My First XL2 Short...

[Pete Bauer]

Yes, there'll be a pull-up (or is it a pull-down...I get confused!) to 60i for TV when you play a 24p DVD, but it will not show interlace artifact because each field is shown by itself during it's 1/60th of a second...a smooth stream of images created by each field. Caveat: I guess newer TVs are coming up with increasingly sophisticated ways of handling various content so a fancy new HDTV might handle things differently.

Encoded to a 24p computer file, you'll still have true 24p for computer viewing (or film conversion, if you use 2:3:3:2, as I understand it). For interlaced, most multimedia file player software will mix 2 fields from a 60i file -- whether they originally "belonged together" in the original 24p miniDV footage or not -- into one full frame. No individual fields are displayed. That's why using the correct pull-down method for your purposes is important.

I think some of the newer software players like MS Media Player may be able to display each field as a 1/2 resolution frame at 60fps to give a smoother appearance like TVs do, but I'm not sure about that. I used a QT player to look at your file; I'll try Media Player when I get home and see if the artifacting still is visible.

If I've gone off track or missed an important point, all you 24p filmmakers out there feel free to put it right!

[David Lach]

Matthew, you don't want to shoot 24p 3:2 if you'll be editing in a 24p timeline.

Here's a kind of visual explanation of both 24p shooting methods, in which every letter represents a field of video, and every group of 2 letters represents a frame of two combined fields. In both cases, the camcorder captures a AA BB CC DD sequence in progressive mode. What changes is the way the camcorder brings it to 60i video for exporting as a regular DV format file:

3:2

AA AB BC CC DD

Here, as you can see, the XL2 creates a pattern of 3 frames / 2 frames, hence the 3:2 definition. As you can tell, in post, if you want to bring back the sequence to the original AA BB CC DD progressive sequence, there is just no way to recreate the BB frame without recompression. Looking at this, you'll see that if you edit in a 29.97fps timeline, you'll have 2 out of every 5 frames that will be interlaced. If you want to bring it back to 24p, the software will need to decompress and recompress in order to recreate the BB frame, which will cause a lost in resolution.

2:3:3:2

AA BB BC CC DD

Here, you can see that the extra frame the XL2 has created to bring the sequence from 24fps to 30fps is the BC one. This method is strickly intended to edit in a 24p timeline, because it does not look very smooth to the eye when watched without removing the extra frame. But, to edit this sequence in a 24p timeline, no recompression is involved, you just remove the fake extra BC frame and you're left with the intact original AA BB CC DD sequence to edit with.

If you plan on shooting in 24p for Web Delivery, you'll definitelly want to use the 2:3:3:2 option. If you plan on editing in a 24p timeline, same thing. 2:3:3:2 again for film transfer or any kind of big screen projection. If you want to create a 24p DVD, this option as well.

The only reason to use the 3:2 method is if you plan on editing in a 29.97fps timeline or deliver it to a regular NTSC TV. TVs show the video as interlaced, one field at a time instead of one frame at a time, so you will never see those interlacing artifacts on a regular TV set. Only when you show it on a progressive capable monitor (computer screen or big screen).

To sum up: 3:2 method to edit in a 29.97 timeline, 2:3:3:2 method to edit in a 24p timeline, regardless of final output. The 3:2 method looks better to the eye unprocessed and left as is to view on a TV set, but the 2:3:3:2 is intended for removing the fake frame and working with the true 24p footage. If you want to output your 24p sequence to TV, you just render a sequence in 60i with a 3:2 pulldown.

[Matthew Cherry]

Ok, I got, mostly. The only thing I'm still confused on is making a DVD. If the DVD is going to be shown on an NTSC TV, wouldn't you have to edit it in a 29.97 timeline? OR are you saying that you would do it all in 24p advanced pull down, editing in a 24p timeline and then allow your compression software to take care of the conversion when making the .mp2 file?

[Marty Hudzik]

<<<-- Originally posted by Matthew Cherry : Ok, I got, mostly. The only thing I'm still confused on is making a DVD. If the DVD is going to be shown on an NTSC TV, wouldn't you have to edit it in a 29.97 timeline? OR are you saying that you would do it all in 24p advanced pull down, editing in a 24p timeline and then allow your compression software to take care of the conversion when making the .mp2 file? -->>>

You are correct. Most mpg encoders will allow you to encode a 24P stream and they will simply flag the appropriate frames that need duplicated. The DVD player sees this flag and adds the 3:2 pulldown on the fly in hardware. In addition the fact that you are actually only encoding 24 frames instead of 60 fields saves you space and can actually increase the picture quality.

Another advantage of this is that if you play it on a PC or MAC it will not add the 3:2 pulldown but just play the original 24p stream. I would guess that on HDTV that are progressive this may happen also but I am not sure.

Good Luck!

[Pete Bauer]

Ok, I posted three screen captures that showed the artifacts as I saw them in my PC's Quicktime 6 Player:

http://www.geosynchrony.com/scratchpad.htm

I was NOT able to open the mp4 file in Windows Media Player, v10 (the most current version), nor in Premiere Pro 1.5 on my PC.

Based on all the nice explanations by David and Marty, it sounds like it wouldn't matter for plain ol' TV, but any 24p output might be problematic now because of the 3:2. You've got those pesky 60i "AB" and "BC" frames in there splitting up your original 24p "BB" frame...no easy way to delete a 60i frame and recover a pure "BB" frame for 24p output. There's probably a work-around but I don't know what it is.

FWIW, I've read many recent posts -- most of them more than once -- about the two pulldown methods and it just didn't quite all gel in my brain. But digging into this little artifact issue on your project has helped a bunch...hopefully for you, too. Thanks for being willing to share your work, which I'll say again I find to be superb!

[Rob Lohman]

Excellent Work Matthew!

[Richard Alvarez]

Nice work Mathew. I think you've got a fair handle on "noir". Check out our trailer at www.nu-classicfilms.com "After Twilight" was shot on 35mm, but we were going for the classic noir look

[Tony Roulston]

I have say how much I admired your short.

I am not quaified to provide critical comment, you have enough of that in this thread, but I want to let you know that I found your short great.

TR

[Matthew Cherry]

Ok, I've decided to expand on this short to make it more of a narrative. I think we've developed a good film noir script and I'm getting excited about doing this. Here's my question.

If I decide to do this, should I reshoot everything so far in 24p advanced pulldown to avoid getting any artifacts? The reason I'm asking is because I might want to try enter and enter the finished short in some small festivals. Nothing major just some local ones.

What would you reccomend?

Matt

[Pete Bauer]

My understanding is that you won't be able to get the 3:2 footage to look right transferred to film without a lot of fiddling around, if at all. (Those pesky frames with mixed "A" and "B" fields again!) So unless someone has a trick up their sleeve, you might end up having to re-shoot using 2:3:3:2.

I actually have a similar problem with a lot of old home Reg 8mm films that I had converted to miniDV. Hopefully someone knows of a practical way to split the fields back into their original progressive frames -- I sure don't! It has been driving me crazy!

If nobody has a good answer in the meantime, I'll give our similar problems more research and thought later in the week. My wife leaves on a trip in a couple days so I can let the Honey-Do list slide a little! (Hope she doesn't read this!) ;-)

[David Lach]

Matthew, if you want the best results for projection, especially on the big screen where resolution counts a lot, yes, you will need to reshoot in 24p 2:3:3:2 and edit (and output) in 24p. Check if FCP will detect and remove the extra fake frame by itself to recreate the 24p sequence, like Premiere Pro and Vegas do, otherwise you'll need to buy a stand alone software like DVFilm Maker to throw that bogus frame away and edit in a 24fps progressive timeline.

Pete, unfortunatelly if you've acquired your 8mm footage using interlaced video, there's not much you can do, aside from recapturing the whole thing with a progressive camera. There's different softwares and plugins to de-interlace your footage and get rid of most artifacts, but that will come at the expense of resolution as frames will be blended together.

I hope we'll some day watch the interlaced format die once we can get affordable full frame 60p (or more) video become the new standard. But for now we have to live with the hastles of interlacing.

[Matthew Cherry]

Thanks David. What about filters. I don't think I would use any for this sort of thing, but I've been experimenting with a Tiffen Black Diffusion to get more of a film look and I do like the results - but that's on my TV. Would using such a filter make the image too soft for projection?

[David Lach]

This is a personal opinion and others might not agree with me on this, but I never use filters unless they are absolutely necessary. If you want a certain look, you can dig deep in the XL2's menu and tweak the image to your liking. If that's not enough, you can manage in post.

But like I said, I'm biased, I do not like to put a potentially cheap piece of glass or worse, resine, in front of high resolution and very expensive lenses. Filters always increase to potential to soften the image, introduce distortion or worse, sometimes completely ruin a shot by creating ugly flares you might not notice when shooting.

I will often use a polarizer when shooting under a bright sky outside, or to eliminate reflections coming from windows or water, and sometimes you will have no choice but to use some NDs, but I would certainly not recommand using any kind of diffusion filters if the intent is big screen projection. Don't forget that blowing up DV footage will introduce softness, so you want to keep your image as sharp as you can between the shoot and the screening. If it already looks soft on TV, it will likely look too soft when projected.

Best solution is to add a diffusion filter in post to your version that will end up on TVs, and do nothing (to soften the image) to the version that will be projected.

[Matthew Cherry]

David,

Would this still hold true if you weren't blowing up to 35mm but instead were going to be digitally projecting it?

Matt

[Matthew Cherry]

I'm not sure I'd have to shoot that way for projection... I'm still a bit confused on that although I now understand the difference between the two modes.

Thank you so much, I appreciate the feedback!

Matt