underwater videographer course | equipment | video formats
Video Formats
The number and variety of formats recorded by video cameras can be really bewildering. This section discusses the video recording formats of most relevance to the underwater videographer. Be sure to read and absorb the previous sectionLINK about video concepts and terminology before reading this.
Analogue Video Formats
If this was written up until the mid 90's we would be talking about Video8 or Hi8 video cameras for use underwater. In fact it was only recently that the PADI Underwater Videographer course syllabus was still citing these as the primary formats for consumer use. Video8 and Hi8 were both analogue formats with a 4:3 aspect ratio and capable of surprisingly good video. Analogue means that the signal was recorded as an infinitely variable magnetic signal. If you copy an analogue signal repeatedly then the quality will gradually degrade as you may have seen from multiple-generation copies of VHS cassettes.
Our family owned a Canon Hi8 camera from the late 80's and it did a fine job of a few weddings and holidays, but never found its way underwater. But in 2005 a customer on the liveaboard I was working on brought along a Hi8 camera in an Amphibico housing and his video was the best I have ever seen from a hobbyist. It proved to me that it really isn't the gear you own that's important, but how you use it.
Pic of Hi8 camera
Although time is more or less up for these formats, there may still be some Video8 or Hi8 rigs kicking around on Ebay for a very cheap price. These might be just the thing for total beginners or those wanting to record a single event or on a very limited budget. But the vast majority of users these days should now be looking at digital formats.
Digital Video Formats
To cut a long story short, digital video is simply a signal made up of a series of "1"s and "0"s. This means that once the video signal is recorded it can be copied or transferred between formats without any loss of quality. And a digital video signal allows video to be easily processed on computer equipment in post-production.
Compression
Digital video is compressed to allow it to fit into a smaller bandwidth and storage space than it would if it was uncompressed. Uncompressed video is used sometimes in post-production but the files are seriously large.
A codec is used used for encoding and decoding the video, and the word "codec" is derived from the words "encoder" and "decoder". Encoding is the act of compressing the video during the recording or rendering process, and decoding is the reverse when the video is played back, edited or transferred to another format.
Intra-frame compression, also known as spatial compression, refers to compression of the image within each frame itself. In uncompressed video every pixel in every frame is fully described even if it is identical in color and luminosity to other pixels in the picture. Much like a bitmap still image. If intra-frame compression is applied, pixels of an identical or similar color and luminosity are indexed so that they don't have to be fully described, much like a JPEG still image. This reduces the bitrate and if a lossless codec is used (such as such as HuffYUV or Lagarith in post-production) then there is no loss of quality from the original. If a lossy codec is used such as DV then there is some loss of quality.
Inter-frame compression, also referred to as temporal compression, means that the video is compressed over time. Instead of defining every frame, only changes between frames are recorded, and so great savings in data rate can be made. Additionally predictions can be made based previous frames, thereby allowing further compression. Video is split up into GOPs (Group Of Pictures). Each GOP begins with an I-frame, which defines the full frame of video. B-frames. P-frames.
Picture of group of pictures with I frame and B and P frames.
Although inter-frame compression can achieve enormous reductions in bitrate, it does make the codec's job more difficult in decoding the video and so more computer power is needed. And difficulties can occur in post-production as a result of incomplete GOPs where the video is split away from I-frames.
The term compression ratio means the ratio of the size of the original data to the size of the compressed data. So a compression ratio of 2:1 would compress a video stream to half its original size.
Drop Outs (maybe as side bar)
A drop out is an error on a tape where the video signal is lost, perhaps due to a problem with the camera or tape when the video was recorded, or due to degradation of the tape, or because of a problem during capture of the tape. In the analogue realm tape errors would generally result in some retention of video data, but the signal would be degraded in quality. However in the digital world a drop outs or "lost packets" are catastrophic. Frames are skipped entirely or with interlaced video only one field might be captured correctly.
CAPTURE side bar - Capture is the process of transferring data from the recording media such as tape to the media it resides on for editing, such as a computer hard drive.
DV
DV is the most common standard definition digital video format. Most of the world's standard definintion camcorders use DV in its miniDV variant. This is the same as DV but on a small, compact cassette tape.
Picture of miniDV cassette
The DV standard encompasses both 4:3 and widescreen 16:9 aspect ratios.
PAL DV is 720 x 576 at 25 fps with a pixel aspect ratio of 1.0926. PAL widescreen simply widens the pixels to a PAR of 1.4568 to achieve a 16:9 DAR.
NTSC DV is 720 x 480 at 29.97 fps with a pixel aspect ratio of 0.9091. NTSC widescreen has a PAR of 1.2121 and a 16:9 DAR. See below for a full comparison table of video formats.
MiniDV is an excellent format and an awful lot of underwater DV footage has been broadcast on TV and delivered on DVD. Even top dollar productions like Blue Planet contained a share of footage captured on miniDV that could not be recorded with the bigger, more expensive cameras used for most of the footage.
Picture of TRV40
Picture of VX2000
DV has no inter-frame compression, only intra-frame spatial compression, and so each frame of video is fully described within itself and could stand alone in its own right without relying on other surrounding frames. This makes it a dream to work with in post-production where it can be easily and reliably edited on computers that are modestly powered by today's standards.
However miniDV can be quite prone to
errors and drop outs, especially on aging equipment or on equipment
that is not stored and maintained well. This is the sort of image
that would quite often crop up from time to time with my DV
cameras, typically if the heads were dirty or damp, or at the start
of an old tape.
Black and white stripes pic and maybe video too
I was once shooting a topside promo video at a swimming pool and discovered that my Sony DCR-VX2000 had recorded those black stripes across the first 10 minutes of video, for no apparent reason. The tape was brand new and the tape head had been recently cleaned. I guess it was due to the very high humidity. There was no other option than to ask the models to repeat the scenes, which of course was embarrassing. After running a miniDV cleaning cassette through again, and switching to a new tape, the video was fine, thankfully.
And here is what happened right in the middle of one of my favourite shots of rainbow runners.
Rainbow runners pic and maybe video too
Sadly there's not a lot that can be done with a drop out like this other than to consign it to your virtual cutting room floor.
Digital 8
Digital 8 is equivalent to a DV signal but carried on the larger cassette tapes that were originally designed for Hi8. Digital 8 cameras are therefore usually bigger than miniDV cameras but they are also often cheaper and the Digital 8 signal is typically more robust than miniDV because of more tape area per data, meaning that tape errors and drop-outs are less common.
DVD
There are camcorders available that shoot straight to DVD, usually to a mini DVD writeable or rewriteable disc. The main advantage of these is that they can played back on a DVD player without transfer or editing of the footage. The video on these cameras is highly compressed and the quality is not as high as DV. They have not been popular for underwater use but some housings no doubt exist.
DVCAM
DVCAM symbol
DVCAM is a professional format. The resolution, quality and bitrate of the video are identical to miniDV, but the data is spread over more tape so that it is less prone to errors. Even DV and DVCAM cassettes are interchangeable, but DVCAM cassettes are allegedly manufactured to higher specifications than regular DV. A "1 hour" cassette that would last 62 minutes recording DV would last 4X minutes recording DVCAM.
DVCAM has been a very common format for use as a "second unit" camera where the first unit camera might be a much bigger camera such as digital Beta. Because of the cost and small size, DVCAM is also very common as a professional or prosumer video format for underwater use, where the small size means a relatively small housing.
Picture of PD150 or PD170
Digital Beta
Digital Beta, or "digi Beta" is a very common standard resolution professional format. Before the advent of HD, most documentary work and ENG (electronic news gathering) footage would have been shot on Digital Beta. The cameras are too large and expensive for the amateur underwater videographer to consider but the format is worthy of mention here as it is often crops up in discussion of video formats.
HD
HD is the most common high definition professional video format. HD, sometimes referred to as "full HD", is 1920 x 1080, square pixels, intra-frame compression, no inter-frame compression. HD is to the high definition world what digi Beta is to the standard definition world. The majority of high definition programming is shot on HD and broadcasters such as the BBC, National Geographic and Discovery Networks state that a majority of their high definition production, typically 75 or 85% be shot on full HD. As a result, HD stock footage is more valuable and marketable than footage shot in more compressed high definition formats.
Picture of Sony F900
HDV
In the first few years of high definition, the format most commonly available to the consumer has been HDV. The HDV standard encompasses a range of resolutions and frame rates, but the two most commonly used are 1080-50i (PAL land) and 1080-60i. These are also often written as 1080i50 and 1080i60. HDV uses a combination of strategies to squeeze that 1080-line 16:9 signal into a similar bitrate to DV, but still retain good quality.
Firstly there is the interlacing, as described above, which effectively halves the bitrate compared to a progressive scan signal.
Secondly the video is only 1440 pixels wide (compared to the 1920 of full HD) but the pixels are anamorphic (rectangular) with a PAR of 1.33 and so the DAR ends up at 16:9. The human eye is more appreciative of high resolution in the vertical plane than the horizontal plane, which is why this approach succeeds.
Thirdly HDV uses a lot of inter-frame compression. It is recorded as an MPEG2 transport stream (.m2t) with a long-GOP structure. Specifically 1080i50 HDV only provides a full I-frame every 12 frames, and 1080i60 every 15 frames.
A common criticism of HDV is that the amount of temporal compression introduces unacceptable motion artefacts. Personally I have never found that to be an issue. In fact I can honestly say I have never even noticed a motion artefact in my HDV video, but then most of my footage is fairly static or of slow moving critters.
As described above in the section on compression, HDV requires more computer power than DV for processing, and the long GOP structure can be a real headache to work with. During or after capture, many editors will convert the m2t signal to a higher-bitrate intermediate format such as Cineform or Prores which has no temporal compression. Personally I edit and archive in native m2t format to save on storage space and avoid any quality loss.
Even though I currently use an HDV camera, I am not a huge fan of HDV. Certain things have cost me an inordinate amount of time and anguish in my post-production career so far. Early DVD writing technology was one of them. Early IEEE-1394 hard drive enclosures on Windows was another. And I put HDV firmly in this category. Establishing a reliable post-production workflow for HDV has cost me a huge amount of time and expense, and I know I am not alone in this. It is still nowhere near as straighforward to deal with is DV was. Aside from that, I still believe that my standard resolution Sony DCR-VX2000 DV camera takes more attractive video (and audio) than my Sony HVR-Z1P HDV camera in every respect except the resolution.
DVCPRO (Panasonic one)
DVCPRO is Panasonic's low end professional high definition format, encompassing a range of resolutions. There are some advantages to DVCPRO over HDV. Panasonic's first DVCPRO camera, the HVX200, was capable of progressive scan recording, albeit only at 720 lines, at a time when competing products from Sony etc. were not, so this made it attractive to shooters who wanted the film look of 24p. The camera could also be undercranked or overcranked at any frame rate from XX fps to XX fps, which made it attractive for capturing footage intended for slow motion. The format also has no inter-frame compression and so can be easier to deal with in post-production than HDV. And finally the cameras use P2 solid state memory cards for storage instead of tape, which has the benefit of rapid capture and random access of data. P2 cards were originally very expensive and rather limited in capacity, but more recently reductions in cost and increases in capacity have made them a more attractive proposition. More about Panasonic cameras later. LINK
AVCHD
AVC, or "Advanced Video Codec" is also known as H.264 and MPEG4 part 10. AVC is a more efficient compression format than the MPEG2 format used by HDV or DVD, which means the same quality in less space, or of course higher quality in the same space. AVCHD simply refers to AVC used at high resolutions (XX x XX and above - CHECK THIS). AVCHD is typically recorded not to tape but to a hard drive inside the camera, or to a solid state memory device such as an SD cardCHECK. The standard is flexible, meaning a variety of resolutions, frame rates and compression ratios are supported. AVC often goes hand in hand with AAC audio (Advanced Audio Codec) which is similarly efficient in its compression.
By 2008, AVCHD has become very important as a consumer video format and is found on many lower-end high definition video cameras. AVC has also gained popularity in other areas such as the internet. In late 2007 Adobe began to support for AVC/AAC in their Flash Player, putting AVC in a very strong position to become the web video standard of the future. The videos you (I hope) see in this course are AVC/AAC delivered in Flash Players.
AVCHD was initially treated with cynicism as a serious high definition acquisition format, but bitrates are climbing. Rates in excess of 20 mbps are now available in consumer cameras and the quality gap between AVCHD and HDV is now very small. AVCHD is likely to remain an important video format for a long time to come.
XDCAM EX
XDCAM EX is Sony's answer to Panasonic's DVCPRO and others, and an exciting development for professional and ambitious underwater videographers. XDCAM has been around for a while as a professional format, and with the XXXX camera, Sony brings the format within reach of the prosumer market. The video records to solid-state Express memory cardsCHECK, which are cheaper and more widely available than Panasonic's P2 cards. The format records up to 1920 x 1080, 24/24/30p, CHECK at 35 mbps. It has temporally-compressed long-GOP structure. Sadly the BBC has decided not to include the XXXX in their list of cameras approved for acquisition of a full high definition production LINK, but National Geographic and Discovery Networks LINK have. More about the Sony XXXX later. LINK
RED
RED has been creating a big stir in the industry. Their aim is to bring true high-resolution digital cinema within the reach of the (relatively) low-budget film maker. The RED One camera is a modular design and the format reaches 4k resolution at XX mbps, but can also be used to shoot 2k which is what many users are doing. Unfortunately there have been reliability and workflow problems with the early units but RED explain that they are a development company and are offering free upgrades to future models. Those who have seen RED footage in its full resolution always seem to report the experience as "mind-blowing". A Gates housing is in development for the RED One camera and units have already been taken underwater by Mark Thorpe LINK and others. A full RED underwater rig is likely to cost in the region of USD40k CHECK, so it's definitely not a setup for the budget-minded hobbyist.
Scarlett
Scarlett is a new camera under development from RED that may be promising to the underwater videographer due to it's proposed compact size, high resolution and relatively low cost. It's the little brother to the RED One camera. Resolution is rumored to be 3k (substantially above full HD) and the first deliveries are slated for late 2008.
Here is a comparison table of many of many video formats used in video cameras. It is by no means exhaustive.
|
Common Resolutions (i=interlaced, p=progressive) and frame rates |
Pixel Aspect Ratio (PAR) |
Display Aspect Ratio (DAR) |
Codec, Compression |
Video Bitrate |
Highest Audio Quality |
Video 8 PAL |
720 x 576i, 25 fps |
XX |
4:3 |
Analogue |
- |
|
Video 8 NTSC |
720 x 480i, 29.97 fps |
XX |
4:3 |
Analogue |
- |
|
Hi8 PAL |
720 x 576i, 25 fps |
XX |
4:3 |
Analogue |
- |
|
Hi8 NTSC |
720 x 480i, 29.97 fps |
XX |
4:3 |
Analogue |
- |
|
DVD-RW PAL |
720 x 576i, 25 fps |
XX |
4:3 |
MPEG2 |
XX mbps |
16 bit, 48,000 Hz, AC3 |
DVD-RW NTSC |
720 x 480i, 29.97 fps |
XX |
4:3 |
MPEG2 |
XX mbps |
16 bit, 48,000 Hz, AC3 |
DV / miniDV / DVCAM / Digital 8 PAL |
PAL: 720 x 576i, 25 fps |
1.0926 |
4:3 |
DV, intra-frame |
XX mbps |
16 bit, 48,000 Hz, PCM |
DV / miniDV / DVCAM / Digital 8 NTSC |
NTSC: 720 x 480i, 29.97 fps |
0.9091 |
4:3 |
DV, intra-frame |
XX mbps |
16 bit, 48,000 Hz, PCM |
DV / miniDV / DVCAM widescreen PAL |
PAL: 720 x 576i, 25 fps |
1.4568 |
16:9 |
DV, intra-frame |
XX mbps |
16 bit, 48,000 Hz, PCM |
DV / miniDV / DVCAM widescreen NTSC |
NTSC: 720 x 480i, 29.97 fps |
1.2121 |
16:9 |
DV, intra-frame |
XX mbps |
16 bit, 48,000 Hz, PCM |
HDV 720p |
1280 x 720, 24/25/30p |
1.0 |
16:9 |
MPEG2, long GOP |
25 mbps |
16 bit, 48,000 Hz, MP2, 384 kbps |
HDV 1080i |
1440 x 1080, 50/60i |
1.333 |
16:9 |
MPEG2, long GOP |
25 mbps |
16 bit, 48,000 Hz, MP2, 384 kbps |
AVCHD 720p |
1280 x 720, 24/25/30p |
1.0 |
16:9 |
AVC |
Upto 21 mbps |
AAC |
AVCHD 1080i |
1440 x 1080, 50/60i |
1.333 |
16:9 |
AVC |
Upto 21 mbps |
AAC |
DVCPRO |
1280 x 720p, 24/25/30 fps |
1.0 |
16:9 |
DVCPRO |
XX mbps |
|
XDCAM EX |
1920 x 1080 |
1.0 |
16:9 |
XDCAM, long GOP |
35 mbps |
|
HD |
1920 x 1080p |
1.0 |
16:9 |
HD, intra-frame |
50 mbps |
|
RED |
4000 x |
|
16:9 |
Redcine |
|
|
Scarlett |
3000 x |
|
16:9 |
Redcine |
|
|
< back to video concepts & terminology | video formats | forward to video cameras >