Tagged with tester - Personal View Talks http://www.personal-view.com/talks/discussions/tagged/tester/feed.rss Fri, 17 May 24 00:35:42 +0000 Tagged with tester - Personal View Talks en-CA PTool Questions http://www.personal-view.com/talks/discussion/456/ptool-questions- Thu, 21 Jul 2011 19:37:18 +0000 stoney 456@/talks/discussions , I'm a total noob tester. I know. RTFM. But there's no fuking manual!!! Going through piles of postings? Are you kidding me? So here it goes. Please feed me more info. I will continue to update this 1st post as I learn more. But this won't cover anything in depth. Just enough knowledge to get started for noob testers.

Basic Info


First read GH1 Firmware Hack FAQ. Then Bayer Filter, YCbCr, Chroma Subsampling, Intra Frame Coding, Inter-frame (P-frame) Coding. See @driftwood's post about various topics. @cbrandin's post corrects some misinformation.

Patches For End Users


MJPEG Size Adjustments: I guess this is where anamorphic settings done. See Anamorphic Cinemascope in Native 1920x810 MJPEG Mode. An ingenious idea. Instead of stretching image during post-processing, let the camera stretch image frames horizontally where the frames are vertically stretched by anamorphic lens. 1920x810 MJPEG + 1.33x anamorphic lens combo gives undistorted 2.33:1 image straight out of the camera. 2x anamorphic lens + 1280x960 MJPEG combo gives 2.66:1 undistorted image. See MJPEG 4:3 for 2x anamorphic shooting.

MJPEG 720p Compression
  • E1 >= E2 >= E3 >= E4: http://timelapser.net/about/ says "MJPEG quality settings are a tradeoff between image quality and recording duration. Setting E1 Table: 4 makes a big difference. My understanding is that the E1 Quality/Table setting is used initially and E2/E3/E4 are a fallback if the camera can’t keep up. As we’re only recording at 2fps this is unlikely to be a problem, but I recommend using settings that descend down from your E1 quality setting. E1 Quality: 400 will give you super-high quality jpegs, but only about twenty minutes of recording time at 1080p."

MJPEG 480p Compression
  • F1 >= F2 >= F3 >= F4: Prolly similar to the E* family

AVCHD Compression
  • Video Bitrate FSH/SH: for 1080i60 FSH and 720p60 SH
  • Video Bitrate FH/H: for 1080i60 SH and 720p60 H
  • Video Bitrate 24H: for 1080p24 H
  • Video Bitrate 24L: for 1080p24 L

AVCHD GOP Length: I frames, B frames, P frames. Vitaliy recommended a factor of a targeted mode. e.g. GOP 3/6/12 for 24p. Official Low GOP topic is covering the topic well.

Microphone AGC 0 Setting: ORLY? Almost disabling AGC? Nope. This is a setting for level 1.


Patches For Testers


AVCHD Compression
  • Video Bitrate 24H ADD: How is it different from the end user's Video Bitrate 24H setting? @Vitaliy Very different. You can try to use it for testing, but normally you don't need to.
  • Overall Bitrate & Bitrate 2: This applies to all video modes? @Vitaliy Yes. It is some time related constant, in fact. Not bitrate. But it can affect bitrate.

AVCHD Research: Video Buffer, Video Buffer 24p, Audio Buffer: Increase them as increasing Video Bitrates?

AVCHD Encoder
  • Quantizer tables: http://www.personal-view.com/talks/discussion/366/pro-avchd-quantization-process No freaking idea. Maybe @driftwood's post helps.
  • Quantizer scaling tables: See @driftwood's post
  • Frame buffer size: It covers all modes.
  • Encoder frame limit: What?
  • Bitrate settings: High Top, High Bottom, Low Top, Low Bottom? Always keep lower bitrate settings within 95-70% of high bitrate settings. Otherwise, you'll see a dramatic bitrate drop in the low modes. There is currently a known drop in bitrate in the first frame set or GOP - known as the blip. Still to find out a definitive answer as to why.

ExTele check disable: I don't understand the description, "Disables ex tele converter checks, so it won't reset as usual"


Misc


For OSX users, mtscopy.app makes it easier to copy MTS files. Install virtual WinXP or WIn7 and launch StreamParser.]]>