Best 24p native solution with hack

[alexberman]

It seems that with the hack there are three ways to get 24p natively with a G13:

1. Shoot with 24p Native checked in ptool (unstable)
2. Shoot 24p in 60i and do pulldown removal (time and processor intensive)
3. Shoot 25p in 50i and slowdown footage when you have a final edit (minor pitch shift)

Since 25p in 50i does not introduce any interlacing artifacts from pulldown, isn't option 3 the best solution? It is the fastest because you must preprocess all footage to remove pulldown in 24/60i with neoscene or voltaic. It is the most stable because it does not rely on the 24p Native hack. And it is the highest quality because it is true progressive without interlace from pulldown cadence.

The pitch shift seems to be the only drawback, but changing the speed of the finished clip minutely (which is routinely done on every Hollywood movie for european DVDs) is far preferable to preprocessing pulldown on EVERY clip with neoscene, compressor, or voltaic.

Assuming you want 24 native frames (not broadcast), does anybody have any other ideas why 25p in 50i is worse than option 1 or 2?

 

[MikalZ]

What you are basically saying is why doesn't everyone shoot PAL? It's a good question and I've wondered this myself.

The reason I personally haven't adopted this workflow is that I don't like dealing with AVCHD directly, and would not want to work with the raw clips on a 50i timeline. This means the clips all still have to go through neoscene, and if I am doing that anyway I may as well have neoscene output 24p clips.

But that said, your point about pulldown artifacts (which I have seen), and clip conversion speed is well taken. I may try the 25p workflow at some point.

 

[svecher]

Since 25p in 50i does not introduce any interlacing artifacts from pulldown, isn't option 3 the best solution? It is the fastest because you must preprocess all footage to remove pulldown in 24/60i with neoscene or voltaic. It is the most stable because it does not rely on the 24p Native hack. And it is the highest quality because it is true progressive without interlace from pulldown cadence.

Proper 3:2 Pulldown removal does not introduce any interlacing artifacts. Deinterlacing does. Some modern NLEs (CS5, Eduis) do pulldown removal on the fly and do not require pre-processing with NeoScene or Voltaic. This approach requires fairly modern hardware and software, but [at least to me] that's more preferrable than 25p-24p conformance voodoo.

However, for me I shoot with native patch + combination of additional factors and try to avoid the native death scene (static high detail) whenever I can. If I must shoot a static scene I have a "reliable settings" card, where native patch is checked off.

 

[noirist]

Proper 3:2 Pulldown removal does not introduce any interlacing artifacts. Deinterlacing does. Some modern NLEs (CS5, Eduis) do pulldown removal on the fly and do not require pre-processing with NeoScene or Voltaic. This approach requires fairly modern hardware and software, but [at least to me] that's more preferrable than 25p-24p conformance voodoo.

Unfortunately Edius does not correctly remove the 3:2 pulldown in GH1 footage. My impression is that it correctly deinterlaces 60i to 30p but then incorrectly blends the frames to get 24p.

However, for me I shoot with native patch + combination of additional factors and try to avoid the native death scene (static high detail) whenever I can. If I must shoot a static scene I have a "reliable settings" card, where native patch is checked off.

What settings do you use for native and for reliable non-native? Thanks.

 

[mojo43]

Unfortunately Edius does not correctly remove the 3:2 pulldown in GH1 footage. My impression is that it correctly deinterlaces 60i to 30p but then incorrectly blends the frames to get 24p.

I have been using Edius 5.10 over the past while and it seems to me that it correctly removes pulldown, but only when you "Print to file" from the project (all other ways were unsuccessful).

 

[alexberman]

In the FCP environment it seems that 25p in 50i processed through Log and Transfer with conform at the end is reliable and quick, allowing you to edit in prores without the additional overhead of removing pulldown. Thanks for the heads up about Edius.

 

[c.cook88]

In the FCP environment it seems that 25p in 50i processed through Log and Transfer with conform at the end is reliable and quick, allowing you to edit in prores without the additional overhead of removing pulldown. Thanks for the heads up about Edius.

What do you mean by "conform at the end"? Right now I'm using compressor to remove pulldown... is there an easier way?

 

[tonpole]

Proper 3:2 Pulldown removal does not introduce any interlacing artifacts. Deinterlacing does.

What do you think pulldown is? Deinterlacing and frame dropping. Your losing quality by not shooting 24p native.

 

[cbrandin]

Not true. First of all the interlaced frames the GH1 produces are not actually interlaced frames, they are progressive frames wrapped in interlaced frames. Second, the "dropped" frames are duplicate frames added to the stream to achieve the 60i wrapper instead of a 48i wrapper. The only difference has to do with whether the 4:2:0 color subsampling is applied to each interlaced half-frame separately of both half-frames at once.

Chris

 

[noirist]

I have been using Edius 5.10 over the past while and it seems to me that it correctly removes pulldown, but only when you "Print to file" from the project (all other ways were unsuccessful).

"Print to file" doesn't remove the pulldown correctly either. Try editing Blackout's original GH1 footage in Edius and see if you can get Edius to export it in 24p without shimmering artifacts. I've tried all the obvious ways (including various ways of exporting from a 60i project, a 30p project, and a 24p project) and none work correctly.

http://ediusforum.grassvalley.com/forum/showthread.php?t=16016

 

[mojo43]

"Print to file" doesn't remove the pulldown correctly either. Try editing Blackout's original GH1 footage in Edius and see if you can get Edius to export it in 24p without shimmering artifacts. I've tried all the obvious ways (including various ways of exporting from a 60i project, a 30p project, and a 24p project) and none work correctly.

http://ediusforum.grassvalley.com/forum/showthread.php?t=16016

Ok, just tried his file.... I have no problems removing pulldown with his file. What am I missing here?

 

[mojo43]

Just checked again... no problems whatsoever... weird! What version of Edius are you guys running?

 

[noirist]

I'm running Edius Neo Booster 2.50.

You were able to import the original MTS file http://blackout.com/ediusproblems/original-gh1-NYC-1080.MTS directly to Edius and export it in a 24p format without any shimmering? If so, please publish your project settings and your export parameters. Thank you!

 

[mojo43]

Yes I was able to (unless I am missing something)

Software: Edius 5.11
Project settings:
Hardware : Generic OHCI
Video
Frame size : 1920 x 1080
Frame rate : 23.98( 24000/1001 )
Pixel aspect : 1.0000
Field order : Progressive
Audio
Rate : 48000Hz
Sample size : 16 Bit
Channel : 2
Setup
Render format : Canopus HQ Standard
Over Scan Size : 3 %
Audio Reference Level : -20.0dB

Remember that I have tried the ACHD Conerter as well as converting from the bin and this does not work for me. The only way that I have found that works is using the "Print to file" option. I am wondering why it hasn't worked for others as I have read threads about this and have not had any problems....

Let me know what you come up with.

 

[mojo43]

Maybe I should have asked this first... what do you mean by shimmering?

 

[noirist]

Maybe I should have asked this first... what do you mean by shimmering?

Look at the trees in the background -- in the original MTS footage they look normal but in the 24p export from Edius they are shimmering.

 

[mojo43]

Oh I wasn't looking at that footage. I was looking at the one supplied by blackout with all of the people walking by. I converted that one and posted it on youtube. Take a look and let me know if the shimmering is there. I unfortunately don't see what you are talking about...

IMPORTANT NOTICE: No media files are hosted on these forums. By clicking the link below you agree to view content from an external website. We can not be held responsible for the suitability or legality of this material. If the video does not play, wait a minute or try again later.

I AGREE

 

[noirist]

You converted the right clip. Take a look at the tree in the center of the background.... See how it shimmers? That's an artifact introduced by Edius that's not in the original footage.

 

[mojo43]

Wow, I never would have noticed that. I would have thought that was just aliasing... sorry about the confusion. I have posted three projects (not paid) all ready and never noticed anything wrong. Thanks for the heads up!

 

[tonpole]

The only difference has to do with whether the 4:2:0 color subsampling is applied to each interlaced half-frame separately of both half-frames at once.

Chris

So whats the difference? And how would one determine this?