. (I suppose the Samsung is optimised for reducing MPEG-2 and MPEG-4 artifacts and noise, and not the type produced by the H.264 of AVCHD.)
As I mentioned, I'm not intending to deride the X900, but rather to establish from those experienced in current machines what I can reasonably expect now from a camcorder in this general category/price area. The Canon XM2 was a prosumer SD camera that recorded to mini-DV. If I remember rightly, the bitrate of the mindDV codec is 15Mbps.
By contrast, the X900 is recording in HD (four times as many pixels) at 24Mbps. Obviously, if we were comparing two machines of the same vintage, the comparison would be idle.
However, the XM2 was released at a time when the highest-end pro DSLRs had 4 or 5 MP sensors, and when processors were many times less powerful and slower than current chips used in cameras/camcorders, and RAM very expensive (think of the spec of a typical c.2000 PC versus one from 2012 - a huge step change in power and speed). Additionally, I believe the AVCHD codec (and indeed MPEG4) is more efficient than that of mini-DV.
For all these reasons, I'd assumed that a top-end consumer (or prosumer?) HD camcorder released in 2012 would surpass the old XM2 is all areas. And this assumption had been underlined by various rave reviews on camcorder/camera review sites, which wrote of the X900's excellent low-light performance. However, this is all too evidently not the case in (even relatively good) artificial light, even though the X900's performance in daylight outside clearly surpasses that of the XM2 in most (not quite all) respects.
Earlier today I did some tests in manual mode, in a living room lit by reasonably good exterior daylight. I used the X900's lens at its widest angle so that the aperture could be at its maximum. I found that the X900 was having to use maximum aperture (f1.5) and about 6dB of gain to boot. In the same light, the old XM2 was needing to use no gain. So this is at the root of the problem - the X900 is having to apply substantial gain (high ISO in still camera parlance), which inevitably results in noise in the post-sensor amplified signal, with noise suppression processing then resulting in the predictable visual artifacts. And the X900 was using 1/50 as the shutter speed (I hadn't switched on the option to use 1/25, since this is really of no use when there's any movement).
The XM2 also has three sensors, so each of those sensors is receiving a third as many photons as enter the lens (as with the X900). This aspect doesn't therefore explain the wide disparity in performance between the two camcorders in other than reasonable daylight outdoors. In still camera terms, the sensors of the X900 appear to have a very low base ISO, meaning that they are relatively insensitive, such that substantial gain is needed relative to sensors that are inherently more sensitive. Given so much advance in sensor technology over the last 12 years, this also puzzles me, not least since Panasonic makes very decent sensors for its micro four-thirds still cameras.
It might be that there are good reasons for all of this. As stated, my intention is not to criticise the X900 in any absolute sense but to enquire of other members whether the other two cameras in this category - the Canon Legria G10 and Sony CX730 - are likely to give substantially better performance in lower light conditions. I've set out my problems with the X900 in this area so that any member with the relevant experience can assist. Or perhaps a member has been through similar travails and found that only by spending substantially more can an HD camcorder work well in decent artificial light - any wisdom along these lines would be of much interest.