HDV vs HD

HDV vs HD: A Primer
Article Focus:Emmy(r) Award winner, Tim Kolb discusses HDV and its place in the greater scope of HD production formats. Since the inception of the HDV format, we have seen many responces regarding the HDV format and how it fits alongside HDCam, DVCProHD, XDCam and other formats in the greater scope of the HD picture. In this article (which is taken from an HDV Format forum responce), Creative Cow's Tim Kolb gives what we feel is an exemplary job of dissecting the subject and clarifying the many components of the HD universe, HDV being one of those parts....There are a lot of interesting comparisons among various HD formats and the winner in each HD "category" - as well as the degree of how much better any one tape format or workflow is over another - is open to interpretation based on each user's experiences at this point. I'll do my best to throw out some general info without "coloring" it too much with my opinion...
OVERALL HD FORMATS: "HD" is a general term and it covers all the Hi Def formats including HDV really. But most people, at least most working professionals in film or broadcast, when they are referring to HD are talking about material acquired on HD Vision, HDcam, DVCProHD, Viper, or HDcam SR (or acquired on film and converted to one of those formats) using a picture size of 1920x1080 or 1280x720 and color sampling of at least 4:2:2 (HDcam is 3:1:1, but that's the exception).
COMPETING HDV FORMATS: HDV refers to a tape format that was really designed to be the replacement for DV, but for once the industry replaced a format with another format that uses the same tapes...hurray! On a DV videocassette, you can get two types of HDV footage: 1440x1080 frame size (as opposed to HD formats with 1920x1080), which has the same data rate as DV (25 megabits per second) but uses MPEG compression to create 15 frame (NTSC successor); or another popular HDV format uses 12 frame (PAL successor) groups of pictures or "GOPs." The basic compromise here is image quality as the image compression is very aggressive, a necessity to fit that much picture on a DV cassette.JVC's HDV uses a 1280x720 image size which reduces the data rate to 19 megabits per second and reduces the GOP to 6 frames. Progressive scan on a JVC camera also makes a difference in an increase in perceived sharpness...but on the flip side progressive scan can increase motion artifacts from camera or subject motion, particularly horizontal.
Since HDV is Transport Stream MPEG (and not studio profile) its color sampling specification is 4:2:0, as a DVD or satellite signal would have. This is effectively equal to the color undersampling that DV does with the 4 pixel blocks shaped as 2x2 in 4:2:0, vs 4x1 pixel blocks in DV's 4:1:1.
Both types of HDV have an 8 bit color depth, which indicates 256 possible values for each color channel.
DISSECTING SOME HD / HDV FORMAT SPECS:With the other HD formats mentioned, each frame is handled individually instead of compressed as a group. This increases quality, but there are some trade-offs in this group as well.
HDcam actually undersamples color to 3:1:1 (remember plain old DV is 4:1:1...3:1:1 is a bit better) and even though it records 1920x1080 images, it's really only laying 1440x1080 to tape to make the data fit.
DVCProHD uses 4:2:2 color sampling which is certainly better than 3:1:1, but Panasonic also needs to record a reduced res image to cram all the data on tape. The 1280x720 that the Varicam DVCProHD camera does so well is actually only 960x720 on tape.
HDcam and DVCProHD are both 8 bit color depth formats.
The Viper and HDcam SR both use far less compression, but require dual-link SDI to pump the video signal to a deck or a server and the equipment is currently not terribly cost effective for many projects that don't require impeccable quality and which have been budgeted accordingly.
In general, because of the nature of HDV, all the cameras that are currently available do not have interchangeable lenses and are just generally less expensive than the more "production-oriented" HD tape formats (for lack of a better term). The configuration of the camera can have as much or more to do with image quality than the compression and, generally, HDV cameras are configured to be inexpensive to fit their intended market.
JVC has two HDV cameras on their way which will have interchangeable lenses and will have what many of us old dogs in the industry might consider more professional "configuration attributes" based on what we're used to working with - but for price points that would range from slightly to substantially higher than we see in the existing HDV camcorder landscape.
HD / HDV POST PRODUCTION:The issue gets more confusing when post production is introduced as you can take your HDV footage and edit it "native" as MPEG or you can convert it into more conventional HD file-types and edit it with other types of HD content - or you can easily output it to other HD formats. Since HDV primarily uses FireWire as its main "pipeline" and other types of HD use HD-SDI, converting the footage can be an advantage in those situations. Some users would prefer to edit their HDV footage as HDV - it's been an on-going discussion as to the advantages and disadvantages of each approach.
PROS & CONS OF "NATIVE" MPEG EDITING:
CON: MPEG as a codec wasn't really designed to be edited as the long "GOPs" require re-compression if they are cut somewhere within the group (which happens more often than not, of course - particularly with the 12 and 15 frame HDV).
PRO: Editing native assures that changing to some other format doesn't introduce any conversion artifacts into the pictures.
CON: MPEG takes more processor power to compress than to decompress, as it was designed as a distribution format. CoDecs (COmpressor/DECompressors) designed for editing tend to be "symmetrical" - or have a much more "even" processor load between compression and decompression. HDV native editing tends to be slower at previewing effects and transitions for this reason.
PRO: High quality "symmetrical" editing codecs tend to create larger files than the HDV MPEG files.
CON: Most of the available "intermediate" codecs (which you would use to convert HDV to for editing if you didn't use MPEG) tend to be of much higher quality over repeated generations of compression. This does not mean that your footage will improve so much as it has the potential to degrade far less through the editing/post production process. Some higher-end HD editing systems have 10 bit color codecs that increases the color depth to 1,024 possible values per channel. For extensive color correction or effects work, this 4X deeper palette can hold quite an advantage.
Also because of the processor loading, it can take a bigger system to edit MPEG with the same effectiveness as an intermediate codec, even though the intermediate files will almost certainly be larger.
PRO: With intermediate codec systems, output to FireWire for mastering back to the HDV camera or deck often entails additional steps.
CON: Many intermediate systems can output to virtually any HD tape format if you plan on mastering to something other than HDV.
There are also "this application vs that one" arguments on this as well, which I've left out of the discussion - and based on the software you use, these topics are searchable using the Cow's search engine.
This is just a scratch at the surface of the basic qualities of some HD and HDV formats...development won't stop anytime soon and new technology will continue to increase in quality and cost-effectiveness, and the lines will almost certainly become even more blurry.
...whoever said "You stop learning when you die.", may have actually worked in our business.
Note:Editing native or with an intermediate codec is actually an ongoing debate within the confines of the "other" HD formats as well. Apple uses the DVCProHD codec internally to handle Panasonic DVCProHD material allowing use of FireWire for I/O and keeping transcoding steps out of the process. The DVCProHD codec makes relatively small files so the speed of editing even on a software-based Mac can be rather responsive. HDcam and DVCProHD native editing are both available on various systems. Those who favor using an intermediate codec point to the fact that the DVCProHD codec is limited to 8 bit color and every time the codec re-processes the footage (every transition, color correction, effect, etc. results in the codec "re-compressing" the footage again), the material goes through the 1280>downsampled to 960>upsampled to 1280 for playback-process. The same type of processing happens with the 1920>1440>1920 processing of HDcam, with the addition of the limitation of 3:1:1 color sampling.
-->