Sony DCR-TRV950 Camcorder Review, number 2

The DCR-TRV950 has been reviewed and dissected by most, if not all, of the mainstream camcorder press with nearly unanimous praise. Unfortunately, many of these sources tend to state the obvious and don't explore the nuances of the products they review. They also fail to compare products to other competing models.

The first review of the DCR-TRV950 was by Robin Liss, editor in chief of Camcorderinfo.com. Her's is an excellent review spelling out all of the specifications, features, and functions of this camera. Robin's review concludes with an insightful opinion of the video quality, audio quality, and ease of use for the TRV950. I want to go beyond what Robin said in her first review and look at what we've learned about the DCR-TRV950 in the past six months

Background and Specifications

The DCR-TRV950 has 3 CCDs producing quality colors. Each CCD is 1/4.7 inches in size, with 1.098-mega pixels maximum resolution (680k max for video, 320k used for image stabilization, and 1 MP available for stills). The CCDs are labeled as "Super HAD". What this refers to is hole accumulation diode, which is an advancement in CCD design whereby multiple micro lenses overly the sensor to focus light and images onto the device with higher sensitivity by increasing the virtual aperture area. The "super" part refers to a series of micro lenses that are finer in construction and closer spaced to each other to diminish the ineffective area on the sensor created by the spaces between the lenses. Sony claims an increase in sensitivity from roughly 2 mv/um2 to 50 (compared to an "average" HAD CCD). While this seems logical on paper, the real and perceptible difference achieved is much less clear. It is best to judge the performance (in this case, light sensitivity) in real world use, or tests, as opposed to the theoretic benefits of this design.

The lens specifications for the TRV950 are 3.6 to 43.2 mm with an f range of 1.6 to 2.8. The lens thread diameter (for filters or add on lens accessories) is 37 mm. Sony has not gone with the designer "Carl Zeiss" lens with the 950, and no special claims are alluded to in regards to the lens. The lens is adequate and appropriate for a camera in this range, but not exceptional. This is not to say that any camera with a "designer lens" is beyond reproach, neither is it to say that the absence of designer type lenses is a strongly negative factor, they are often over hyped.

The forward mounted, on camera internal microphone, does leave something to be desired. Some noise is transmitted from the camera works to tape (as is the case with most on board microphones). On a positive note, however, external accessory mics are readily available as an upgrade. There are two ways to hook the mic up; first with a Sony brand mic via the smart shoe, and second, via the microphone input jack (although mini) located up front. Also nice is the ability to set for 12 or 16 bit audio manually.

Video performance specifications are listed as 680k pixels net for video, 1.098 MP resolution maximum for stills. Of importance is that video output is limited to 530 lines of horizontal resolution (as is the case with almost all video cameras). Making sense of all of these numbers is not so easy at first blush, however, it does boil down to a simple principle. The best video performance will be limited to the weakest link in the chain. In this case, the 530 horizontal lines of video resolution is that weak link. Strictly speaking, this is the top resolution, and all the pixels in the world will not raise this bar. Hence the theoretic maximum of 680k net pixels as opposed to 1.098 MP times three as arithmetic might suggest. The extra pixels, however, are utilized in the image stabilization function, which is the best I have seen thus far. While all of this might sound depressing, it is not entirely bad news as this refers strictly to maximum video resolution, not color accuracy, or saturation which is superior in 3 chip cameras in general, and extraordinary in the 950, in particular.

Minimum illumination for the 950, stated in lux, is 7. Notice that this number is higher (or worse) than the TRV900 which the 950 is replacing. Sony made efforts to explain this away by boasting about increased sensitivity with the "Super HAD CCDs" that are used. This issue has raised the most controversy and debate over the 950 so far.

Another feature of the 950, worthy of further discussion is the zoom's toggle control. In my opinion, this is the 950's Achilles heel. The problem being that the control is overly sensitive without a gradual transition. I find this analogous to the difference between analog and digital controls (this is one instance where digital is not a good thing). This seems to be an oversight, or mistake, on Sony's part, as this would be very easy to remedy, preferably when in development, before the final product had reached the market. Fortunately, add on auxiliary controls can be readily had, and solve this problem. Unfortunately, these add on's are bulky, costly, and therefore detract from the otherwise compact nature of the DCR-TRV950.

An exciting feature of the 950 is the availability of a wide screen 16:9 ratio mode for recording. As the video industry is marching in this direction, this is a wonderful option to have, particularly if high-level amateur, or semipro, filmmaking is the intendedf use for this camera.

Controversial on its' face is the addition of Bluetooth wireless network capability to the DCR-TRV950. While this is a neat wireless connectivity protocol, I am not sure what useful purpose it could serve in this, or any other digital video camera. The purists, myself included, have no real use for it, as it is too slow for transferring video data. It does not add to the quality of the video, or the ease of use of the camera. I personally think that it is a useless feature but I haven't had a ton of experience with it. If you are interested in the Network features, take a look at Robin's review.

The other features and controls of the TRV950 are very well described and explained in Robin's review so I won't cover them here.

Resolution Truly Defined

I want to review some principles of resolution because there are a lot of misconceptions to be dispelled. Factors important in terms of useful resolution measurements are:

1. Effective, noninterpolated CCD resolution in pixels.
2. Recording mode, either progressive or interlaced.
3. Resolution capability of viewing device or monitor.
4. Presence, or absence of a scaler, or line doubler, inboard on the viewing device.
5. Conversion formulas from pixels too lines (progressive scan has higher pixel counts by corollary, compared to interlaced), too X and Y pixel parameters (example: 1280 X 1024 refers to a single image resolution, equivalent to 720P in moving picture format, or terminology, and calculates to 1.3 MP resolution for stills).
6. Frame rate, which differs for PAL and NTSC standards. Film is rated at 24 frames per second. Progressive scan recording are generally rated at lower frame rates (15 fps is typical) which makes for a jerky appearance to video that results, but allows for capture of higher resolution stills and better slow motion effects. These misconceptions come from a lack of understanding of these principles and what the upper limits, or maximum capability, of currently available monitors are. Ten years ago, we would confidently state that the best possible resolution would be 480 progressive scan as, outside of prototypes, this was the best available. Resolution output higher than this would be wasted (although many manufacturers add about 10 percent above the maximum for comfort, I guess, hence the magic number of 530). Today, we are still limited by this upper limit as evidenced by the maximum stated resolution for video stopping at about 530 lines. Some cameras, including the DCR-TRV950 have sensors with resolution capability higher than this limit, however the extra resolution is used for purposes other than video resolution like still mode and image stabilization.

So what does all of this mean to us, the consumer? Well, it means that beyond a certain limit, higher resolution does not mean what you think it means (and also, btw, what the marketing people would like you to believe it means, as this is how they sell more cameras). So, the take home message is look at the resolution specs with something of a skeptical eye when deciding on a camera. It is certainly not hard to imagine that a camera with the highest numbers could have a terrible picture, and conversely, one with more modest numbers might well have a superior picture. Use your eyes, not your brain when assessing picture quality.

All of this will change, probably within the next ten to fifteen years, once high definition penetrates this market. Once this happens, there will be an evolution in this industry that parallels the one that will occur in broadcast television. For those who have seen HDTV, it is a quantum leap improvement over standard definition. On the down side, all of the current equipment will pale in comparison and become obsolete. On the up side, the video quality will be spectacular. Don't worry now however, it will be, conservatively, 10 to 15 years before high definition will be available, and affordable for consumers as there are numerous obstacle to implement a conversion to high definition. Format wars, copy protection, compression codec wars, and need for a much higher capacity storage media are just a few of the obstacles to be overcome.

The Low Light Performance Debate

The most prolific criticism of the DCR-TRV950 has been in regards to the low light performance, and minimum lux rating of 7, worse than the DCR-TRV900s rating. While the extreme low light capability of this camera is slightly lower than some of the top competitors, the critics have blown this issue way out of proportion.

The rationale for the lower performance in extreme low light conditions stems from the smaller size of the CCDs (1/4.7 inch), which will result in lower light sensitivity when compared to a larger CCD of the same design. Sony, however, claims to have addressed this shortcoming by virtue of its' "Super HAD CCD" (see description above). However, my feeling is that the real world performance in extreme low light situations is, at worst, slightly noticeably less than the best performers available (i.e. the VX2000), but, with some adjustments to shutter speed, not perceptively different. Even more obvious, and to the point, even if it were a major problem, it is very easy to correct by adding external light or adding a camera mounted accessory light and giving forethought to the setting in which you are shooting (which we all should be doing anyway). Ultimately, the proof is in the pudding, prospective buyers of this, not inexpensive, camera should be open minded, try it out, and make the appropriate changes in the settings. Then, and only then, can you make an informed and educated decision for yourself.

Still Picture Mode, a Neat Gimmick

A lot of time has been spent in discussion about the still picture features of newer digital video cameras. In fact, the manufacturers have made this feature a cornerstone of their marketing efforts. For reasons that elude me, this is apparently an important element in the consumer's decision-making process when shopping for a digital video camera. Even more surprising is that this is a feature of importance in the higher end prosumer digital video camera market as well.

The issue needs to be put in proper perspective to be better understood. While this is a nice gadget feature, it is not comparable in performance to dedicated digital still cameras and should not be considered as though it is. The TRV950's still picture maximum resolution is about 1MP. This is adequate for prints up to about 4 X 6 only, and is comparable to digital still camera performance of 10 years ago. Today's digital still cameras are clocking in at 5.0 mega pixels for about $800. With these cameras you can print 11 X 14 size photos indistinguishable from film cameras. As such, monetarily, this feature is worth, maybe, 50 dollars (or how much it would cost to acquire a still camera of 1 MP resolution today). This leads to my final argument on this topic, raised in the form of a rhetorical question; would you make a decision for purchase of a 2500-dollar digital video camera based on the merits of an out of date, 50-dollar gimmick? I wouldn't.

Conclusion

I find the DCR-TRV950 to be an exceptional product that, on balance, provides the highest quality video, and most enjoyable experience of use of any currently available camera in its' price range (and maybe beyond). That being said, the TRV950 is not without flaw, as no such product ever could be.