Multimedia
Newsletters
Blogs
Whitepapers
Events
DCP Directory
Mill Directory
Edit Calendar
Advertisers
Reader SurveyHDV editing using Avid’s Liquid 7.1, Part 5
Sep 7, 2006 4:20 PM, Steve Mullen
In the previous installment, we examined one of the fundamental debates in working with HDV: should we edit it natively or by using an intermediate codec? After covering this topic, we now almost fully understand how a timeline can hold multiple formats. However, we have not considered one issue. How can multiple aspect ratios be used in the same timeline? Let’s make this an even harder question. How can multiple aspect ratios automatically be used in the same timeline?
In fact, let’s push this to question to the extreme. How can multiple aspect ratios automatically be used in multiple timelines that have different aspect ratios? For example, I have two HD sequences on my PC. One sequence was created as HDV 720p30, and the other sequence was created as HDV 1080i60. The 720p30 sequence contains 720p30, 720p24 (yes, 24p), 1080i60, 16:9 DV, 4:3 DV, and 4:3 MPEG-1 shot with a cell phone. The 1080i60 sequence begins with multiple Sony Z1 clips followed by the entire 720p sequence. I simply dragged the 720p sequence from Liquid’s Sequence Rack (Bin) into the 1080i sequence.
All these, of course, are 16:9 sequences. I also have two 4:3 DV sequences – one that will become an anamorphic DVD and another that will become a letterboxed DV tape. And, while I am not fond of converting 16:9 to pan-and-scan SD video, later in this edition we'll learn how to this, too.
The first step, of course, is to capture all of our source materials — whether 16:9 or 4:3. When all of our clips have been captured, we will perform a magic step. A step that must be done before clips are placed into a sequence. However, I’m not going to reveal the secret yet.
The next step is to create the sequences we need. For each sequence, we can choose from a list of DV and HDV presets: NTSC 4:3, NTSC 16:9, HDV 720p29.97, and HDV 1080i59.94. I could also select 720p23.98 from the Region 60 presets and then select the M2V codec to create a 720p24 HDV sequence. In either case, the resolution for the Sequence will now be defined.
When we define a new sequence, we also define the image aspect ratio and pixel aspect ratio. The pixel aspect ratio will be used when conversions are made between square and non-square pixel video.
My 720p and 1080i sequences have square-pixel and non-square-pixel aspect ratios, respectively. My NTSC 4:3 DV and NTSC 16:9 DV sequences, however, both have a resolution of 720x480 and a non-square-pixel aspect ratio. (The DV format uses a WIDE flag to indicate 16:9 anamorphic video.)
As each clip is placed into a sequence, we would like the NLE to automatically handle it correctly. When the clips have the same pixel aspect ratio, the issue is simply how to match resolutions. When, however, the clips have different pixel aspect ratios, the issue is more complex. Pixel aspect ratios and resolutions for both clip and sequence must both be considered. However, we do not have to do the “considering,” as Liquid will do it for us.
When all your clips have been captured and graphics created, select all of them. Now, right-click and select Properties… and then click on the V-tab. You will have access to two pull-down menus as shown below.
As shown below, by leaving the selection at “Use original” we the use the information from the source clip.
Now, make the selections shown below.
Once this has been done, the following is automatically done:
1. 16:9 to/from 16:9 — uniform scale based upon “X”
2. 16:9 to 4:3 — uniform scale based upon “Y” yielding an anamorphic image
3. 16:9 to 4:3 WIDE — uniform scale based upon “X” yielding a letterboxed image
4. 4:3 to 16:9 — uniform scale based upon “Y” yielding a horizontally centered image.
When we want to place HD clips into a 4:3 sequence and we do not want either an anamorphic or a letterboxed image, we must take a different tack. After all your clips have been captured and graphics created, select all of them. Now, right-click and select Properties… and then click on the V-tab. Next make these selections.
When these clips are used in HD sequences, the following is automatically done:
1. 16:9 to/from 16:9 — uniform scale based upon “Y”
2. 4:3 to 16:9 — uniform scale based upon “Y” yielding a horizontally centered image.
When an HD sequence is moved into a 4:3 sequence, a “center-cut” 4:3 image is created automatically.
To implement Pan-and-Scan, first define all clips as “Fit X.” These can be used in an HD sequence. Then drag the HD sequence into a 4:3 sequence. Next, drag the 2D EDITOR GPU filter onto the clip(s). For 720p, set Horizontal Size to 200%. Now you can keyframe Horizontal Position between +50 (full left) and -50 (full right) to create pans over time—as shown below.
Now that we understand how to work with multi-format timelines, there are two additional topics to consider. First, while stereo audio has been acceptable for standard-def DVDs, as the world moves to high-definition DVDs in 2007, stereo is not going to cut it. Productions must have 5.1 surround sound. Liquid provides a near complete 5.1 audio creation capability.
Second, exporting MPEG-2 is a very time-intensive process as we all know. Liquid provides “smart GOP splicing” technology. Can this save export time? To answer that, we need to understand how this technology works.
Both topics will be covered in upcoming issues of HDV@Work. While this series is not a review of Liquid, it does seem appropriate to make a few comments about my experience editing with Liquid. I have been reviewing NLEs since the days of SuperMac's DigitalFilm, the VideoVision Studio from Radius, and Premiere 3.0. From me, Liquid earns a “my favorite NLE, but …” rating.
The “but” stems from: (1) the lack of keyframing in the main color corrector; (2) the lack of up-to-date, complete documentation (for example, the keyframeable color corrector is not documented); (3) audio editing problems; (4) lack of true 16:9 DVD support; (5) and too many design inconsistencies/bugs. Unfortunately, these kinds of issues involve the type of investment that companies avoid because they do not add sizzle that can be marketed. Nevertheless, in my opinion, the only new function Liquid requires, beyond P2 DVCPRO HD support, is integrated support for burning high-definition DVDs. (However, in a future installment I’ll show you how the latter limitation can be overcome for under $50.)
My “best” rating comes from: (1) Liquid’s "any-in/any-out" capability; (2) integration of almost every production task into a single application; (3) Liquid’s reliability when editing video; (4) display presets that work as well with a laptop as with multiple displays; and (5) an interface that is a true joy to use. The heart of this interface is an extensive, customizable set of toolbars that enables Liquid to function almost “mode-free.”
What are modes and why do I, and others, consider them an invention of the devil? First, it is important to distinguish between two kinds of modes. When you select a group of cells in a spreadsheet and enter Chart Mode, you are linking data to code that uses the data to perform an entirely different function. This is like Trim Mode in an NLE.
The type of mode with which I am concerned made its appearance with the first primitive text processors. These applications had no GUI. Everything was controlled from a keyboard. To insert characters into existing text, you first pressed the INS key to turn-on Insert Mode. (If you have a PC, you likely still have this key on your keyboard.) Once you entered Insert Mode, the keyboard continued to insert text.
When you wanted to replace existing characters, you pressed the INS key again to enter Overwrite Mode. Once in Overwrite Mode, the keyboard continued to overwrite text. Thus, whenever a fast touch typist became distracted after making a correction, they could wipe out vast amounts of text before looking up and seeing the OVR indicator was on! (Just for fun, next time you are using Microsoft Word, press the INS key and/or click the OVR button in Word’s status bar.)
The first Avid was designed in the DOS days and not surprisingly has Extract/Splice-in and Lift/Overwrite modes. And while Apple boasts that FCP does not have a “modal timeline," it does use a modal toolset. (Premiere likewise has a model toolset.) Simply put, a modal interface is one where what the computer does upon your pressing a key or clicking a button depends on what mode the interface is in. It's up to you to remain aware of the mode.
A non-modal interface does only what you directly command. You need only decide what you want to do and press a key or click a button to accomplish the task. You need not consider what mode is currently active. For example, the more modern Edius 4 interface has toolbar buttons for ripple and non-ripple deletes and ripple and non-ripple inserts. Interestingly, since PC keyboards offer both del and backspace keys, Liquid’s programmers missed the opportunity to support a mode-free non-ripple (del) and ripple (backspace) delete command.
Update: A luma (Y) signal ranges from 16 to 235. The Cb and Cr components range from -112 to +112, each of which is offset by 128, yielding a range of 16 to 240. This explains why Avid supplied both “235” and “240” values to me in describing its “2vuy” 4:2:2 uncompressed codec.
