3:2 pulldown supported?

[Jeff_DML]

I assume yes but searching these forums and the options I have not found out for sure. The reason I ask is I am seeing some stuttering playing movies from DVDs. When I play them on another player the stuttering goes away. When I play a 720p movie I got from OTA is plays smoothly on VLC.

If so is there a option I need to set? I dug through the options and couldnt find anything.

thanks
jeff

Edit: Didnt see any glaring problems when I display the messages

[Jean-Baptiste Kempf]

what do you mean? Cropping? Aspect ratio?

[Jeff_DML]

basically 24 fps being displayed 60i, EDIT: guess more accurate since we are talking computer world/progressive 24->60p

http://www.dvdfile.com/news/special_rep ... lldown.htm

sometimes is called film mode detection

[Jeff_DML]

from the 138182 spec

D.2.2 Movie timing
A decoder can implement 3:2 pull down when a sequence of progressive pictures is encoded. Each
encoded movie picture can independently specify whether it is displayed for two or three video field
periods, so “irregular” 3:2 pull down source material can be transmitted as progressive video. Two flags,
top_field_first and repeat_first_field, are transmitted with the Picture Coding Extensions and adequately
describe the necessary display timing.

[DJ]

There seems to be a number of errors in your reference as I understand this.

1. The video on a DVD is stored as 23.976 FPS with a 3:2 pull down flag (when the project is complete most Telecine artists choose to delete the flag) and in a fashion that equates to three progressive frames and two interlaced frames on most DVDs. But with the pull down flag removed the video will run at 29.97 FPS. I say most DVDs because this the most common. But really it up to the Telecine artist and how he sets up the equipment. This is NOT 480i as this standard is a 856x480 aspect ratio or 1.78:1 but 480p or i and is flawed in that 856 is not divisible by 16 so DAR is used to store 720x480 that expands to 856x480 (anamorphic video) in a player, However transcoding these will only yield an 854x480 picture so a few lines of pixels are lost if you want to maintain the aspect ratio. I have heard the term 480i/24, but the term is misleading. Also in the transcoding process an inverse telecine could be applied to make the end result fully progressive, but the end result is 23.976 or the audio would not sync picture and the picture would end up being jumpy or drift over time. An inverse telecine is really not necessary if the pull down is always 3:2 as the player can always interpolate the proper number of frames (progressive scan), however for other rates an inverse telecine would be necessary to make the end result progressive. In early players the math used was not quit right for a 23.976 FPS and a 3:2 pull down and the picture was jittery but this was not the only problem as the audio and video clocks needed to be fully locked, so when the math was corrected and PLLs (phase Locked Loop) were employed to lock the clocks, the term "Progressive Scan" was born.

In computers you can NOT use PLLs so video can never fully "Progressive Scan" from a DVD without usin a transcode process. Considering that all digital video is lock to audio for sync, if the sound card does not accept control from the program or it does not control the video well, there will be problems that range from video that won't play to jerky video to simple jitter. Also the refresh rate for the video enters into this equation as it should be a multiple of the frame rate.

2. The conversion is made to 23.976 FPS and not 24 frames.

3. This method provides backward compatibility to most 4x3 TVs that require interlacing along with black borders for 16x9 content that is displayed in a 4x3 frame. By this method (3:2 pull down) and others described here a 4x3 TV that wants an interlaced frame can display a progressive frame.

I'm not sure I agree with some of the other things this guy says as the process is not as complex as he seems to want to make it, at least as stated in the URL you provided. The only time I have run across line doubling is, in picture reconstruction or enhancement in combination with error interpolation, you can use sum and difference to create depth of field. More like a 3D image. The only success I have witnessed in this area is Sony who became king of proprietary interpolation methods.

In error interpolation you use a read ahead buffer and when a error is encountered the information before the error and after the error are compared and a new buffer is created with the most likely information so the error is avoided. In enhancement the buffers are always in effect and the lines of the picture are constantly being scanned. To make this more effective in enhancement the lines are doubled and the sum and difference compared. The new picture is then output with the added information (resolution not necessarily lines). This process could only be done in dedicated hardware. To the best of my knowledge another manufacture has never attempted this.

[Jeff_DML]

looking through the code it looks like the flags are handled properly, thanks