The original title of this post was “HDV Sucks.” Now, I’ve changed the title to something a little less harsh, but have no doubt, this is a contrarian post.

Isn’t it amazing that the same little tapes that hold an hour of standard definition DV footage can also hold an hour of HDV footage. It seems like a miracle. It seems too good to be true. We’ll explain how this works.
First, let’s keep the definitions straight. Standard defintion (SD) and high definition (HD) refer to the resolution of the frame, whereas DV and HDV refer to recording formats (or how the video is stored) on tape. There are several flavors of HD and several recording formats for both SD and HD. This article is only concerned with HDV and DV the most popular formats for consumer and prosumer video cameras. Both formats record to miniDv tapes. Though some HDV cameras require special hi-fidelity tapes, they will work either way.

The difference between standard defintion (SD) and high definition (HD) is a matter of pixels. HD video has more of them. We generally use the word “resolution” to describe the number of pixels in an image. For example, the resolution of a frame of SD video is 720×480 and the resolution of a frame of HD video is 1440×1080. You may have seen the term “1080i” thrown around. This refers to the 1080 vertical lines of resolution in the image with interlaced frames.

So, there is a lot more information in a frame of HD than there is in a frame of SD. Yet, somehow they require the same amount of tape. The data rates for both HDv and Dv are around 25 Mbit/sec (or 3.5MB/sec). To put the data rate in context, here are some other data rates worth knowing.

Cables:

• USB 1.0 - 12 Mbit/sec
• Firewire - 400 Mbit/sec
• USB 2.0 - 480 Mbit/sec
• Firewire800 - 800 Mbit/sec

Video Formats:

• DVD video (MPEG-2) - up to 9.8Mbit/sec (average is closer 6.5)
• Dv and DVCPro25 - 25 Mbit/sec
• HDv - 25 Mbit/sec
• DVCPro50 - 50 Mbit/sec
• DVCProHD - 100 Mbit/sec

Compression

The key to getting more pixels into the same amount of space is to use more compression. If you are concerned with image quality, it is important to learn the basics of how compression works in order to understand how it affects the recorded image.  There is a simple way to understand compression.  The encoding of the video is a description of the images that the player needs to put it on the screen.  When the video is compressed, the description is shortened by abbreviating and simplifying repeated sections and small details.  The data rates listed above can be thought of as the length of the description.

As video becomes more and more compressed, you will start to see more and more compression artifacts.  These are the boxy inconsistencies, you see on poorly encoded youtube videos.  Sadly, DVDs that try to pack a lot of extra features onto a single disc, often pay the price by cranking up the compression, sacrificing image quality.  These will often show up most notably in the dark areas of the frame.

MPEG compression, the kind used in DVD and HDV (but not miniDv), compressed the video in three dimensions.  That is, rather than just taking each 2D frame and compressing it by replacing of similar color by a simplified version, MPEG also, compressed in time.  If some area of the frame does not change from one frame to the next, it doesn’t need to described twice in the datastream.  This means, that MPEG compression artifacts can sometimes be more jarring as they jump in time.  A part of the frame may stay unchanged during a slow pan over a detailed area and then jump to catch up.  I have shot a fair amount of HDV and have not found this to be terrible, but it’s not great.

Why capture breaks in HDV

Now that we know something about the compression, we can discuss why there are so many difficulties capturing HDV.  If you have tried to capture HDV to Final Cut Pro, you may have noticed that often, FCP will choke and claim that there is no time code.  Or, the capture will take  a couple tries and produce several extraneous video files.  Recall that HDV is compressed in time as well.  That means, some frames are pieced together from previous frames.  Because of this, it can be hard for the capture to device to be clear about what is in a frame if you start at a random spot.  This is not really an excuse for the software and hardware developers but it at least gives something of a reason.  I have found that with both Sony HDV decks and Sony HDV cameras, the tape will cue up, then drop time code for an instant before sending data over the firewire cable.  This little hiccup immediately causes FCP to throw up a warning and abort.

If anyone know a good workaround for this, leave a comment below.