Sony's HDV Debut
It had been a long, quiet summer on the HDV front. Then, a few days before the opening of IBC 2004 in Amsterdam, came Sony's HDV cam-corder announcement. Via the Net we quickly learned Sony's new HDV camcorder looked much like the consumer camcorder pre-announced at Cebit by Sony just prior to NAB 2004.
From the press release, we learned a few more facts. The HDR-FX1 would have three 1/3-in. CCDs with a 16:9 aspect ratio. Each CCD would have 1.12 megapixels arranged in a 960×1080 matrix. It would have an MSRP of $3,700 and go on sale within the USA in November.
As I searched for more information, the quest became complicated by the fact that the press release described the consumer version, the HDR-FX1, while the company was showing a pre-release version of the professional camcorder, which does not yet have a model name. To keep certain features of the professional version secret, Sony taped over the menu control.
Looking for more details, I found that the only Sony location with detailed information was its Japan website (www.sony.jp/products/Consumer/handycam/PRODUCTS/HDR-FX1). Therefore, this report builds on this site, information from folks who attended IBC, my good friend D.W. Leitner, who attended a Sony presentation, and Sony USA's Consumer Marketing group.
By the time you read this, there no doubt will be extensive marketing materials available on the HDR-FX1, so I'll focus here on seven camcorder topics that have the most relevance for professional shooters.
There may be two versions of the FX1. One for Region 60 will support 1080i60 (1080i59.94) plus 4:3 and 16:9 NTSC DV. The other, for Region 50, would support 1080i50 (1080i50.00) plus 4:3 and 16:9 PAL DV.
Stereo audio and 1080i video will be encoded using MPEG-1 Layer 2 and long-GOP (either 12 frame or 15 frame) MPEG-2, respectively. The 25Mbps encoder output will be recorded as PES (Packetized Elementary Stream) data. The PES data structure is different from the Transport Stream structure used by the JVC HDV camcorders and D-VHS decks. This means the FX1's i.LINK output cannot be recorded to D-VHS. Moreover, Sony has already acknowledged the FX1 is not yet compatible with Bluray DVD recorders.
Software, of course, can be enhanced to support PES because TS and PES have a hierarchical relationship. Video and audio Elementary Streams (ES) are combined, along with frame presentation timestamps, into packetized streams (PES). Transport Stream data are generated when PES data are divided into packets of exactly 188 bytes. The bottom line is that until Apple, Avid, Adobe, Pinnacle, Sony (Vegas, VAIO, and XPRI), and Ulead release the 1080i HDV support they all announced at IBC, you might not be able to edit 1080i HDV.
The HDR-FX1 has an F1.6 to F2.8, Carl Zeiss Vario-Sonnar T* (anti-reflective coating) lens that accepts 72mm filters. The 12X zoom covers a range of f = 4.5mm to 54.0mm (35mm equivalent: f = 32.5mm to 390mm.) You can zoom by variable servo control, a lever, and a nonperpetual zoom ring.
Sony developed a new 1/3in., Super HAD CCD for its HDV camcorders. Each CCD has 1,012 (horizontal) by 1,111 (vertical) elements (1,120,000 pixels) that provide an effective pixel count of 1,070,000 pixels (972 horizontal by 1,100 vertical. Vertical smear level is rated at a very low -107dB. Each element has a 2:1 aspect ratio.
This aspect ratio explains how Sony can work with an image that is 960 pixels wide by 1080 pixels high. To obtain an image aspect ratio of 16:9, the pixels from these CCDs could not be square. One ninth of 1080 is 120, which when multiplied by 16 indicates that there should be 1920 pixels in each row. By making each pixel twice as wide, only 960 elements are required.
Sony's use of CCDs with micro-lenses partially accounts for the FX1's light sensitivity value of only 3 lux. Each CCD element has a tiny lens (shown in the left half of Diagram 1) that gathers light and focuses it on the center of the CCD. (This implies that the extra width of each element, on either side of the centered lens, does not play a big role in increasing light sensitivity.)
Both the JVC “HD-1” and Sony's “HD-2” HDV definitions utilize Main Profile@High1440 (MP@H-14) MPEG-2. (See “Profiles and Levels” in the July 2004 issue.) This format accepts a vertical pixel count of up to 1,080 and a horizontal pixel count of up to 1,440. JVC works with 720 by 1280 while Sony works with 1080 by — 960? That's certainly not what I expected. In my article “The HDV Sweet Spot” last month, I wrote “It is customary to employ CCDs that have at least the number of columns as the maximum number of columns supported by the video format.” I certainly expected a CCD element count of at least 1,440 pixels. What happened? It turns out Sony is using a technology that can provide 1,440 pixels from CCDs with only 960 elements.
By offsetting the green CCD one half-element spacing from the red and blue CCDs, a source of additional luminance information is created. By combining output from all three CCDs, horizontal resolution is increased by up to 150 percent. And, indeed when 960 is multiplied by 1.5, the result is 1,440. Naturally, that leads to the question of how “real” is the extra resolution obtained by using pixel offset technology.
The simple answer is that when resolution tests are performed, the horizontal resolution will be that expected from a CCD 1,440 elements wide. The complex answer is that with pixel offset technology, effective horizontal resolution is a function of the colors, the color patterns, and the motion of colored objects in a scene. (When pixel shift isn't employed, effective horizontal resolution is fixed — not variable.) To demonstrate that resolution is dependent on what's in a scene, I created several diagrams that show effective horizontal resolution of a very imaginary scene — a picket fence that is imaged by tiny CCDs. (Green is displaced half a pixel from red and blue.) While not at all mathematically accurate, the six cell values (0 to 6) in the diagram's top row represent the combined output from each column of green plus red and blue CCD elements. The gray level in each cell is proportionate to the cell's signal value. So, if you look closely at the top row of each of the Diagrams 1 and 2, you'll see a grayscale pattern. You should see differences in the clarity of the fence pattern.
The 4:2:0 chroma data are obtained from sub-sampled chroma information from the three CCDs. The two color components Y-R and Y-B are obtained from G-R and G-B, respectively.
Conversion of the three signals from the CCDs is accomplished by a 14-bit A/D. The digital data are then processed by a 14-bit DXP (i.e., DSP). The wide-word-length converters and DXP hopefully will prevent highlights from blowing out. The DXP also supports gain settings of 0, +3, +6, +9, +12, +15, and +18dB. Shutter speeds range from 1/4 second to 1/10000 second. It seems very likely the FX1 correctly employs the ITU Rec. 709 HD color space.
Two-channel audio can be obtained from the built-in stereo mic or via a 1/8in. mini-jack. An audio level control adjusts the level for both channels simultaneously. While some will object to this limitation, the Pro version does have dual audio-level controls. The Pro version also has dual XLR connections.
DV audio is carried as PCM data while HDV audio is carried as MPEG-1 Layer 2 data (MP2). This precursor to MP3 is, like MP3, a “perceptual” encoding system that discards audio information that DSP computations indicate will be masked by other audio information. In short, it is a lossy encoding system that may not meet field-recording requirements.
While the FX1 offers a CinemaTone mode, Sony USA has yet to reveal any details. From the Japan site, however, it appears this mode alters gamma in two ways. First, a “knee” is established at around 80 IRE that gracefully prevents highlights from exceeding about 100 IRE, thereby preventing the total loss of highlight detail when portions of the image are too bright. This function simulates the increased light latitude expected from film.
Second, the portion of the gamma curve below 80 IRE is altered to compress the video signal slightly by shifting levels downward about 20 IRE. Doing so increases the richness of mid-tone colors. Of course, this also crushes blacks slightly, thereby decreasing shadow detail.
According to Sony USA, CineFrame 30 uses 2:1:2:1 pulldown to create video that has a temporal resolution of 30fps. Diagram 3 shows how 2:1 pulldown accomplishes de-interlacing of 1080i60 video. (Red text indicates the 30 samples per second.) While the downside of CineFrame 30 is less effective vertical resolution (about 540 lines), the upside is an image with no interlace combing on progressive displays.
In CineFrame 24 mode, the camcorder creates “24fps” video. Sony claims CineFrame 24 works by dropping every fifth interlaced frame over the period of 30 frames, thereby creating 24 interlaced samples per second. Although 24 samples per second are obtained, clearly the sampling interval is not constant as it would be with a real 24fps camera. Nevertheless, CineFrame 24 should yield a film look that will be acceptable to some. (A far more acceptable film look would be obtained if the consumer camera's CCD frame rate were set to 24fps. The sampling interval would then be constant and would truly have a frame rate of 24fps.)
Also according to Sony, as shown in Diagram 4, 2:3:2:3 pulldown is applied to the 24 samples to generate 1080i60 video. (Six times each second, four samples are converted to five frames, yielding 30 frames.) By applying pulldown, quasi-24fps video is carried as 1080i60 HDV video.
How difficult would it be for the Pro version of the FX1 to support 24p video? Perhaps not difficult at all — if the CCDs' frame rate can be set to 24fps instead of 30fps. CCDs are inherently progressive; thus, every CCD readout yields 1080 lines. For interlaced scanning, “row pair summation” creates a 540-line field from each scan. However, were summation disabled and every other readout discarded, each 1/24 of a second, all 1080 lines from the CCDs would become a progressive frame stored in a buffer. From the buffer, alternating “fields” of 540 odd and 540 even lines would become available as needed. (The fields, of course, are not interlaced.) This mode of operation is very similar to the 1080/PsF24 recording used by Sony CineAlta cameras.
To record 24p to tape as 1080i60, pulldown must be applied. Although 2:3:2:3 would work fine, a 2:3:3:2 cadence would be optimal. With 2:3:3:2 pulldown, when 1080i60 is transferred by i.LINK to your NLE's hard disk, one of every five frames — the judder frame containing fields from two frames — can simply be discarded. The result is 24p video. If 2:3:2:3, which has two judder frames, were used by the Pro camcorder, software that employs reverse 2:3:2:3 pulldown can recover the 24p.
One of the most important features of the HDR-FX1 is its 3.5in., widescreen color LCD that has 250,800 pixels. The color LCD viewfinder has 252,000 pixels. To aid focusing, you can push a button and the LCD zooms by a factor of four.
Although the HDR-FX1 looks quite large, it weighs only 2kg (4.4lbs.). Analog component output, i.LINK (HDV and DV), and a LANC port are some of the connectors supported by the FX1. Although a 720p30 tape can be played and output as 1080i via component analog, the 720p30 Transport Stream cannot be output by the FX1's i.LINK port. Three batteries are available: the NP-F570, NP-F770, and NP-F970 provide 65, 130, and 205 minutes of recording time, respectively.
Without doubt, the HDR-FX1 will be the camcorder of the year. It will likely initiate a rapid move to HD the same way the Sony VX1000 drove the move from analog to digital. The FX1 also offers the bonus of “over-sampled” NTSC 4:3 and 16:9 DV in a better-than-VX2100 package. I expect Sony will easily sell every FX1 that it can manufacture over the coming months.
Contributing editor Steve Mullen is owner of Digital Video Consulting, which provides consulting and conducts seminars on digital video technology. Mullen can be reached at
email@example.com. His website is
To comment on this article, email the Video Systems editorial staff at firstname.lastname@example.org.