Canon Vixia HF10 Camcorder Review
The Canon HF10 is among the most anticipated camcorders in recent years, due no doubt to the outstanding reception of its HDV predecessor, the HV20. There are several key differences, however. The CMOS chip is not the same as last year’s Canon HD camcorders; in fact it’s smaller, with a higher pixel density. The HF10 records in the AVCHD format, heretofore riddled with performance problems. This is Canon’s third try on an AVCHD camcorder, and the format appears to finally be coming into its own. Full 1920 x 1080 output, a new 17Mbps bitrate, and 60i/24P/30P frame rates make this the best AVCHD camcorder we’ve seen to date. Prepare to be surprised.
Video Performance* (10.5)*
The Canon HF10 does not use the same imaging chip as its popular predecessors. The HV20, HV30, HG10, and HR10 all use a 1/2.7-inch CMOS with 2,960,000 pixels, which produces outstanding results for color and sharpness. There is some worry, naturally, about a turn away from that, especially when the new imager has been reduced in size. The HF10 features a 1/3.2-inch CMOS with a gross pixel count of 3,310,000 pixels – that means more pixels on a smaller surface area. We’ve seen this strategy in the past from Sony, viewed mostly as a means to make the specs look more impressive on the tags at Best Buy and Circuit City.
Here’s the problem. When surface area on an imager shrinks, but includes the same or a greater number of pixels, the size of the individual pixels has to shrink. Typically, smaller pixels mean a reduced ability to gather light, which can hurt low light, as well as an increase in noise. There are ways manufacturers can combat the negative effects, such as using better materials and smarter processing, but the danger of worse performance is real.
Even as the imager was shrunken, Canon has made two improvements this year that have raised hopes. The outputted resolution has been increased from 1440 x 1080 to 1920 x 1080 (though we saw from the Panasonic HDC-SD9 that this makes no improvement on our resolution test if the gross pixel count is the same). Secondly, the bitrate increased from 15 Mbps on the HG10 to a new high of 17 Mbps. This is expected to increase how motion is rendered, as well as decrease compression artifacting.
For the in-the-lab portion of testing, we shoot our DSC Labs ChromaDuMonde under near-ideal conditions – an even 3000 lux of light. The results look excellent, nearly identical to the Canon HV20 and HV30. The color balance is great, with just the right amount of saturation. In very close examination, the HV30 seems to produce slightly sharper edges, but you have to be paying very close attention.
It may be that this is the AVCHD camcorder people have been waiting for. Take a look at the 200 percent crop between the HF10 and last year’s HR10, an AVCHD camcorder recording to DVD. We’ve also thrown in the HV30 chart. Note how much the resolution has improved since last year.
Against the competition, the Canon HF10 produces a sharper, less saturated image than the Panasonic HDC-SD9. If you like the Panasonic’s vivid colors, you can boost saturation on the HF10. However, there is no tool on the Panasonic SD9 to lower the saturation. The Sony HDR-CX7 looks very good on these charts, but in motion testing, it doesn't perform as well.
We also shot the chart in other frame rates and gamma settings offered by the HF10. Shooting in 60i with the Cine color mode engaged, the highlights and shadows are compressed and the middle range is expanded. This has the effect of nicely desaturating the colors, but losing details in darker areas of the shot.
Shooting in 30P and 24P has little effect on color and sharpness in this light. On the whole, 24P looks much better than on previous Canon AVCHD models. If you’ll recall, we slammed the HG10 and HR10’s performance in 24P mode, and we weren’t alone. The new bitrate seems to have made a positive impact. The HF10 is the first Canon AVCHD to offer 30P. We did extensive testing on the HV30 earlier this year and were quite pleased. The HF10 matched it in all important regards.
Outside of the lab, we found several interesting results. Shooting in 60i mode outside on a cloudy day, we saw very little difference between the HV20, last year’s Camcorder of the Year, and the HF10. This includes color, motion, sharpness, and compression artifacting. No one could be more surprised than us, the very people warning others off of AVCHD for the last two years. We shot identical scenes with the HV20 in one hand and the HF10 in the other. Looking very, very closely, we could see some instances when the HV20 looked sharper during panning. There was more ghosting along high contrast areas on the HF10, as well. These differences were so slight that the most consumers will not see a difference.
In indoor shooting, with moderate light, again we saw very little difference. The sensitivity in low light is reduced with the smaller imager, but the sharpness looks great.
Amazingly, in the particular low light shot below, the HF10 looks better than the HV20, despite the smaller imager. The sharpness is improved, and there is visible noise. This actually contradicts the hard numbers testing we ran in Low Light testing (further down the page), which indicates higher noise. This may only mean that the HV20’s noise is more discolored (the tiny red, green, and blue pixels), while the HF10 produced a more monochromatic noise in greater amounts.
In summary, the HF10 is best AVCHD performer we’ve seen yet, and may finally be the model to crest the summit established by HDV.
Video Resolution* (21.94)*
The video resolution is tested by shooting a DSC Labs video resolution chart at an even, bright light, then examining the playback footage on an HD monitor. We shot in all three frame rates. At 1080/60i, the Canon HF10 produces a horizontal resolution of 675 line widths per picture height (lw/ph) and a vertical resolution of 600 lw/ph. This is a good score, better than any AVCHD camcorder we’ve seen to date. In fact, since we introduced this test, this is the first camcorder to produce such a high horizontal resolution in any media category. It’s important to note that this exact test did not exist when we wrote some of our professional camcorder reviews.
In 30P mode, the Canon HF10 produces a horizontal of 650 lw/ph and a vertical resolution of 650 lw/ph. This score is higher than the 60i score, particularly the vertical.
In 24P, the Canon HF10 produces the same results as 30P: 650 lw/ph in both horizontal and vertical.
Low Light Performance* (14.27)*
Our low light testing takes place in three stages. First, we shoot the DSC Labs ChromaDuMonde chart at an even 60 lux and 15 lux, then compare to the results with the competition. As mentioned above, the expectation is for a decreased performance compared to the last year’s Canon AVCHD models because the imager size has been reduced and the pixel count has increased. But we have been surprised before.
At 60 lux, the Canon HF10 looks good, but not as good the HR10, which has the larger 1/2.7-inch CMOS. The HF10 clearly has a higher resolution, due to the higher pixel count, but the noise is more obvious. The color retention is better on the HR10, as well as the HV20 and HV30, all of which share the same larger imager. Interestingly, our out-of-lab shooting indicates that sometimes the HF10 appears sharper than the HV20, so it can depend on the situation and the dominant color tones in the shot.
Switching over to 30P mode, the improvement is obvious. The overall exposure is higher and the colors are more vivid.
24P mode is even better, obviously, though the trailing that occurs in low light can be distracting.
At 15 lux, the Canon HF10 produces a lot of noise, as is the case with most consumer camcorders. The Canon HR10 doesn't look great at 15 lux, but it definitely produces less noise.
The second part of the low light test involves shooting an X-Rite Color Checker chart at an even 60 lux, then exporting frames to Imatest imaging software for analysis on color accuracy, noise, and saturation. We found the Canon HF10 produces a color error of 11.9, which is less accurate than the HG10 and HR10. The noise measures an average of 1.13 percent. This confirms what we were expecting – the new, smaller imager produces more noise than its AVCHD predecessors. The HG10 and HR10 produce an average 0.8 to 0.9 percent under the same conditions. Interestingly, the HV20 and HV30, using HDV compression, produce more noise than any of them. Finally, the saturation of the HF10 measures 74.86 percent, which is lower than any of the older Canons.
These same tests were performed in the camcorder's alternate frame rates. In 1080/30P mode, the HF10 produces a color error of 9.3, a noise score of 1.63 percent, and a saturation of 91.32 percent.
In 1080/24P, the camcorder produces a color error of 8.95, a noise score of 1.095 percent, and a saturation of 94.81 percent.
The third test measures sensitivity. We shoot the DSC Labs ChromaDuMonde chart while slowly and steadily lowering the light, until the camcorder can produce a maximum outputted exposure of 50 IRE (according to a waveform monitor). Shooting in 1080/60i mode, the Canon HF10 is able to produce 50 IRE at 10 lux. The HG10, HR10, and HV30 are able to do the same thing in only 8 lux, which again demonstrates that the smaller imager is having a negative impact on low light performance.
The sensitivity test was also performed in 1080/30P and 1080/24P. In 30P mode, the Canon HF10 is capable of producing 50 IRE at 6 lux. In 24P mode, it produces the same results with 4 lux of light.
The HF10 is equipped with Optical Image Stabilization (O.I.S.), the preferred form of shake reduction in the camcorder world. OIS functions by separating the lens element from the body of the camcorder. The less-effective Electronic Image Stabilization (E.I.S.) operates by creating a digital buffer around the sensor, sacrificing resolution.
We tested the competence of the HF10’s OIS using our custom-built shake emulator at two speeds—Speed One and Speed Two. Speed One simulates typical stationary handheld shake, while Speed Two’s frequency is more intense, generating a shake that is more akin to a light jog with the camcorder in hand. At Speed One, the HF10 exhibits a 75-percent shake reduction, while Speed Two yields a mediocre 51.85-percent reduction. The HV30 performs slightly better at Speed One, but both camcorders are nearly identical in terms of image stabilization. The Panasonic HDC-SD9 is currently the top performer, but we’ve also seen a pretty magnificent performance form the new JVC GZ-HD6.
Wide Angle* (9.6)*
We tested the HF10’s maximum wide angle using a vertical laser with OIS disengaged and the zoom pulled back to its fullest degree. Footage was then later interpreted on an external monitor to obtain a true reading. The HF10 displays a maximum wide angle measurement of 48 degrees, which is just two degrees shy from the HV30.