Internal codec compared to external ProRes 220

I've taken 2 frame grabs from the original footage on your Google Drive.


IMAGE 1. (00001.MTS)
Here is a frame grab from the AVCHD file from the camera.
Notice the moire in the two speakers.

AVCHD from camera.jpg





IMAGE 2. ( PIX_001 mov)
Here is a frame grab from the ProRes file from the Pix.
Notice there is no moire.

ProRes from Pix.jpg


I've been using the PIX240 on most of my shoots.
During the online you can see the difference.

Bill Kerrigan
Montreal, Quebec, Canada
http://www.kerrigan.ca
http://fs700user.com/
 
Last edited:
Here is a frame grab from the AVCHD file from the camera.
Notice the moire in the two speakers.

Billl, I'm puzzled by this, because that moire is NOT in the original clip. Did you scale the clips / frame grabs at some point?

If I look closely at 100% magnification, there is slight moire in both clips -- the MTS and the PIX file. In fact I would say that it is slightly worse in the PIX file. But it's nothing like your grab above. If it had been, I certainly wouldn't have come to the conclusion I came to in my review.
 
Here's a ghastly screenshot showing the limits of the low bit-rate avchd codec. I shot this in 60p stupidly, but even with a lower frame-rate rustling leaves would've killed the codec totally.\

Edit: how do I post full-res screenshots?

vlcsnap-2012-09-24-14h01m48s81.jpg
 
Last edited:
Kerrigan: Your frame grabs are from playing the clip in the google doc player, which is pointless since it recompres the video files using the Youtube compressor for preview purpose. You need to download the file and play it locally. I confirm that there is no moire in the original file.
 
You are right Jean-Philippe...
But, I'm curious why the Google doc player created moire only in the AVCHD clip.
 
But, I'm curious why the Google doc player created moire only in the AVCHD clip.

Because it has less fine detail in the AVCHD clip to work with and I'm sure it does a quick and dirty encode job just to show it off to you as quickly as possible.
 
You are right Jean-Philippe...
But, I'm curious why the Google doc player created moire only in the AVCHD clip.

Because it has been deinterlaced

Often Google/Youtube will assume transport streams are interlaced and deinterlace before re-encoding . Not only that, the type of deinterlacing is horrendous (drop field)

Download the re-encoded mp4 versions at 1080p, and you will notice 1/2 the resolution is missing (since it's only made up of 1 field, instead of 2 for progressive content, that's why there is aliasing and moire patterns as the progressive frame is interpolated from 1 field)

Be careful with certain software and NLE's as well (many will do the same thing - deinterlace, unless you set them up correctly)

e.g. 1:1 crop of the re-encoded 1080 mp4's, prores right, avchd left . Notice the yellow text of the engineering manual
23k3tkh.png
 
Because it has been deinterlaced

Often Google/Youtube will assume transport streams are interlaced and deinterlace before re-encoding


Thank for answering PDR...
But wasn't the PIX footage also progressive?
 
Thank for answering PDR...
But wasn't the PIX footage also progressive?

They are both progressive. But note the PIX footage is in a MOV container.

It's what youtube/google "think" of the transport stream (MTS is a transport stream, think of it as a "container" that holds video, audio, metadata and other data streams ). It's misinterpreted as interlaced, so it applies a deinterlace. The footage in the MOV is properly identified as progressive.
 
Back
Top