i have the LX1 and i only shoot raw. i also rarely go over iso80. iso200-400 is hardly usable but for snapshots it is ok.
but even iso80 shots can be quite noisy, especially shadow-noise is a problem and it contains lots of uncontrolled color-noise.
the fine thing is, that with raw you have much more control over the noise. i prefer to shoot raw and then denoise it with some software afterwards, the result will always be much better than the original jpeg.
i know from my own experience that the LX2 sensor is slightly less noisy than the LX1 sensor, but this can be a result from some slight noise-reduction applied already on raw-files. especially shadow-noise is blurred in the LX2.
JPEG also look different from both cameras, as they have different VENUS processing engines. but i wouldn't take neither as RAW always gives me better quality.
i attached you a 100% crop from my LX1. on the left is the original jpeg, on the right is the raw processed with ACR4.0 and some neatimage applied for noise reduction.
have a look at the wooden structure. it is much better defined in the raw-version.
also look at the ring-binder in the background. the jpeg contains lots of color-noise. this is even better visible in the iso400-shot.
ISO80:
ISO400:
even at iso80 some slight neatimage processing can boost image quality a lot. i never filter all the noise, but rather half of it, so that the image still looks natural unlike an aquarell painting.
most of the time i use ACR4.0 for raw-processing as i find that 4.0 makes some natural grain-like noise out from LX1-files, whereas the newer version 4.3 has some other demosaicing and stronger noise-reduction which you cannot turn off. this way the noise can become quite a mess. therefore i still prefer 4.0.
i also tried other raw-converters and they produce different noise-structures. raw shooter had some block-structure, bibble and raw therapee produce much finer noise, which can be hard to deal with.
my former camera was a canon G5 which most of the time was also claimed to have a noisy sensor.
i could make some comparisons between the G5 and the LX1, and the LX1 was not much noisier than the G5 of course with RAW-files as basis.