Image Quality (tentative first impressions)
It is too early to make any definitive statements about the S100's image quality, partly because we have a lot more testing still to do ahead of a full review, but also because the camera sitting on my desk in front of me is the third sample that I have looked at, and the third with what appears to be a slightly decentered lens. Quality control issues can affect cameras from early production batches (and inevitably, review samples tend to fall into this category) and I hope that we have simply have been unlucky with the samples that we have received so far.
This isn't simply blind optimism (nor is it brand bias). No man is an island, nor any website. I am reserving judgement on the S100's optics for another reason: I have seen images from the S100 posted by owners and trusted competitors which appear better than those that we have been able to capture. At the time of writing, we are expecting another sample from Canon and we will rush it into the studio and out into the real world as quickly as possible. If necessary, our previously-published studio samples will be replaced, a fresh samples gallery created, and progress on a full review resumed (containing, of course, a full account of our experience).
Despite what I hope is only a temporary setback, I am able to make some judgements about the S100's image quality - specifically as regards its resolution and high-ISO performance. The lens issues that we've experienced with our early S100 samples are noticeable in the critical environment of our studio, but as usual are less problematic in 'real world' day-to-day shooting at a range of different focal lengths and distances. That said, pending further testing, the images on this page should be regarded as preview samples for now.
Looking at samples from the S100 next to those from its predecessor the S95, and cameras like the Canon PowerShot G12 and Nikon Coolpix P7100 it is clear that the S100's 12MP CMOS sensor does genuinely deliver more detail. Across its entire ISO span, the S100 out-resolves the S95 at equivalent settings - not by much, but it is noticeable. At their respective highest ISO settings the gap is narrower, but the S100 is certainly no worse than its predecessor at ISO 1600 and 3200 and - crucially - its files are captured faster, and the camera is ready to shoot again more quickly. The S100 is fairly noisy at ISO 6400 but like the S95, the S100's noise reduction treads a sensible line between noise smoothing and detail retention, and in a pinch I'd be perfectly confident making postcard or letter-sized prints at this setting.
Although we have to reserve final judgement for now, I'm encouraged by what we've seen from the S100's homegrown CMOS sensor in our preliminary shooting.
Thoughts so far...
With the S100, Canon has updated a popular design without apparently meddling too much with the ingredients that made the S90 and S95 so popular. Owners of either camera will have no trouble picking up and using the S100, and even first-time buyers should find the S100 easy to get to grips with. There are raw-shooting enthusiast compact cameras on that the market that offer finer-grained control over their settings and operational ergonomics, but those custom functions that are offered (like Control Ring and RING FUNC button customization) are sensible and well thought-out. I'm not sure whether the addition of GPS or an uprated video capture mode will tempt many photographers to trade in their S90 or S95, but if they do, they'll find that they've upgraded to a camera that at the end of the day produces slightly better critical image quality, slightly faster than they're used to.
My only concern about the S100 at this point (and remember I'm writing this before we have been able to carry out our full gamut of studio and real-world tests) is the camera's lens. The three cameras that we have used have exhibited optical characteristics consistent with slight decentering of their lenses. Naturally, as I've already noted, this is much more noticeable in our studio image samples than in most 'real world' shooting (can you spot it in the images on this page? I doubt it). Depending on the focal length, focussing distance and scene though, it can be detrimental to critical image quality. I hope that we've simply been unlucky with the samples that we've been sent so far. Watch this space.