Digital Video Content Creation Guide

Last Update : October 7, 2003


00. Define Source & Target

01 Get the analog video into digital form

02 Transfer the digital form to the computer

02.1 analog -> DV-AVI or MPEG file
02.2 miniDV -> Firewire -> DV file (AVI format)
02.3 DVD+/-R(W)|RAM -> MPEG file (VOB format)

03 Clean up

03.01 video edge noise
03.02 interlace
03.03 inverse telecine
03.04 black level
03.05 trim
03.06 audio

04 Encoding

04.01 real
04.02 MPEG-1/2
04.03 QuickTime .mov, MPEG-4 .mp4
04.04 Windows Media 9 .wmv

04.04.01 MPEG-2 -> WMV
04.04.02 DV -> WMV

05 Delivery

05.01 streaming .rm
05.02 streaming MPEG
05.03 download .rm, .mpg, .mov, .mwv
05.04 Access control

06 Bandwidth

06.01 GigE
06.02 100Mbps
06.03 802.11a
06.04 802.11g


00. Define Source & Target

There are many ways to deliver the digital video. Think about some of the questions below and use the diagram to decide what route is best for you.

There are many advantages/disadvantages of each method. The flowchart use special consideration criteria that is not changable to narrow down the choice for you quickly. You can find comparison studies and marketing info on vendor's websites if you want to know more.

If you have other restrictions, you may have to sacrific some factors and use an alternative.

Questions about Source :

miniDV? VHS tape (not copyprotected)? Laser Disc? DVD (not copy protected)?

the author gave me the file - what's the extension .mov, .rm, .mpg, .wmv?

Self-created? Public Domain? Got permission from copyright owner(s) already?

[ If it is a "non standard" format - PAL/SECAM instead of U.S. NTSC? 8mm/16mm film? consult your support people ]

Questions about Target : who? how?

Just students on campus? Off campus? All got broadband? Some only have modem access? General public and others in your discipline may be interested?

A comparison of different bandwidth is the end of this article.

click on thumbnail below get a popup of full size image


Some old documents that you may be interested also, they are outdated, but some background information probably still applies


01 Get the analog video into digital form

Unless you already got the video file from the author or has it on digital tape (miniDV, or Digital 8, microDV, or DVD usually, it need to go from analog (e.g. VHS) to digital first.

[ If some people give you full size DV, DVCAM, DVCPro, DVCPro50, Digital S, or DVHS tape instead, consult your support person for those ]

You have several choices.

Hook up the VCR to

  1. Digital camcorder with analog in
  2. Digital capture box/card with analog in (such as Pinnacle/Dazzle, Canopus, ADS)
  3. Digital VCR with analog in (like SONY GDV-xxx, DSR-xxxx, JVC HR or SR-xxxxx)
  4. DVD+/-RW|RAM VR deck with analog in (e.g. Philips DVDR-75/77/80/85/985, Pionner, Panasonic, or Toshiba)

It doesn't matter which way to do it, you end up with either a miniDV tape, DVD+/-R(W) |RAM disc or directly becomes a DV or MPEG-1/2 file on your computer.

You just push PLAY on the VCR, and RECORD on the recording device.

There are some minor catches on the quality settings (SP or LP, 32 vs 48 bit audio, 384kbp or 256 kbps audio), use the highest setting when possible.

02 transfer the digital form to the computer

If you are not using one of the method to go from analog directly to a computer file, you have to get from digital tape/disc on your hard disk somehow.

Several possibilities again

  1. use IEEE-1394/Firewire/iLink connection on your computer (like the one in R40) to capture the miniDV tape's content.
  2. put the DVD+/-R(W)|RAM disc into your computer with the appropriate kind of DVD drive.
  3. use a digital recorder drive like FireStore, DataVideo, DSR-DU1, Videonics to transfer from miniDV to hard drive directly

02.1 analog -> DV-AVI or MPEG file

If you go straight from analog to MPEG2 on DVD+/-R(W)|RAM or hard disk, e.g. MyDVD 4.5.2 or VideoStudio 6/7, you don't need to transfer file.

VideoStudio 7 direct to MPEG capture plugin

Micrsoft Movie Maker 2.0 Capture task

MyDVD 4.5.2 record to disk


Windows Media 9 Capture Utility

02.2 miniDV -> Firewire -> DV file (AVI format)

This is the most common and most flexible way.

You do this using one of the many video editing program, e.g.

  1. Adobe Premiere 6.5, Pro aka 7.0
  2. Ulead Video Studio 4.x, 5.x, 6.x, 7.x
  3. Microsoft Movie Maker 2.x
  4. Pinnacle Studio 7.x, 8.x
  5. Roxio (MGI) VideoWave 4.x, 5.x
  6. others like AVID Xpress DV 3.5, Apple iMovie/FCP 3.x/4.x/Express), Sonic Foundry/SONY Vegas Video 3.x/4.x Pinnacle Edition 4.x/5.x, Canopus Edius 1.x, Ulead MSP 5.x/6.x/7.x, various Discreet/Media 100 tools.

Basically, you push record and you end up with a AVI file with your DV content.

Here are the catches

  1. drop frame - if your computer have other things going on during the capture, it may not be fast enough to capture all 30 frames every seconds. Make sure you check after capture whether there is drop frames.

    Video Studio 7 (see its FAQ (popup), no screen estate left for this) and Microsoft Movie Maker do not report that, just watch the video, it is obvious if there are missing frames.

    Here are the screeshots of the drop frame counter in Video Studio 5 and Premiere 6.5

  2. older programs doesn't support the OpenDML extension, you are limit to 2GB per file (~ 11 min). Most new program doesn't have this problem.
  3. more than one AVI-DV format, there are Type 1 or Type 2 files. It is never a problem with in the program you capture the file, it is usally cause hiccup when you try to use the file in another program. If that happen, just choose the other type.

    Here's a screeshot of the new project page in Video Studio 5. You can set the Type of DV file it capture there.
  4. Field order - it may be labelled as A/B, lower/upper, odd/even. Usually the program set it correctly in its preferences.
  5. Remember to rewind past where you want to start recording. There is a few seconds of lag time before you hit record on the computer before it actually starts in some programs (without pre-roll). So don't use the first 10 seconds of a tape, just record blank on that. (don't leave that "blank blank" otherwise it missed timecode)

02.3 DVD+/-R(W)|RAM -> MPEG file (VOB format)

There are two ways to do this.

  1. Use a DVD extraction program to "rip" the files to the hard drive.

    Make sure you have cleared the copyright with the owner first.
    And there is no copy protection on the disc.

    Consult your legal department and your service provider if either of the circled box said yes or something like CSS/CPPM and/or no Region 1 listed.

  2. Drag all the .VOB files from the VideoTS folder to the computer and rename them .MPG. (To do: They will play just fine. Not sure whether it works ok for feeding into other program, test)

The video will be in 1GB trunks, you have to combine them back into one big file.

  1. Drop to DOS and do
    copy [/b] 1.vob+2.vob 12.mpg
    (To do: only tested on .vob ripped from DVD+RW, need to test on other MPEGs)
  2. Use a MPEG tool like tmpgenc to merge th files (To do: works on regular MPEG-2. test VOB off DVD+RW)
  3. Use a video editor like Adobe Premiere to combine the pieces (To do: test VOB)

03 clean up

Now you get either .avi or .mpg file. If you are lucky, you are ready to go. However, most of the time, the video need to be "cleaned". You can do it now or during encoding, depends on which software have the options you like the most.



03.01 video edge noise

source from bad VHS deck, satellite and the like usually have problem with noise at the edges that are very annoying to the viewers. They don't show up in TV because of the overscan (10-20% are cut off)


Note: the marked area in middle are due to interlace/telecone instead to be discussed in teh next section


There are several ways to deal with the edge noise

  1. crop them out, but that change the size.
  2. crop them out, and then stretch back to original size (720x480) (ToDo: Scott said it works in Media Encoder too, test)
  3. use masking function to cover those with black

    e.g. crop and stretch in Premiere 6.5

    in Windows Media Encoder 9

    here is an example of using tmpgenc to mask out the top 8 pixel to eliminate the "running ants", and 2 pixel from bottom for the "on/off moving step


03.02 interlace

Unless the original is a progressive source, deinterlace the video.

Pick the method that looks best. Becareful about some scheme double the data rate (ToDo: how to test the effect, they still plays fine)

Here are the choice in tmpgenc, available algorithms differ from program to program

03.03 inverse telecine

If your original is from film, most likely it is 24fps and have been 3:2 pull down to 30fps 60 fields. Now you have to inverse telecine it back to 24fps.

The UI in tmpgenc looks like this

Windows Media Encoder 9, use A,B,C,D naming convention, you will have look up some reference on how those works

03.04 black level

Because the way DV handle color, to optimize the video's black, sometimes you have to change the settings for processing if it is not automatically take into account.

The toggle for DV in tmpgenc is in Special setting.

In Helix, under video filter (To Do: check doc, I think that's the same thing)

03.05 trim

It is 25Mbps for DV or usually 4-6 Mbps for MPEG-2, trim the part you don't need..

The shortcut keys are handy here (I and O for Premiere, F3 and F4 for Video Studio)

It is harder to use just numbers like Media Encoder 9

It is best if the program you use have audio wavefront display, it is must easier to trim sometimes. low end consumer product like VideoStudio doesnt' have that feature (ToDo: what's the shortcut key, if any, to mark in/out in tmpgenc?)

03.06 audio

Some of the source's audio is low in volume. The ThinkPad speakers are not very powerful, try to maximize the audio. Best if there is auto normalization. If not, usually it is no more than 300-400% boast or +12 dB.

In tmpg, use change volume

In Helix, use audio gain; Premiere is audio mixer

(ToDo: check whether DirectX equalizer transform plugins work or not in wme)

04 encoding

Finally, we got a clean video we can use. You can then convert the .AVI or .MPG to a deliverable. (.MPG actually can go straight to IPTV if that's what you want (ToDo: make sure renamed/merged .VOB works) )

The common formats are

(ToDo: AOL licensed On2 VP6, the spec looks good, check license whether it will become open source like VP3)

04.01 real

It is best to use Helix Producer rather than the one internal to your video editor.

You have to decide how many different bandwidth to accomodate different users in real.

(ToDo: check wmenc and IPTV on how to set multistream and multi bitrate MPEG)

Don't forget to add clipinfo

and hit encode, it is about 4x realtime on R40.

04.02 MPEG-1/2

If you start with a DV-AVI file

You don't need this step if you capture directly to MPEG with DVD+/-R(W)|RAM, other hardware like Canopus Amber, or software like Video Studio 7.0, MyDVD 4.52


Adobe MPEG Encoder v1.1 main page

04.03 QuickTime .mov, MPEG-4 .mp4

If you start with a DV-AVI file

to do the conversion

04.04 Windows Media 9 .wmv

Convert from DV to WMV is eaiser than from MPEG-2 to WMV.

There are at least 3 ways to do the later.

04.04.01 MPEG-2 -> WMV

You need the "right" decoder installed.

Play your MPEG in Windows Media Player and check the property page.

Doesn't work
Video decoder: Ligos
Audio decoder: InterVideo (ToDo: 4.0, have not tested with 5.0 yet)

1st pass work, but choked on audio codec in 2nd pass
Video decoder: InterVideo
Audio decoder: InterVideo

Tested to work
Video decoder: DVD Express (ToDo: probably 5.0 something, do not have 6.x, try get a disc from Aptiva to try)
Audio decoder: CyberLink (ToDo: 3.0, got 4.0 somewhere, don't have 5.0)


04.04.02 DV -> WMV

Pretty much just go through the wizard.

you can change the details of this setting latter

Here's where you fine tone the compression

you have to decide whether to keep the rectangle pixel 720x480 or convert to square pixel 4:3 640x480

and DRM that you have to have your service provider to perform that.

remember it is a 2 pass process

05 delivery

05.01 streaming .rm

you need to do two things

1. put the .rm file in \\netmeeting\rtsp\departmenname

2. create a metafile .rm|.ram/.rpm (depends on whether it play outside/standalone or embedded in a webpage) or .smil (if it is a complex presentation) on \\acfiles\www-home that points to the file like


05.02 streaming MPEG

IPTV is the replacment for IBM VideoCharger and Concurrent MediaHawk for streaming MPEG-1 and MPEG-2 contents.

1. put the .mpg on \\iptvserv

2. configure the file for delivery via the webpage control panel

05.03 download .rm, .mpg, .mov, .mwv

you can put the files for download on

  1. public server like \\acfiles\www-home for web delivery
  2. an internal server like \\acad1 for internal delivery
  3. blackboard

05.04 Access control

  1. you control web delivery via the .nsconfig like all your other files
    for all academic users, it would be
    <Files *>
    RequireAuth dbm="/pub/ns-home/userdb/wfu" realm="WFU Academic Computing"
    < /Files>
    see (pop-up) for details
  2. internal NT/2000/Celerra server like \\acad1, use XP's permission
  3. consult your support personnel for blackboard control


06 Bandwidth

798MB test file

06.01 GigE

ftp ~ 42 seconds (19.56MB/s)

http ~ a minute something

06.02 100Mbps

ftp 3 min 23 sec, avg 32132 Kbps (4016.520 KBps)

smb 3 min 36 sec

http 19 min 25 sec (693.5KB/sec)

06.03 802.11a

ftp ~ 9 min, 1.x MB/s

http ~ 25 min, 600KB/s

06.04 802.11g

ftp 12 min 30 seconds, 1117.24Kbytes/sec

http 25 min, 550 KB/s