Canon XL2 - Post-converting workflow

[Manietape]

Hello everyone,

I'm shooting a film with a Canon XL2. I'm in the process of converting the tapes. I'd never worked with a Mini-DV before, and before moving on to post-production, I realized I'd learned some things about improving the image quality, especially considering the possibility of a large-screen display.

First, I discovered deinterlacing. I record at 24 fps with my camera. I don't notice the horizontal lines when I pause or play my footage. However, I learned that this process is essential even if my camera records progressive video. To do this, I watched a series of videos on Hybrid. However, I noticed that they were outputting as .mov. It didn't make sense to do color grading and correction with a processed image on Davinci, as I'm outputting with a .dv extension. Wouldn't it make sense to work with this raw footage? Furthermore, I couldn't think of anything else other than deinterlacing the entire hour-long video clip on Hybrid. After editing in Davinci, I didn't know how to transfer the parts I'd included in the edit. I thought about editing the sequences I'd use on a single piece of video. But as I said, outputting with a .mov extension didn't make sense. What do you think I should do?

I'm also considering the image enhancement process in Topaz. Some sequences are more grainy than others. I'm planning to work on these separately. But should I enhance all the converted 1-hour video segments like I did with Hybrid? Should I also color correct the rendered footage? Finally, wouldn't printing a 4K output and then editing it again in Davinci be a disadvantage over the rendered footage? I've seen Topaz's Davinci plugin. I think image enhancements can be performed through this panel. But does it offer 2K or 4K output?

I've asked a lot of questions, so I apologize for that. I haven't really been able to articulate the workflow. I appreciate your help in advance.

Regards.

[Alwyn]

Here's my take.

How are you downloading your video? On more basic DV camcorders, we use WinDV or Scenalyzer to transfer the DV. It downloads as DV AVI (Interlaced, 720x576+25fps/720x480+29.97fps, around 13GB per hour).

If you're shooting Progressive, there's no need/you shouldn't deinterlace, because there's no interlacing to "de".

If you're interlaced then yes, you probably should deinterlace so it's 24fps Progressive, like a real movie. Hybrid can deinterlace using QTGMC.

If you're going to edit in Resolve, I would just deinterlace then export from Hybrid in a format that Resolve accepts. I haven't used Hybrid for a long time but IIRC it can output FFV1 and ProRes, both of which Resolve will read.

If you don't need to deinterlace, just import your DV file straight into Resolve and edit away. Resolve will accept DV-AVI.

After editing in Davinci, I didn't know how to transfer the parts I'd included in the edit.

I don't know how to use Resolve but generally, you'd just work on each clip/segment on the timeline until you've made up your movie, then export it. You could export the bits you've worked on and then bring them back onto the timeline later but you should export as lossless so you don't lose quality.

Re your noisy video, NeatVideo is an excellent denoiser and has a plugin for Resolve so you can denoise just bits of your clips.

IMO you should just edit it all at the native res (720x576/480) and then, if you wish, upscale at the final export step.

As far as Topaz goes, can't help.

[Alwyn]

Just had a look through the manual. 24P is probably what you want for your film (not 24P 2:3 or 24P 2:3:3:2).

It would be interesting to see what video format the camera actually outputs; you can use MediaInfo to get access to the files attributes: go to the "Text" view and copy and paste it here. Or you could attach a file here; around 15-30 seconds would be good for analysis.

Resolve accepting 24P DV could be an issue...

[johnmeyer]

I didn't look at the manual, but I seem to remember that a lot of these early semi-pro DV camcorders which first offered "24p" actually wrapped it into a 29.97 stream. I just spent ten seconds and asked ChatGPT and it confirmed what I remembered:

The Canon XL2 offers a "24p" mode, which simulates 24 progressive frames per second.

It uses a technique called progressive frame capture with 2:3 pulldown, recorded in a 60i (interlaced) signal.

This means the camera captures images progressively (not interlaced), but records them within a 60i interlaced stream, compatible with standard DV equipment.

I'd have to look at a short sample clip containing motion to know what tool to use in order to end up with true 24p. You most definitely will need to do that conversion before you proceed with any editing.

[edit]According to the same source (ChatGPT) all you need to do is apply standard IVTC (i.e., pulldown removal). It recommended Telecide, but I always use TFM/TDecimate, because that's what I've used for 20+ years.

[videoAI]

@Manietape

Hybrid isn't the only tool you can use.
FFmpeg with GUI can help you in many areas.

try clever FFmpeg-GUI, it supports a wealth of features, including IVTC/Decimate


[Attachment 89124 - Click to enlarge]

[Alwyn]

I just spent ten seconds and asked ChatGPT and it confirmed what I remembered:
The Canon XL2 offers a "24p" mode, which simulates 24 progressive frames per second.

That's not the whole story. It also "offers" true 24P, as per the manual (AI obviously doesn't/can't read manuals):


[Attachment 89132 - Click to enlarge]

As I have already said, the OP should use 24P (not 24P 2:3 or 24P 2:3:3:2).

No need for IVTC.