Your workflow with video

[sorvju-f]

I am just learning finalcut and thinking if I can find better workflow for my work. I normally do batch capture, edit it to time line, taking sequense to color and do the adjustment. After that I go to voice editing.

Do you collect the best clips somewhere and how? At the moment I don't.

How about back-up...I have only cassettes as back-up?

What is today the best way to add your final work to web? I remember that David had great quality with DivX.

Jukka

[David White]

I batch capture all the tapes and then systematically go through each clip to trim the best ten seconds or so. Since I do things backwards by editing the video to the music, the selection of an audio track is an important consideration in my compositions. Once an audio track is selected the clips are arranged on the timeline and further trimmed to coincide with the downbeat of the music. Once satisfied with the flow of the video the clips are cropped and colour corrected and titles are added. For larger projects like the "Visions of Lembeh" DVD, each track is edited as a separate entity and then re-assembled into the final project for rendering and output to DVD. Otherwise it takes an interminable amount of time to load the project. Once the project is complete a copy is made to tape and then stored with all the other tapes in an on-site insecure facility...what am I thinking?

Although DivX is a great codec for internet distribution there are few sites that will host DivX files and few users to download them. It seems as though Flash has won the battle for supremacy. Sites like vimeo convert your .mpg files to flash before hosting on their site.

Since starting out on this underwater videography adventure I have gained a new found respect for editors. Ten days of diving in Lembeh resulted in ten hours of film which required ten months to edit. However, I am very pleased with the results.

[Jonathan Bird]

Here is my basic workflow. Key words are bold.

1. I begin the shoot with either a script or at least an outline of the segment I'm shooting. From that I create a shot list of shots I must have. Of course I will get shots I didn't think of, and some of the shots on my list might not happen for one reason or another, but the shot list helps me have a good idea that I will get all the shots to tell my story. Because my work is documentary in style, I am always telling a story, so I need to be sure I have all the shots to tell it. This is different from a style such as that David White often creates which are more like underwater music videos (no narration, all video set to music). He has a totally different approach to shooting/editing.

2. I shoot everything on location to tape. I do not spend precious time on location doing things like logging tape, capturing material, etc. I do sometimes go back and check shots at the end of the day to be sure I have things I think I have, particularly things that are very important. If I get a shot that is very rare or hard to get, I may capture it to my laptop just to be sure I have it on a hard drive in case of a tape drop out. This is rare.

I have no interest in hard drive/flash card cameras because I do not have time for the media management issues associated with dumping cards, backing them up on location, archiving them at home, etc. Tape is cheap, reliable, archival and proven. HDV tape is fantastic in terms of error correction and reliability--it's rare to get a dropout. (But when you do, gulp, there goes half a second of GOP).

3. When I get home, I sit down and go though each tape methodically. My capture process is to go through one whole tape in Final Cut, logging the in and out points of each shot I might need, naming the shot, rating the shot according to how good it is, putting comments in for it (like "best shark shot") and then SAVE.

Once the tape is logged and the project saved, then I do a batch capture on that tape. The machine does that on it's own, but you need to keep an eye on it because FCP is stupid and bails out on an error, rather than just skipping the shot and going to the next one.

Once the entire tape is captured, SAVE again. Now move on to the next tape. I tend to capture more shots than necessary (hard drive space is cheap) but not longer shots than necessary. (What I mean is that I might capture 3 different takes of a shot, but I don't generally just combine them and capture one long scene because it's wasteful and less organized to hunt for later. I like each take in it's own file. It's a little more work but more organized and more efficient with drive space. Also, then I can delete takes I decide not to use later to free up more drive space.)

By the way, buy the biggest hard drives you can afford....they fill up. If you buy huge drives, you will never need to delete any media to create space. When I run out of room on a drive, by that time drives are cheaper and I can buy a new one that is 4 times as large as the previous one for the same money, move the media and keep capturing. Then I have all my media available all the time. Also back the drive up with a duplicate, and you will never cry that you lost your media.

Back up your FCP project files and if you do lose your media, at least you can re-capture from the tapes because the original project file has all the information in it.

One other tip: name your tapes logically in whatever way you like and don't forget to put that name into your editing program when you are capturing. It isn't much help later if the you need to re-capture footage and all your tapes are named "untitled"--you won't know which footage came from which tape. Naming your tapes and being extremely careful about using those names during capture is very important.

4. One of the advantages of the logging/capture process is that it forces you to watch all your footage. Once all the footage is captured, I have now seen everything and I have a better idea what the segment will look like in my head. I usually sit down and re-write the script based on my knowledge of the strengths and weaknesses of the footage.

5. Now I record a "scratch" narration track by reading the script into a microphone.

6. I edit a rough cut of the segment using my scratch narration. Of course, as I am editing, I come up with changes in the script due to how things cut together. So I mark changes into the script as I am working. Sometimes I re-read the changes into the scratch narration for pacing, or sometimes I just leave it out, with a mental note of it and a piece of video in there which is about right for the missing/changed narration.

7. Once I am happy with the rough cut, I go into my sound booth and do the final read of the narration. If someone else is narrating, I need to send the script off to him at that point. I am a lot looser with my procedure when the narrator is me, because changes are easy. If I am having someone else narrate it, I always put ALL the changes in script into the edit as scratch narration to be sure it works before I pay someone to narrate the finished product.

8. Once the final narration is in place and the picture is tweaked, we call this picture lock because this is the point at which it would go to the scorer and we guarantee that the picture will not change.

While the scorer is scoring, I work on the sound. Sound work with underwater material can be quite complicated depending on how much work you want to put into it and how good you want it to sound. For some, the audio from the camera of the cameraman breathing and some music is fine. For higher production value, there will be an underwater "bed" sound, plus synchronous sounds of the divers breathing, the fish might "swish" as they swim by, etc. This is all tedious foley work that takes time...perfect to do while you wait for the score.

In the case where I am using royalty-free or other pre-scored music, I will do the audio work first, then deal with the music. I try to cut the music to the picture as much as possible. (This is a skill which requires a degree of musical knowledge, patience and practice where you carefully cut and loop pre-recorded musical tracks to alter their length so they fit your edit). Music can only be cut to a degree before it sounds wrong, so sometimes once the music is close to the right length, you then tweak the picture again to match it up with the music. Done right, most people will think the music was scored to picture.

9. The last step is to watch the whole thing though to tweak the audio levels between narration, sound, and music and to catch any last minute things that don't look right, such as bad edits, or dissolves that are too long, or something. The final mix is very important. Music must duck under the narration, and all the sound must render the narration clear, but you can't lose the subtle work you did in the synch sound foley stuff. Mixing takes a while!

10. Output/Master to format of your choice (tape, DVD, web, etc.). This can also be vary time consuming. Rendering mpeg files for the web or for DVD, or down-conversion to standard def is quite computationally intense and can take a long time, depending on the length of your project and the speed of your computer. This is a good time to have a fast machine. Output to tape can often still involve a render if you are dealing with HDV which must be conformed (another word for rendered). Output to BetacamSP, HDCam or some other professional format will require dedicated hardware with additional setup that can be frustrating at best!

The whole editing process, from tape capture to finished segment on an 8 minute Blue World segment typically takes me about 5 working days.

Jonathan

[Jonathan Bird]

Since I do things backwards by editing the video to the music, the selection of an audio track is an important consideration in my compositions.

David, although it may be backwards to a guy like me who does more documentary-style production, editing to music is a very common practice in music videos, "musical interlude" sections of films and even some commercials.

Another "backwards" style is sometimes done in documentary production where the story is more driven by the pictures than the narration. Sometimes if I have a segment that has very powerful footage that tells its own story without much need for explaining things behind what is going on, I edit the footage first, then write the script around the result. It often results in a much better looking piece, with less redundant narration. As a bonus, the script practically writes itself once the edit is done, and this kind of editing often goes together quickly because you are just stringing together the logical progression of good shots, then describing what is happening. I edited the Yap Feeding Frenzy segment for Season 2 of Blue World in this way. It was not tightly scripted before the shoot. I had lots of footage from multiple camera angles. So I edited it the thing together with the strongest footage in a logical order and voila! It practically doesn't even need narration. Not all segments are like this though.

My film Endangered Mermaids (http://www.oceanicresearch.org/television/manatee.html) was edited (by Art Cohen) almost entirely this way.

Jonathan

[Jonathan Bird]

I completely left out color correction.

For obvious reasons, I do as little color correction as possible as it is lossy to the picture quality (though nowhere near as lossy as you might imagine.) I tend to tweak color with the 3 way color tool in FCP. I have never learned to use the new Apple Color app. Someday!

Most of my color correction is to correct for bad white balance, improve contrast a little or, most commonly, to make the white balance between two different cameras match better. One of the biggest problems in doing two camera underwater shoots is that the two cameras rarely get the same exact white balance--much harder to do underwater than on land. So the picture is different and when you cut them together, one camera looks redder or bluer than the other one.

Jonathan

[sorvju-f]

Jonathan

Some questions:

Once the tape is logged and the project saved, then I do a batch capture on that tape.

I have noticed that when the time rolling changes in tape FCP can't regognize which period of time you want to capture...means that the shots can be totaaly different that you want. I don't know what are the princibles camera is starting new time rolling from zero.

I do not have time for the media management

Do you have some library if you want to use some earlier filmed shots?

Output to tape can often still involve a render if you are dealing with HDV which must be conformed (another word for rendered)

I have tried several time to record HDV back to tape...no luck so far ( both FCP and Premiere ). It starts, but taking only start of the shots and then frozening. It might be that with lap top file sizes are simply too big to do it. Maybe RAID disks and super processors?

In the case where I am using royalty-free or other pre-scored music

What are the best places to get royalty free music?

or dissolves that are too long

I have heard that pros are trying to edit and cut so that transfers are not needed...are you using them regularly?

Blue World segment typically takes me about 5 working days.

:D Pro do it so...mine takes 5 months up to five years:D

It practically doesn't even need narration.

I feel that any voice is important to have with shots...are you using music if not narration?

Most of my color correction is to correct for bad white balance

What is the best way to adjust WB in FCP...Premiere had own tool for that?

Thanks for very detailed answer

Jukka

[Jonathan Bird]

I have noticed that when the time rolling changes in tape FCP can't regognize which period of time you want to capture...means that the shots can be totaaly different that you want. I don't know what are the princibles camera is starting new time rolling from zero.

If you allow a blank section of tape between time code, the camera will reset (well, consumer cameras will...). So you must be sure that if you remove the tape or play it back, you return it to the end of the last scene when you start recording again, so it will pick up where the last time code left off. This is really important because as you said, if you have redundant time code on the same tape, batch capture may grab the wrong sequence.

Do you have some library if you want to use some earlier filmed shots?

I name the clips logically and leave them on the hard drive. I basically capture all the good stuff and never delete it, so it's always available on the hard drives in the future. Good for stock video requests.


I have tried several time to record HDV back to tape...no luck so far ( both FCP and Premiere ). It starts, but taking only start of the shots and then frozening. It might be that with lap top file sizes are simply too big to do it. Maybe RAID disks and super processors?

I have never had an issue with playing back and recording HDV to tape. It has to "conform" (i.e. render) and then it goes out just fine. You are capturing to an EXTERNAL drive, right? You never want to put media on the internal drive--it will drop frames and bail out of the process.


What are the best places to get royalty free music?

There are a ton of them on the internet and they allow you to listen before you buy. You get a deal if you buy more from them. I have used nearly a dozen, like music2hues, Davinport, royaltyfreemusic.com and others.

I have heard that pros are trying to edit and cut so that transfers are not needed...are you using them regularly?

Good editing is always pleasing to the eye and doesn't call attention to itself. Dissolves have a very specific use (they are used to show a change in time or place, never to "mush" two scenes together that don't cut well). In a good edit, 98% of the transitions will be straight cuts. There are obviously some times when you break the rules, such as video montages with flowing music or stuff like that. But, in general, dissolves are to be used only for specific purposes. And they should be short. I typically use 20 frame dissolves (2/3 of a second in 60 field/sec video, which would be equivalent to about 16 frames in 50 field video). Dissolves should never be longer than a second unless you are intending to confuse your viewer.


feel that any voice is important to have with shots...are you using music if not narration?

Yes, when the narration goes away, it is always nice to have the music come up or into the scene, but not always necessary if the sequence in question is extremely interesting and holds it's own with good nat sound.

What is the best way to adjust WB in FCP...Premiere had own tool for that?[/QUOTE]

FCP doesn't have a WB tool per se, but learn to use the 3 way color adjust--it's extremely powerful.

Jonathan

[sorvju-f]

Thanks for very detailed answer!

Jukka

[Red Sea Pictures]

If you allow a blank section of tape between time code, the camera will reset (well, consumer cameras will...). So you must be sure that if you remove the tape or play it back, you return it to the end of the last scene when you start recording again, so it will pick up where the last time code left off. This is really important because as you said, if you have redundant time code on the same tape, batch capture may grab the wrong sequence.

Another technique is to 'stripe' the tape. Record the entire tape with the lens cap on, continuous time code will then be permanently embedded along the entire length of the tape.

[Jonathan Bird]

Absolutely true, but it does have two disadvantages:

1. It is a pain and takes an hour per tape. "Blacking" 20 tapes before a dive trip would be a pain.
2. It puts wear on the tape before you have even used it once, so you have a higher potential for dropouts.