This phase involves everything after production including television system conversion, capture, editing, pre-processing, output, encoding, integrating metadata, interface design, creating the web pages where it will be published, and posting to streaming servers. This is where the fun begins... turning plain video into a quality interactive multimedia experience. No more worrying over pre-production oversights, no more sweating the production disasters. Now you get to be creative with graphic design and build a presentation that viewers will engage with and retain more, that is if you choose to embed your video in a Web page.

Acquire footage or digital files

Some of you will be capturing your own raw footage to edit yourself. You can then proceed to the Capture section. Others will be getting footage or an edited master from a customer or Post house. A Post house will likely be able to capture to an MPEG file so it's definitely worth asking for one to negate conversion and capture. If you're outputting to tape or DVD, request MPEG-2 (full size). If not, MPEG-1 (half size) will do for streaming files. Set up an FTP space somewhere to transfer these big files at lightning speed. Your personal web host usually provides FTP service. Otherwise put it on a network drive and transfer it with Windows Explorer. If the Post house or production company doesn't digitize, request a miniDV copy of the footage so you can capture with your DV camera via Firewire. If that's not possible, request a master copy in whatever format they use and have a local conversion / duplication company convert it to your tape format or CD.

Convert system format

If production and / or editing is done overseas, the television system format of the tape (NTSC, PAL, SECAM) may need to be converted to your system. For example, if you're in the US, getting a PAL BetaSP tape shot in Europe FedEx'd and having a video conversion / duplication house locally do the conversion to NTSC miniDV works well. Do it yourself and you can get it down to 5 days instead of 2-3 weeks. Asking the Post house to convert it to NTSC and transfer it to miniDV can take awhile as they'll probably have to send it out. Therefore, it's much faster if you can get the Post house to capture the files to MPEG-2 and post them to your FTP server, taking as little as an hour instead of 5 to possibly 21 days. Here's a list of countries and the television systems they use. 

Capture

• Use AV hard drives – Firewire or RAID
• Research your capture card based on your budget if not using Firewire – consumer, prosumer, pro?
• Conserve CPU power – disable all anti-virus programs, screen savers and other apps during capture process
• Optimize your hard drive – defrag before capture
• Capture at highest quality – NTSC DV 720x480 (0.9), 29.97 fps; PAL DV 720x576 (1.067), 25fps; both 16-bit audio
• Avoid dropped frames – test and improve your system until you reach a zero dropped frames standard
  • Use an external drive (not your C: drive!)
  • Backup projects to DVD+RW or tape, verify, then delete from external drive
  • Defrag often
  • No other applications loaded while capturing, and from a fresh boot
  • Or outsource capture to a video conversion / duplication house, especially for long-form videos

Acquire adequate workstation and peripherals

• External USB or RAID hard drive – 120-200Gb, 7200+ rpm
• DVD+RW or tape drive for backup
• RAM – 1Gb for each open application
• Firewire card (if not already built-in)
• A-to-D breakout box for analog sources (Canopus ADVC-110)
• 3-CCD DV camera (or cheaper miniDV camera to playback)
• TV monitor

You'll need to acquire an adequate workstation (and peripherals) with enough RAM and disk space to handle video capture and editing (see suggested workstation components below). A video-capture card is necessary to capture audio and video if Firewire isn't built-in. It's better to configure the workstation with Firewire to begin with. Otherwise, there are basically two types of capture cards: capture from an analog tape deck or camera, and capture from digital video decks or camcorders using FireWire (sometimes referred to as I-Link or Lynx). This interface is capable of data transfer rates up to 400 Mbps so it's perfect for handling the amount of data needed to transfer real-time video. Both analogue and digital cards capture video in real-time. However, an analog capture card will either degrade the capture via compression or will create raw .AVI files with enormous file sizes. Firewire is definitely the way to go, and comes built-in to most PCs now—certainly multimedia workstations—and negates a capture card altogether.

Dropped frames during capture are something you want to avoid. Removing dropped frames is tricky and can make a voiceover stutter if you edit out those frames. Optimize and test your system until you reach zero-dropped frames. They're usually a result of a configuration problem, trying to capture video that exceeds the capture rate for your system, or more likely a fragmented hard drive. Backup (then delete previous projects) and defrag your drive often. You stand a better chance at zero-dropped frames if you have contiguous disk space to write to.

TIP: Don't bother with USB 2.0. Firewire is the standard transfer protocol in the DV realm. Stick to Firewire peripherals such as the external hard drive and the DVD+RW drive.

TIP: Canopus breakout box is a great alternative to an analog capture card, and has an added bonus. The breakout box has Firewire in / out, analog audio and video in / out, and S-video in / out. You will probably get a VHS tape at some point to capture, maybe even a cassette tape. This box allows you to input analog sources yet output Firewire for capture. The bonus is, in your Premiere project file you can preview video on a TV monitor at full frame rate without hardware acceleration! Perfect for checking video levels going out to tape.

Editing

• Confirm appropriate pixel width is specified
• Skip the fancy stuff – stick to cuts and quick fades. Focus on content, not effects
• Use titling in an external graphic if possible, not in the video – think 160x120
• Output requirements – AVI, MPEG-1, MPEG-2, DVD, CD, tape?
  • determined in pre-production
  • crop on output vs. in the timeline unless there's multiple crop requirements in your source files
• Notate timecode in Advanced Script Indexer text file for markers / scripts
• Export a frame of video for the Audio Only video area and web page link
• Have someone transcribe the video for a downloadable transcript that will also provide text for closed captioning
• If importing graphics with Premiere, specify the default duration before importing
• Specify default transition / duration before inserting them

Editing activities include arranging various sections of captured clips into a timeline, inserting transitions, incorporating graphics or visual effects, applying pre-processing effects, mixing and synchronizing audio, and possibly adding titles. Bear in mind the limitations of the Web and avoid unnecessary motion that will make streams 'smudgy', namely fancy transitions... stick to cuts and quick fades.

Pre-processing

Pre-processing improves the quality and levels of source files. This is where you'll fix marginal audio and video. Look at the quality improvement in the adjacent clip. Hopefully the production crew did a good job with audio quality and video levels. Many times however, audio levels are too low and video contrast needs adjusted for PC delivery. Pre-processing prior to encoding also reduces the burden on the compressor, i.e., increasing black levels to eliminate grain.

Adjusting Color and Brightness

• Brightness – Controls the overall luminance value and can help dimly lit video.
• Contrast – Increases or decreases the range of difference between the darkest and lightest portions of the picture. The more 'washed out' it looks, the less contrast you have. Increasing it sharpens it up.
• Levels – Uses a valuable histogram to evaluate video levels such as black, white, and grey points. Very useful as the flat levels on the histogram show exactly where to position the black and white points—they should be adjusted to where the histogram starts and ends the waveform. Even if you're on a horribly calibrated monitor, you'll output proper video levels.

• Gamma Correction – Used to adjust video brightness levels for playback on Macs which look brighter than PCs. Also useful for a global brightness adjustment without applying effects to each clip (decreases processing time).
• Saturation – Helpful when you get bad color video, i.e., they didn't white balance. You can pull the saturation down, almost to black and white, so the bad color is not so overwhelming. It beats green video!

Video levels for tape / Web

Video is much darker on a PC than it is on tape / TV. And video on a Mac is brighter than on a PC. So, if you're outputting to tape as well as a stream, you'll want to pre-process them separately (create separate Premiere projects). Keep your MPEG-2 for tape unadjusted brightness-wise even though it looks dark. This is why it's a good idea to have a cheap TV monitor included in your system to check video going out to tape / TV. On your transfer file for streaming however, go ahead and brighten it up to where it looks acceptable on your monitor, but be sure to calibrate your monitor first.

Audio Pre-Processing

It's very helpful, most times necessary, to export your audio track(s) from your video editing software to a .WAV file and open it in a dedicated waveform editor (audio application) such as SoundForge. This allows you to see the waveform up close and determine levels quickly. Always check audio levels of your source footage and not just turn your speakers up (or down).

• Normalization – Maximizes the volume level without causing distortion if set properly, providing the best possible signal for encoding. Very useful on every project.
• Volume – The overall level of the signal. Certain portions of the audio may need to be adjusted so the volume of the presentation is consistent. Better done through normalization and / or compression.
• Compression – Evens out the dynamic range of the audio. Very useful for controlling level spikes and adding 'presence' or 'punch'.
• Equalization (EQ) – Various frequency ranges can be boost or cut in the same way that the treble and bass on a stereo are adjusted. It's always better to cut than boost as boosting adds distortion.
• Resampling – Files may need to be resampled to a specified sampling rate. This is sometimes necessary if you shoot DV in LP mode which may record your audio at 32k instead of the normal 48k. You may also have music tracks sampled at 44.1k. You should keep all audio tracks in your editing software the same sampling rate, so you may need to resample a few files up to 48k.
  
  Premiere will take care of multiple sampling rates if you select Best in the Enhanced Rate Conversion dropdown in Audio Settings.

Output

During pre-production (Output Requirements), hopefully you ascertained what output formats you'll be targeting. Output your edited video to:

• AVI source file for encoding / (CD if required)
• DV tape for a master tape archive if you have full size video (720x480)
• MPEG-2 for DVD (if required)

For the .AVI streaming file, output to the highest size you'll be encoding to, be it 320x240 or 240x180. Use the MPEG-4 v2 codec with maximum settings (the least amount of compression) so file sizes are manageable. Uncompressed isn't necessary and the file sizes would be huge, even for just a 10-min. video. Settings for the audio output should be 16-bit stereo at 48k.

Tip: At encoding, specify mono audio in profiles to conserve bandwidth. But keep this output file stereo for tape, DVD, or CD.

Getting the Proper Image Parameters

• Cropping – Most video files have black bands or tearing around the edges and need cropped at some point, whether it's done in the timeline or on output. This usually occurs during capture and PAL / SECAM conversions. When you crop in Premiere, you'll need to check the Scale To option to get it back up to your output size (240x180). Pay attention to the aspect ratio, either maintaining the original 4:3 (or 16:9) aspect ratio and / or keeping the dimensions to multiples of four for codec compatibility. For example, you could crop 2 pixels from the left and right, and 0 from the top and 3 from the bottom, which gets you the 4x3 aspect ratio. Sometimes that's not enough so crop 8 from the sides and 6 by the top and bottom. Otherwise you're stretching the video in either the x or y axis if the Scale To option is checked.
• Image Size – Digital video is typically captured at a larger frame size (720x480 NTSC, 720x576 PAL) than you are likely to output for streaming. For streaming, output to the highest size you'll be offering online with square pixels. For tape or DVD output to full size with the appropriate pixel width. For CD, output 640x480 with square pixels.
  
  You might get widescreen video at some point. You'll want to output to 320x180 with square pixels for a 16:9 stream version. Crop 40 pixels off each side to 240x180 for a 4:3 version. Leave the Scale To option unchecked. Sometimes you'll need to vary the amount on each side so the subject is centered as shown below.

Varying the left / right values to center the subject for a 4:3 version

• Pixel Aspect Ratio – This sets the size of the pixels. In your Premiere Project Settings during editing, match the pixel aspect ratio of your footage (NTSC 0.9; PAL 1.09) for DV video. When outputting for tape or DVD, retain these pixel aspect ratios. When outputting a 'transfer' file for streaming, use square pixels.

Pixel Aspect Ratio Options

• De-interlace – The U.S. television standard, National Television Standards Committee (NTSC) and the European standard, Phase Alternate Line (PAL) deliver video in an interlaced fashion. Each frame is broken down into two fields: odd lines and even lines (NTSC has 550). These lines are displayed separately, odd lines first then the even lines, each set taking about 1/60 of a second to display. This method is referred to as interlacing. The de-interlace function combines the odd and even fields or deletes one set to make progressive scan video. Progressive scan (used by computers, new DVD players and HDTVs) delivers entire frames sequentially. De-interlacing isn't necessary for video sized at or below a height of 240 because you can't see it, but it's very noticeable if you're outputting to tape, CD, or DVD. You can de-interlace at output or encoding. For more detailed information, see Television Standards.
• Inverse Telecine – Used to discard the additional 6 frames the 3:2 pulldown process adds to film-to-tape transfers since film is shot at 24 fps and NTSC video is 29.97 fps. Not necessary for PAL video (25 fps).

Calibrating your monitor

Notating marker timecodes

Before you close your editing software and get to encoding, notate the timecodes where you'll want to place markers for chapterization (jumping to different parts of the video) if you choose to implement them. You may have separate clips you've laid into your timeline and can simply zoom in to get those in / out points. Or, you may have a single hour-long clip where you'll need to log the video by watching the whole program and writing down the spots where you want to put markers. Write these timecode markers directly into this Advanced Script Indexer (ASI) text file for importing later. Save this file as "markers_scripts.txt" to your hard drive and don't change the formatting so ASI will import it correctly later. Keep a copy of this file for entering your markers in future projects. It's much easier than entering markers and scripts through ASI.

Transcribe

Once the edit is signed off and approved, it's a good idea to have someone transcribe the video to allow downloading from the site. It will also provide the text for the SAMI file used for closed captioning, as well as aid in adding markers and scripts. Translations to different languages should be transcribed and proofed if language use will be enabled in closed captioning.