JVC GZ-HD3 Camcorder Review
A CamcorderInfo.com unbiased, scientifically tested review of the JVC GZ-HD3 high definition hard disk drive camcorder.
Video Performance* (7.75)*
The JVC GZ-HD3 shares some of the same imaging system with its big sibling, the GZ-HD7. Both have three 16:9 1/5-inch CCDs, each with a gross pixel count of 570,000. The effective pixel count has been reduced in the GZ-HD3, however, due to the loss of the Fujinon aspheric lens. The GZ-HD3 instead uses a Konica Minolta lens, which hits the sensors in a different way. The effective pixel count is 430,000 (down from 530,000 on the HD7). All JVC three-chip camcorders use pixel shifting – a method of off-setting one of the chips – in order to increase total resolution.
The overall video performance was similar to the GZ-HD7 in most respects. At a casual glance, even a good look, the GZ-HD3 and GZ-HD7 would be hard to tell apart. When we looked very carefully at areas of fine detail, however, the GZ-HD7 clearly had the sharper image, likely due to the better lens. Color performance was essentially identical.
Once again, we found the GZ-HD3’s video quality to be somewhere between HDV and AVCHD. The GZ-HD3’s image lacked the resolution of both the Canon HV20 and the Sony HDR-HC7. There was also a clear increase in compression artifacts from the GZ-HD3. However, it looked slightly sharper than the Canon HG10 and much sharper than the Sony HDR-CX7.
Color performance was sometimes a weak area for the GZ-HD7 in our Great HD Shootout with the top HD camcorders. The GZ-HD3 performed pretty much identical to the GZ-HD7. Under perfect indoor shooting, the colors were saturated (no surprise for a consumer camcorder), but the colors also tended to skew toward green a little. We white balanced with a few different methods, and they all came up a little green. In mixed indoor/outdoor lighting, the camcorder did not fare well, though it did a better job that most Sonys under the same conditions. Outside, the JVC GZ-HD3, like its predecessor, warmed skin tones too much.
One thing that troubled us most about the GZ-HD3’s image was the poor rendering of whites. The camcorder showed a proclivity to push the whites too hard, blowing out certain channels in areas where very few camcorders had trouble. For instance, at 3,000 lux the DSC Labs Chroma DuMonde chart should not be blowing out. This did not appear to be an issue with oversharpening – some of the other tell-tale signs were missing. The whites were just too bright.
The camcorder also offers a Sharpness control of +/-5. Boosting the sharpness pushed the in-camera sharpening to make edges look sharper. (It does not actually increase resolution.) When we pushed the sharpness up to +5, the fine grain noise increased a great deal, more than it was worth for the minor increase in perceived resolution.
At a sharpness of -5, the image actually looked out of focus. This is not a setting we would recommend for general occasions.
Overall, the GZ-HD3 will not disappoint casual shooters. The image looks sharp and the colors are acceptable under most conditions. But the discerning videographer will find the small problems – notably, blown out whites and overly warm skin tones – add up to an image that can be improved upon by other camcorders.
Video Resolution* (13.125)*
The video resolution is tested by shooting a DSC Labs video resolution chart and watching playback footage on an HD monitor. This test measures the final output of the resolution, rather than the advertised resolution of the chip. We use the highest quality output the camcorder offers. In the case of the JVC GZ-HD7, this was the HDMI. We found the camcorder was able to produce a horizontal resolution of approximately 525 line widths per picture height (lw/ph) and a vertical resolution of 500 lw/ph. This was the same resolution score as the GZ-HD7.
Low Light Performance* (2.59)*
Our low light testing is a multi-stage affair. First, we shoot the DSC Labs Chroma DuMonde chip chart at an even 60 lux and 15 lux. At 60 lux, the JVC GZ-HD3 managed admirably at retaining color information. The 1/5-inch CCDs are certainly not the largest in the world, and bigger chips have done less. The noise, however, picked up a tremendous amount. This is not the finest grain noise, and it will be quite visible on a big screen TV. Combined with the compression artifacting, it was visibly detrimental to fine detail.
At a close look, the resolution of the GZ-HD7 is higher than that of the GZ-HD3. However, at least at this light level, the lack of an aspherical lens did not handicap the GZ-HD3 in terms of its low light performance. Were we shooting a target from farther away, the case would not have been the same. At 10x zoom, the GZ-HD3 has a max aperture of f/2.4, while the GZ-HD7 has a max aperture of f/1.9.
By comparison, the Canon HG10 (in 1080/60i mode) was darker, but much less noisy. Less noise meant better resolution in the fine details. It seems likely that part of the GZ-HD3’s problem was allowing the auto gain (AGC) to go too high. You can turn the AGC off altogether, but that makes it too dark. While we would have appreciated a slightly brighter image, the Canon HG10 looks better.
The Sony HDR-CX7 was noisy, as well, but the noise was of a much finer grain and easier on the eyes. Surely it would be visible, and even distracting, on a TV, but it has a less deleterious effect on sharpness. The Canon HV20 had similar fine grain noise patterns to the HDR-CX7, though the overall resolution was sharper. Also, the HV20 has the option for 24P recording, which increases sensitivity in low light. (The Canon HG10 also has 24P, but testing showed the video in this mode to produce too much stuttering).
The JVC GZ-HD3 does offer a few tricks to help with low light shooting. The best choice is probably shutter speed control. When the shutter was lowered to 1/30 in 60 lux, the image improved a great deal. Having more time to gather light, the auto gain could be lowered, thus reducing the noise. The reduced noise also meant sharper video. Of course, you’ll have to balance this against the slight trailing that may occur with a lower shutter, but overall we liked this image better.
You can also turn off the AGC altogether, though it’s not recommended. As you can see here, the camcorder loses too much information to make a useful picture.
At 15 lux, the GZ-HD3’s image was quite dark and noisy. Some detail and color can still be made out, but this would not be considered usable footage even for high school-grade productions. The GZ-HD7 did appear to have an advantage here with its aspherical lens, producing a brighter image in the same light.
The Canon HG10, Canon HV20, and Sony HDR-CX7 all have a larger imager, which was likely why they too were able to produce brighter images at 15 lux. Among these, the Canons had the most accurate colors.
When the shutter speed was lowered to 1/30, the image improved somewhat, but not enough to really save the shot.
At 15 lux with the AGC off, the image was entirely black.
The second segment of the low light test involves monitoring outputted IRE levels from the GZ-HD3 and lowering the light until the camcorder achieves a peak 50 IRE. This gives us an idea of the camcorder’s sensitivity. The GZ-HD3 was able to produce 50 IRE at 24 lux. This was not as good as the GZ-HD7, which could pull off the same trick at only 17 lux.
Finally, we shot a GretagMacBeth Color Checker chart at 60 lux and ran frames through Imatest imaging software to determine color accuracy, saturation, and noise. At best, the camcorder was able to produce a color error of 10.4, with a noise level of 87.02 percent, and a saturation level of 0.885 percent. These scores are comparable to competing camcorders.
The GZ-HD3 is equipped with digital image stabilization (DIS). DIS functions by creating a digital buffer around the frame, which reduces the resolution. Optical image stabilization (OIS), found on the GZ-HD7 and Canon HV20, functions by isolating the lens element from the body of the camcorder. OIS systems are superior to DIS systems because they do not affect resolution.
We tested the GZ-HD3’s DIS system by using our custom-built shake emulator at two speeds. Speed one emulates a casual stroll down the sidewalk. Speed two is more along the lines of a light jog or bumpy car ride. The GZ-HD3 produced a 50 percent shake reduction at speed one and a 69 percent shake reduction at speed two. The GZ-HD7 produced a 75 percent shake reduction at speed one while the Canon HV20 took first place with an 83 percent reduction. However, at speed two the GZ-HD3’s DIS system appeared to handle fast motion better—the GZ-HD7 produced a 50 percent reduction while the Canon HV20 lagged behind with 36 percent reduction.
Wide Angle* (10.0)*
We tested the GZ-HD3’s maximum field of view by placing the camcorder on a tripod with the zoom pulled back to full wide angle. We then measured the left and right angles with a vertical laser. After computing the differences of the angle from a true 16:9 monitor, we were able to attain a wide angle measurement. The GZ-HD3 has a maximum field of view of 50 degrees, which is identical to that of the GZ-HD7’s. The Canon HV20 took the gold with a 52-degree wide angle measurement.