoded it and it came out to 76 frames.
I'll go back to my last question -- what decoder are you using? Is it a beta, alpha, or release candidate???? Because it sure sounds like it is losing frames.
I'm using QuEnc 0.59[BETA2] and CCESP V2.67 trial
from: AVS
trim(7948,8023)
But have you tried doing only with the three cells?
V01000000001003
V01000100001004
V01000200001005
Perhaps if CELL:4 was added then CELL:3 (actual 74 frames) will take the next 2 frames from CELL:4 and make it 76 frames to fill in the gap?
It doesn't work that way. Besides -- why would I want to encode only 3 cells?
What DECODER are you using? mpeg2decodedg, DGDECODE, version? date? There are versions that are known to drop frames.
As I said... it isn't 74 frames. It is 76 frames.
I'm using MPEG2DEC.dll, I don't know how to check for version and date, but on right click property it says modified January 7 2003.
I just added cell:4, this time CCE shows 76 frames on CELL:3 on actual encoding, but CELL:4 shows two frames short.
(CELL:4)V01000300004001
from: REBUILDER.ECL
frame_first=0
frame_last=8610
encode_first=0
encode_last=8610 -----------gt; actual encoding from CCE = 8608 frames??
I guess it's a chain reaction to the rest of the cells as well, get 2 frames from the next cell to fill in the gap.
Is that the same MPEG2DEC.dll (modified January 7 2003)you have?
Ooops, thanks Jdobbs for pointing that out about decoders.
I just downloaded DGDecode.dll, tried it and it worked, the stutter is gone in CELL:1/CELL2.
Now I just have to encode the whole movie and see if any more stutter.
I wish on the setup of DVD-RB where it says quot;path to decoderquot; didn't ask for MPEG2DEC.dll
That way I wouldn't have gone and started this whole mess.
Again, thanks Jdobbs for the great help, and I apologize for the trouble.
mchipk.. actually its a good post.. IMO... show cases how important your decoder version is and may help people with similar issues. dvd rb still offers the choice since some people still prefer mpeg2dec.dll.
also a good post since everyone should own around the world in 80 days (jackie chan rules!!) |
|
Post at 2011-4-15 15:32
| This Author's Posts Only
