The 'flagship' model in the range got a physical kick forwards with the K7 in 2009, a new shape with the metal body you desire, a professional build really, and weather resistance. That body was continued with the K5 series from 2010 and remains with the K5II. The recent K3 maintains the high quality and weather sealing, but the body has changed shape somewhat and is very slightly bigger.

The K7 / K5 / K5II all make use of the same battery grip (BG-4), but the K3 has an new optional grip.

From a body / handling / quality point of view all of the cameras I've mentioned are top notch. The sensors and IQ step up from the K7 (14.5 mp) to the K5 (16) and now the K3 (24).

I think it is fair for me to summarise the general 'consensus' on IQ (if thats possible to attain at all), which would be that the K7 offers excellent IQ except for higher ISOs in low light (eg noise is an issue from ISO 1600+) - the K5 improves a lot on this and seems to be our Pentax users camera of choice for high ISO work - and there is some debate about whether the K3 has continued this performance, some have shown poorer high ISO perofmance on the K3 above 6400, and others have expressed great satisfaction with the high ISO performance of the K3.

Of course, there are some other recent bodies that also offer high IQ & low light performance, but I only refer to the 'flagship' options as you mentioned the build quality and coming from the LX.

Smeggy describes a batch process - that is only ideal on a 'batch' level, it is never (cannot be by design) 'ideal' on an individual image level - that is an 'optimised' workflow. It depends on how you operate.

I have thousands of images, but I don't apply a batch automatic 'pre-procesing' to them -I'm content for them to wait until I choose individual ones worthy of attention.

Anyway, batch processing NR based on an ISO variable alone is missing the most important aspect of the process - namely image content. The distribution and spread of different kinds of noise varies widely between subject types, and IMO demands an optimised approach.

Its a personal choice really. Why would I want to pre-process batches of images, the majority of which won't need to be used or seen anyway? Life's too short.

That was a very different thread - this noobie was some kind of lottery winner who said he always bought the best of everything now. The K3 was offered as that - then the K5II - then the K50 - and he ended up with the K500, which the two agreed at the end was 'the best'.

The question for us is - were they right?

Thanks Steve, I went a bit easier on the saturation.

Thanks Alastair also

Philip, in No 4 I had lost the DOF on the right side there. I was pushing it with this lens really (non-macro), handheld at least - it was right on the close focus limit for all these shots, so my own slight body movements were pulling focus in & out. But there was enough of a breeze to make slower shutter speeds a worry as well - a good challenge!

I was being a bit tongue in cheek there Rob Of course, I don't know whats going on 'under the hood' - we can only go by the results we can see. The issue of processing time here is surely only of real significance to a batch-style workflow, where an accumulation of time multiplied by a high number of images could really slow you down - if time is really that important and you couldn't get on with anything else while you wait.

But Topaz DeNoise is really for an 'Optimised workflow', meaning special, unique and maybe multiple processes for one simgle image. Here the aim is to perfect and tailor the adjustments exactly to each unique photo, so I wouldn't be rushing through all that anyway. Topaz takes about 40 secs on my machine which I find perfectly OK.

How do we know if the slowness is being caused by the software being badly written, or just the toll of all the number crunching being taken on the computer?

Not extravagent for him personally - but thats different to it being ideal, assuming the aim is to quicky be able to take great shots and get plenty of enjoyment out of doing that.

Maybe our assistant was too quick to make assumptions about our mans capacity - or maybe his capacity is not relevent? He could just ignore the bits on the camera he doesn't understand, and hopefully learn more over time, sort of 'grow into it'.

The alternative I suppose would have been to try and convince the customer that all the features he didn't understand were actually a good thing for him personally - like they'd help him get better pictures when he did understand it all, so better to buy the expensive one and then he won't have to upgrade so soon.

I mean, should the aa filter simulator have been the rejecting factor for the K3? Is it really just one extra exotic feature that is not good for a 'noobie' ?

Well, I always figure that if its taking a long time then it must be doing some serious stuff! If batch processing and speed are a priority, then a solution like Lightroom is going to be best - Topaz DeNoise works on a single image at a time, although it's settings are stored from the previous use when you reopen it again so can be reapplied with one click to a new image.

The thing about Topaz really to consider, is that it works as a Photoshop Filter plug-in - so the most 'intelligent' way to use it is on a duplicate image layer ... so you duplicate your background layer, then work in Topaz and apply the filter onto the upper layer. Then you can make use of all the great flexibility of layer techniques .... you can use a layer mask to precisely apply the effect to exact areas, and you can also use the opacity of the upper layer to allow the unfiltered original underneath to blend through, say to bring back a touch of the original texture and noise grain, to help with your detail restoration.

In these ways it is a most flexible tool, and so the options for finely controlling the effect after it's applied are probably far greater than in Lightroom.

Here's my standard workflow in Topaz - give it a go and see if it works for your images : these settings are start-points :

Once the plug-in is open, click on Reset All - this will get you to a zero start point.

1. Set 'Overall strength' to 0.07
2. Check (at 100%) midtoned and brighter smooth areas like skies - any fine luminance noise should be gone now
3. Check deep shadows and move 'Adjust Shadow' to 0.20 ... this will blur these areas for now
4. Look closely at the colour of remaining noise patterns - most common is patchy yellow, but you can also sometimes see blue or red specks. For yellow pull the 'Clean Colour' slider to about 0.30. This will reduce the yellow noise and slightly pull down yellow tone overall. Check again for any blue or red specks and pull up on the appropriate slider (often most blue has already gone in the shadow adjustment)
5 Now set 'Recover Detail' to 0.20 and 'Reduce Blur' to 0.15. Look at the texture coming back in and then just balance back and forward with the 'Overall Strength' slider and the detail ones.

Its always a compromise to an extent between smoothness and detail - thats where the layer masking comes in to fine tune things around the photo.

Its terminologies really, isn't it? There is a difference I think - what we consider 'image' files are 'rastered' into a pixel based bitmap type - that can be seen directly on a display device. RAW is not like this.

The JPEG preview surely is created simply in order that we can actually see what the RAW image looks like in order to manage and organise it?

Its hard to know without seeing exact settings - real comparisons are tricky IMO unless the users skills and experience are similar on both softwares - I certainly wouldn't expect to get good NR results with Lightroom myself.

For what its worth, my normal main strength slider setting in Topaz DeNoise is about 0.5 or 0.6, very small really. To this I add a touch of shadow, blue & red colour NR, and a good dose of the main 'reduce colour noise' slider.

Thanks Don, you're always very kind