Posted: Thu 22 Sep 2005, 4:25 Post subject: TMPGenc 2.5 vs 3.0 : pulsing
I set all my settings to the highest quality possible. In this case, 8000CBR 1I 14P.
Today I tried to encode some video in 2.5. While it turned out a bit darker than I was hoping, it looked great otherwise.
So I tried the same in 3.0 knowing that it's known for lightening the video up a bit. It, too, looked great, but a strange pulsating effect was clearly visible. The pulsing seems to correspond with where the I pictures are located. Funny how that doesn't happen in 2.5.
The motion isn't fast or anything, but it is an old camcorder video from the mid 80s so there's a fair bit of video noise. Should I try using some B pictures? Video noise reduction didn't help at all.
Joined: 04 Feb 2003 Posts: 587 Location: Lisboa, Portugal
Posted: Fri 23 Sep 2005, 17:36 Post subject:
TMPGEnc Xpress is called "3.0" but it's actually not based on TMPGEnc 2.5, as far as I can tell, so it doesn't surprise me that it has some bugs not found on 2.5.
Using B-pictures might disguise the problem, by introducing a faster flickering rate, but it probably won't get rid of it completely. My advice would be to contact Pegasys about the problem and, meanwhile, continue to use TMPGEnc 2.5 (non-Xpress).
Actually, I notice the pulsation in the video created with 2.5 as well. Again, it seems to pulse right where the I pictures are. It's like by the time the GOP has reached the 14th P it has degraded too much, so when the next I frame comes around it snaps back to what it should be.
Edit: I did some research on the net. I found limited information, but I did find one place that calls what I am experiencing "grain pulsing."
I did try the "Standard" GOP setting which does 1I 5P 2B and it does mask the effect better. Still, for a CBR of 7400 everything here tells me I should shoot for no B pictures and long GOPs.
Joined: 04 Feb 2003 Posts: 587 Location: Lisboa, Portugal
Posted: Mon 26 Sep 2005, 15:16 Post subject:
Technics wrote:
Actually, I notice the pulsation in the video created with 2.5 as well. Again, it seems to pulse right where the I pictures are. It's like by the time the GOP has reached the 14th P it has degraded too much, so when the next I frame comes around it snaps back to what it should be.
Or possibly the other way around. I would expect that, after 14 P-pictures, with stills or reasonably slow motion, you actually have better image quality (better in the sense that it's closer to the original) than right after an I-picture (because P-pictures can actually refine the image beyond what a single I-picture can show, due to the frame size limit). Make sure the VBV buffer size is set to 224 (the official DVD limit), a lower value could make a bigger difference to the grain.
Also, try using shorter GOPs (ex., I=1, P=4, B=0). This will probably cause a slight loss of detail, but most small detail is probably grain anyway.
Finally, if the clips you are encoding were originally shot on film, and later converted to video, it's possible that you will have some "temporal pulsing", caused by the frame rate conversion. Play the original tape and advance frame by frame, to see if there are ever two identical frames in a row. If that's the case, it's not something you can solve easily.
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
