FORUMS: list search recent posts

Rectangular Pixel Encoding

COW Forums : FFmpeg

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Zeus Agbaosi
Rectangular Pixel Encoding
on Jan 14, 2018 at 1:13:04 am

Hello members

I'm looking forward to getting a recording system for high definition formats for games, maybe like Nvidia's ShadowPlay, but at the moment I'm working with the older happauge and sony capture cards for footage. I had a question about this 720x480 format that is mutual between those cards and even my older replaytv device. When the content is resized and encoded directly from my PC card, when does the rectangular pixel placement actually occur? Is it directly when it first outputs over S-video or when it's received by the capture card of the same interface?
I've heard it said that monitors are pretty straightforward when it comes to receiving output, so it stands to reason that when re-encoding systems are involved, there's more processing going on than for a direct line.

My method for post processing is to use the Handbrake program to deinterlace and instantiate the footage from my capture to a near lossless bitrate, then using Sony Vegas I crop the borders (so its 4:3ish) and finally resize it to a size similar to its original resolution setting from in game. Sometimes I use VLC to do the job, but having the graphical representation of the rendering progress in Vegas helps me stay patient.
What I've noticed is that regardless of what bitrate setting or denoise utility I use, there's still a little bit of Dot Crawl or herringbone to the output. The resize helps for quality, but the eye follows the visual artifacts left behind.

I first thought that it was just Vegas automatically detecting that 10/11 PAR from the encoder, but after messing with some settings in Handbrake, I believe that its the interpolation settings that Vegas applies that's causing the non-square pixels to be so obvious. Frame Blending and Interpolation techniques are an area that I know very little about.

I'd like to know which problem to attack, should I focus on the nature of square vs non square pixels or should I read up a little more about motion blur and how frames fit together. All in the hopes of furthering my own understanding of how encoding works. There's always going to be new problems to face with new hardware.

Sony MPEG RR Engine [Sony VGN PRAV1]
MediaPortal TS {ITU-601} [Happauge WinTV 418 Tuner]

Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2019 All Rights Reserved