|
crowd sourcing the experts |
|||||
| |
|||||
 |
|||||
(Page 2 of 2 pages for this article < 1 2) Filed under: Cameras • Review: Canon Vixia HF11 AVCHD camcorderThis inexpensive consumer camcorder makes surprisingly good HD images.
Camera PerformanceThe 12x zoom runs from 4.8-57.6mm. In 35mm SLR terms, it starts at the equivalent of 43mm. Compared to an EX1, it’s about 2/3 as wide. The optical zoom range can be supplemented with digital zoom to a maximum of 40x or 200x. As you might expect, such magnified images are fairly soft and noisy, and there’s a discontinuity in zoom speed as one crosses the boundary between optical and digital magnification, but the digitally zoomed images are as cleanly magnified as you could hope for. The lens has a bit of barrel distortion fully wide, but not to excess; around midzoom it’s fairly distortion-free, and it’s a bit pincushiony at telephoto—again, not to excess by comparison to most other camcorder zooms. Chromatic aberration is present when zoomed in or out all the way, showing as green/magenta fringing, but it’s minimal. The lens holds its f/1.8 maximum aperture for the first 2/3-3/4 of the zoom range, and ramps smoothly down to f/3.0 at full tele. There’s some portholing in that last quarter of the zoom range, darkening the edges of the image about a stop down from the center. The lens focuses as close as 1cm when wide, but near focus is a meter away at telephoto. Overall, it’s a pretty good lens for its zoom range, especially given its tiny size and price. The HF11 has Canon’s refined optical stabilization, which works very well to take most of the shake out of this featherweight camera. The sensor is a single, 1/3.2” 4x3 CMOS chip (thus a bit smaller than a 16x9, 1/3” chip) using an RGB Bayer pattern. In video mode, it uses 1920x1080 effective photosites, and feeds its image to Canon’s DIGIC DV II processor. Yes, the chip has a rolling shutter, and violent camera motions show “jellocam” distortions. But these are minimized when optical stabilization is on, since it smoothes the high-frequency shakes that make jellocam most obvious. Fast pans still show tilt, but the rolling shutter reads the chip in 1/60 second, so image distortions are no worse than on cameras like the Sony PMW-EX1 and EX3. Practically speaking, the rolling shutter is only an issue with the most frenetic shakycam moves. The best resolution we should expect from a 1080-line Bayer-pattern sensor is 80%, or about 800 TV lines per picture height (TVl/ph). Indeed, that’s pretty much what I see; I’d call it somewhere in the range of 750-800 TVl/ph. The image above that resolution is fairly clean, too, though some chroma aliasing is present and there is noticeable diagonal aliasing, both of which are common characteristics for Bayer-mask images.
A 1:1 excerpt from a DSC Labs multiburst chart, recorded in MXP.The chart shown was shot using the camera’s highest-quality mode, 24 Mbit/sec 1920x1080 PF30, but I saw little variation across frame rates or recording formats. Both interlaced and progressive shooting modes gave me similar results, as did different recording formats: on a still image, as I’ll discuss, even 5Mbit/sec LP mode gave a comparable picture, despite its 1440x1080 recording (which tops out at 787 TVl/ph resolution in the horizontal direction). The camera offers three 16x9, 1080-line formats: 60i, PF30, and PF24. 60i is of course the standard 59.94 fields per second, interlaced; PF30 and PF24 are 30 fps and 24 fps progressive images, recorded using interlaced formatting with 2:2 (PsF) pulldown for PF30 and 2:3 pulldown for PF24. Assuming a constant bitrate across the three frame rates, there is no increase in efficiency at lower frame rates since the signal is always processed as 59.94 interlaced, though of course the inherent redundancy of temporally-correlated fields and repeated fields (in PF24) may improve compression quality somewhat. The camera offers video shutter speeds in 60i and PF30 of 1/8, 1/15, and 1/30 second; in PF24 you can choose 1/6, 1/12, 1/24, and 1/48. All frame rates allow 1/60, 1/100, 1/250, 1/500, 1/1000, and 1/2000. There is no clear scan / synchro scan. Apertures can be set in half-stop values between 1.8 and 8.0, zoom range allowing. It’s hard to get a feel for the camera’s sensitivity, since the exposure system doesn’t let you lock out gain changes nor does it tell you what the current settings are. When ample light is present, the HF11’s images are very clean; when the light level drops, they get noisier, but not as noisy as you might expect. Between the comparatively large pixels (cameras of this ilk often use smaller chips with higher pixel counts) and the DIGIC processing, the camera does as well in low light as many 3-chip 1/3” camcorders. What noise there is is a fairly unobtrusive balance between luma and chroma. As the light level drops even further, the camera can automatically go to a slow shutter (if enabled) or it can boost gain further, or you can turn on the HF11’s LED video light. It casts a cold bluish glow towards the center of the image and lets you get pix at reasonably close distances in what would otherwise be utter darkness. If you’re looking for a full-color simulation of IR nightvision, or a “Cops”-style flashlight-cam look, it’s handy; it’s also useful for simple location scouting pix (the HF11’s small size and weight makes it a great take-along cam for location scouts even if you plan to use a “real” camera for actual production). Dynamic range is very good: I’m able to see about 8.3 stops cleanly resolved in normal shooting modes, and a full 9 stops in cinema mode. In normal modes, the camera is a bit contrasty and uses an auto knee which sets in anywhere from 70% and up. In cine mode, contrast is tamed and the knee is fixed at 70%. Hue and luma don’t show much obvious distortion at the knee point, and colors tend to bleach to white just before clipping sets in; the pix from this camera look very much like pix from an XL H1 or XH A1. That’s the DIGIC image processor again, handling highlights just as it cleans up the noise, and giving the HF11’s images that characteristic Canon look. Color rendition is clean and consistent, if perhaps a bit bold. Backing the saturation off to “-” renders a neutral, more accurate image. Contrasty edges along areas of overexposure can show a bit of purple fringing; it’s noticeable but not so prevalent as to be distracting. The camera does not offer variable frame rates, 720p modes, or 480i/p modes, though the component output can be set to downconvert to 480i, and both component and composite outputs can be set to display either full-frame or letterboxed 480i images. The camera records two channels of Dolby AC-3 audio at an unspecified bit rate. I tested recording with the built-in mics, feeding them music from near-field monitors. In automatic gain, sudden peaks tended to cause clipping distortion until the auto level control caught up; in manual mode the sound was clean, if a bit lacking in bass (which I attribute to the mics). There weren’t any overt compression artifacts, at least none that I can point to objectively. I’d certainly use the HF11 for background audio and interview tracks; I’d want to test it more rigorously before recording feature film sound on it.
AVCHD RecordingThe HF11 records AVCHD (long-GOP AVC, a flavor of h.264) to flash memory, either the camera’s built-in 32 GB buffer or an SD / SDHC card. It offers five bit rates, which it labels with arbitrary letter codes; I list them with their peak data rates, resolutions and recording times on the internal memory:
All bit rates use 8-bit, 4:2:0 sampling and 60i recording.
SD / SDHC cards (SDHC is a revised SD card capable of holding more than 2 GB) need to meet certain restrictions: they must be 128 MB or larger, and must have an SD speed class of 2 or greater (4 or greater for FXP and MXP recording). At press time, a 16 GB class 6 SDHC card is about $25-$30: not as cheap as tape, yet, but still cheap enough that you can carry a pocketful of cards to a shoot and not worry about offloading data in the field. The camera has a simple, tabbed interface in playback mode for sorting between stills and clips.
Tabbed interface shows metadata for current clip at bottom of display.You can even create simple playlists in-camera. The camera connects to a computer via USB; files can be transferred to hard disk via drag ‘n’ drop. The camera can also connect to a Canon DVD burner via USB, and make an AVCHD DVD for every format other than MXP (the 24 Mbit/sec data rate of MXP is too high for a DVD). Final Cut Pro 6 on a 2.3 GHz MacBook Pro transcodes the AVCHD files to ProRes422 via the Log & Transfer window (but not ProRes422 HQ, alas), taking about twice real time to do so. FCP can also transcode to the Apple Intermediate Codec for those who prefer it. In Adobe Premiere CS4, the AVCHD clips are directly usable, and play in draft mode with only the occasional stutter. All clips import as 29.97 interlaced, since that’s how they are recorded. Extracting “pure” 24p from the PF24 clips can be done using Apple’s Cinema Tools, DVFilm’s Maker or other reverse-telecine tools. Fine, so AVCHD is usable in common NLEs. How does it look? AVCHD has a “not ready for prime time” reputation based on early reviews, all of which have been at lower bit-rates. Yet AVC in general is said to be roughly twice as efficient as MPEG-2, meaning that the same quality can be obtained at half the bit rate. The HF11 is the first camcorder to support the 24 Mbit/sec rate, which should allow it—if theory holds—to make better images than 25 Mbit/sec HDV, or even 35 Mbit/sec XDCAM EX HQ. I found that on static images, there was little to differentiate even the 5 Mbit/sec clips from the 24 Mbit/sec clips. However, the quality of the lower bit-rate clips degraded quickly as soon as any motion was introduced; indeed, the quality of moving pictures recorded with AVCHD varied in direct proportion to bit-rate. Slow pans of dense, sunlit foliage held up very nicely at 24 Mbit/sec, while at 5 Mbit/sec the trees were reduced to fuzzy green blobs, only resolving to individual branches and leaves once the camera stopped moving (and once there was no breeze to move the trees themselves!). Intermediate data rates gave intermediate results. Overall, 5 Mbit/sec looked like bad web video, while 24 Mbit/sec looked like it could hold its own against HDV or XDCAM EX. To test this more rigorously, I bolted the HF11 alongside a Sony PMW-EX1 using a spare RED arm, so I could operate the cameras in parallel.
The HF11 attached to a Sony PMW-EX1 using a RED Arm.I then set up a variety of busy scenes, indoors and out, setting both cameras to the same angle of view, same frame rate, and same shutter time. I could then shoot the same scene with both cameras simultaneously, recording the same static images, wild pans, rolls about the lens axis (“dutch” angles), and general codec-stressing shakycam. I pitted the HF11’s five bit-rates against the EX1’s 25 Mbit/sec, 1440x1080 HDV-compatible mode, and its 35 Mbit/sec, 1920x1080 EX HQ mode. I imported both camera’s clips into FCP, transcoded them both to ProRes 422, lined up matching clips in a top-and-bottom splitscreen, and compared the compression. In 60i mode, I found that the HF11’s 17 Mbit/sec images held up almost as well as the EX1’s 35 Mbit/sec images, while its 24 Mbit/sec pix were visibly less affected by compression artifacts. Likewise, the HF11’s 12 Mbit pix almost equaled the EX1’s 25 Mbit pix, while its 17 Mbit pix were better. AVCHD really is roughly twice as good as MPEG-2 in terms of bit-rate for a given quality of image. Now, not every frame of every comparison bore this out: in any given test where the results were close, sometimes the EX1’s images were better than the HF11s, and a few frames later, the reverse might be true. The two clips were created by different cameras, and had slightly different content (due to parallax). I could not guarantee that both cameras started their GOP structures at the same point in time; indeed I cannot say that they even have the same GOP length. Nonetheless, in normal 1x playback the quality of compression appeared essentially equal when AVCHD was recorded at half the bit-rate of the corresponding MPEG-2. In frame-by-frame examination of the clips, the MPEG-2 (at roughly twice the bit-rate of the AVCHD) tended to equal the AVCHD images perhaps 70% of the time; of the remaining 30% where I could point to an unambiguous quality difference, the MPEG-2 images were better perhaps 2/3 of the time, and the AVCHD excelled the other 1/3 of the time. Put another way, AVCHD equalled MPEG-2 70% of the time, surpassed it slightly 10% of the time, and fell slightly beneath its quality 20% of the time, while running at half its bit-rate (17 vs. 35 Mbit/sec, and 12 vs. 25 Mbit/sec). In progressive, the contest tilted more towards MPEG-2, since the EX1 records its 35 Mbit/sec as true progressive, while the HF11 uses pulldown to record both 30p and 24p as 60i. As a result, 17 Mbit/sec AVCHD wasn’t able to hold up to 35 Mbit/sec MPEG-2; and even 24 Mbit/sec wasn’t quite able to equal it, though it came close. Part of the problem with AVCHD, at least as recorded on the HF11, is that it’s always recorded as interlaced, and thus shows interlaced 4:2:0 chroma:
A 2:1 blowup of an HF11 progressive image. Note the “sawtooth” chroma, indicative of interlaced 4:2:0 color encoding. (The HF11 was shooting wide-open at full telephoto, and it focused on the hummingbird’s tail: sometimes shallow depth of field works against you!)Interlaced coding also isn’t as efficient as progressive coding, and even at 30p (where at least there were no repeated fields in either format) the 35 Mbit/sec XDCAM EX footage held a slight quality lead over the 24 Mbit/sec AVCHD. When the EX1 was set to SP recording (HDV-compatible 25 Mbit/sec MPEG-2, using pulldown) the contest was closer, and the nearly-two-to-one advantage of AVCHD held—because, in essence, I was still comparing two interlaced recordings. Yes, it’s not an entirely fair test, since the cameras were different: the EX1 produces more finely detailed images that are harder to compress than the images the HF11 produces. Even so, I was able to produce interlaced pictures using the HF11 that were essentially as good, as far as compression is concerned, as interlaced images from the EX1 recorded at twice the bit-rate. Based on my testing, I’d say that if you’re happy with HDV, or with interlaced EX1 footage, you’ll be happy with high-bitrate AVCHD as produced by this little Canon. The HF11 wasn’t quite able to match the EX1 in 24p, but that may be due to the HF11’s interlaced encoding; I’d like to test an AVCHD camera laying down its progressive picture using progressive, native-frame-rate encoding. Overall? AVCHD is capable of turning out high-quality images, and at least in interlaced modes, the HF11 can record images of comparable quality to HDV and XDCAM EX camcorders. ConclusionThe HF11 intrigued me due to its addition of 24 Mbit/sec recording to the high quality I’d seen from its HF10/HF100 brethren, and in testing it proved itself a capable camera. While it’s clearly aimed at the consumer crowd, it has the ability to produce interlaced images fully usable in a professional environment (at least, if you accept that HDV and XDCAM EX interlaced images are usable professionally, then the HF11’s pictures will be perfectly acceptable to you). Its progressive pictures are slightly degraded by interlaced encoding, but even then its quality is surprisingly good. For a walk-around consumer camera, it’s hard to fault the little Canon. For professional applications, it holds up very well if you respect its limitations: it won’t let you focus, zoom or tweak exposure and audio on the fly with anything like the control afforded by a pro camera, but its basic image quality is such that if you lock it down before a shot, or if you’re happy to let the camera control focus and exposure, the HF11 will capture images that approach those of cameras costing (and weighing) as much as several HF11s put together. Yes, there’s a bit of chromatic aberration (it’s still less than on many three-chip HD camcorders costing ten times as much). Yes, there’s some purple fringing on blown-out highlights. There’s no clear scan, there’s no 50i compatibility, and it records 4:2:0 interlaced color. But it makes an image that’s about as sharp as a single chip 1920x1080 imager can, with color, detail and dynamic range that are close matches for Canon’s 3-chip cameras. That’s nothing to sneer at. If you need something small for travel purposes; something lightweight and shock-resistant; something capable of recording nearly three hours at a go internally at its highest quality, or two hours on a 16 GB SDHD card at 17 Mbit quality (which for talking-head interviews is well-nigh indistinguishable from 24 Mbit/sec quality); something that will intercut seamlessly with XL H1s or XH-series HDV cameras; something that can fit in a large coat pocket and still leave change from a $1000 bill (assuming, of course, you can even find a $1000 bill): the HF11 is well worth a look. Besides, it’s just plain fun to use. It weighs nothing and it fits comfortably in your hand, and its image quality is better than it should be for the price. The following bullet points are based on looking at the HF11 as a camera for professional use. Pros
Cons
Cautions
Recommended For…
See also: Chris Hurd’s HF11 overview and comparison with the HDV-format HV30 Vixia HF11 review at camcorderinfo.com Canon’s Vixia HF11 webpage.
(Page 2 of 2 pages for this article < 1 2) 
You must be registered to comment. This is an effort to reduce spam. Please REGISTER HERE.
|
|
||||
|
|
||||
Redrock Micro’s ultraCage for the C300New Cage Fits New Camera Like A Glove
At the 11th Annual San Francisco SuperMeet, I was able to actually lay hands on the |
DP Alex Buono Talks C300 At SF SuperMeetSNL Veteran, Talks About Using the C300 For Network Television…
Recently I attended the 11th annual San Francisco Supermeet. Beyond some great lectures by some great speakers, there were a variety of different hardware manufacturers and software companies to… |
Working Pros Discuss the Canon C300Rodney Charters, Drew Gardner and Lan Bui discuss the Canon C300
What happens when you get three working professionals together late at night with the Canon C300 camera and liquor? They shoot camera tests. And that’s exactly what happened when… |
Sony’s NX70 camera to receive its missing 29.97p framerate via free firmware update29.97p is a vital framerate for producers in ex-NTSC countries. I’m glad Sony has recognized this fact and is finally adding it to the NX70.
Sony has just announced that the NXCAM camera model known officially as the HXR-NX70 (often followed by a regional suffix) —but colloquially known simply as the NX70— will… |
_thumb.jpg)
To be considered for listing, contact pr (at) provideocoalition (dot) com
Copyright © 2011, HD Expo, LLC a division of Diversified Business Communications. DBA Createasphere
All rights reserved. HD EXPO, High Def EXPO, Createasphere, E-Tech, Entertainment Technology Exposition, 3D Production Workshop, VariCamp, P2 Camp, ColorCamp 101, and Lighting, Filters & Gels for HD are all trademarks of HD Expo, LLC.