Log in

View Full Version : Need help with my first multicam editing project


avz10
27th December 2015, 01:13
I am editing a function, where the video recording was done with 3 video cameras. My idea is to render the video as a DVD and create a disk and to render another copy with BluRay settings, although I do not have a BluRay writer or player, but might get one in future. The people that I am doing this for, also do not have a BD player, but a big HD TV where I think a good quality video would be so much better.

The info of the 3 videos differ (analysed by MediaInfo). I am therefore unsure what settings I should use to get the best quality DVD and BluRay disk (I suppose one will be able to play the BD video by using a memory stick?)

As this is South Africa, we use PAL.
Looking at the 3 cameras, I have highlighted a few areas. The 2 JVC cameras appeared similar and are Everios. My Canon is a Legria HV40, which has since broken.
http://i261.photobucket.com/albums/ii62/avz10/3%20camera%20settings_zps1y2wlpix.jpg
As I only edit every few months, or now haven't edited for nearly 2 years, I found that it was a steep learning curve everytime I would go the Premiere Pro, AE route. I have used Pinnacle up to the last project, but have now changed to PowerDirector 14. This is not to say that I need to stay with these products. My first DVD's that I made , was made with 4 different software components and I can easily again go that route if someone can just give me guidance on what to use.

In the past, I have done a lot of transferring VHS to DVD, capturing and doing basic editing with VirtualDub and the Lagareth codec e.g.
http://forum.videohelp.com/threads/355843-Virtualdub-filters-For-8mm-Film-That-Was-Captured-On-Vhs
It has been suggested that I should create lightweight i-frame codec proxies and do the editing on them. This is a new concept for me, but appears very useful from the bit that I read.
It has also been suggested to render all of them to a single format of lossless YUV avi. 1920x1080p25 with pcm audio.
This is where I get stuck. I will probably know how to do it, but have no idea now where to start. Do I immediately start with e.g. Premiere Pro or PowerDirector or is some of the work done with other software prior to go into an editor?

I am just not sure how the editing process in practice will work-

Do I need to take the clips from JVC 1, edit the proxies or full quality clips and render; the same with JVC 2 and Canon. Then I have three rendered videos. I then restart with the timeline, this time I use three lines of the timeline and then decide what clip from what camera I will use, delete the others and render again. This time I have one video stream, but because it has been rendered twice, will I not lose quality?
Is this correct or have I got it completely wrong.

Can anyone please comment on the settings, the speed; interlaced or progressive; bitrate and what the maximum and average bitrate I should be aiming for?

Thanks so much in advance.

Warperus
28th December 2015, 13:35
1) Resolution. You definitely want 1920*1080 resolution, since all 3 cameras provide it for view. Cannon Legria uses HDV resolution with non-square pixel, that ends in 1920*1080 on the screen.
2) Fps. Your second camera uses 50 fps (progressive scan) and other 2 use interlaced 25 fps (50 fields). You can stick to 25 fps to keep it natural. TV sets present such material perfectly, and you can get perfect "interlaced" part from second camera. Much like cameras do interlaced material, you'd get half lines from each frame from your JVC2 Everio. Would you upload your movie to youtube, it would better to deinterlace everything beforehand and probably stick to 50 fps progressive, but I don't think it's needed for TV (also not every TV set plays 50 fps perfectly, some may fail in this case).
3) I'd say avc/aac in mp4 container. For final decision you better find TV set model and check if it accepts such combination (most modern TV sets do). Bitrate for video like 16Mbit/s should be enough. It's about what you have in source, but final renderer is more robust that camera encoder, so you can try lower bitrates. You can try 8Mbit/s, probably it will look good either. I don't think lower bitrate is useful unless you have to fit your movie into something like DVD.
4) For intermediate AVC-Intra 100 is quite enough for visual transparency. Don't worry about compression artifacts, they'll not be noticable after final render.
https://ru.wikipedia.org/wiki/AVC-Intra
Lossless is good choice for quality, but you will probably notice it is slower in playback due to big size. Don't do it unless you have a good reason to keep quality at lossless level, visual transparency is more than enough for mortals.
5) General process looks like this:
* grading - put colors in line, remove undershoot, overshoot, adjust white balance, drop some really unuseful stuff
* encode what you get from grading in intermediate format
* check what material you have, write a script
* get everything to trimmer, cut some things off
* get all the stuff together, synchronize different cameras

* try full quality intermediate, if it playback preview is smooth, you can avoid proxy creation
* if not, render proxies and edit in proxy mode - generally you don't need full quality for cuts/camera switches, sometimes you can need it for pixel-level effects

* switch cameras according to script or gotchas
* add some titles, text, few transitions
* add voice over if needed
* export timeline from Premiere to encoder
* encode in draft quality, probably proxies
* fix some things, redo rendering
* when it's all ok, render full quality material
* check full quality render, take a look at fine details in complex scenes
* fix some things, redo full quality rendering