[Class A]

@richandfleur
I think you should forget about in-body stabilisation for video and embrace the idea of using a small gimbal like the "Weebill S".

You Tube

You Tube

[Class A]

Here's what I was able to do (without a lot of effort) in DxO.

Thanks for sharing the comparison.

The DxO-processed image looks cleaner but also more "digital". Some areas seem over-smoothed and over-sharpened, particularly in the close ups.
I find that the other image has a more organic look to it and could be improved by changing the WB a bit.

All a matter of taste of course and in the full images the DxO artefacts aren't problematic; also different levels of adjustment would probably produce a less over-processed (in my view) looking image.

[zkarj]

Thanks for sharing the comparison.

The DxO-processed image looks cleaner but also more "digital". Some areas seem over-smoothed and over-sharpened, particularly in the close ups.
I find that the other image has a more organic look to it and could be improved by changing the WB a bit.

All a matter of taste of course and in the full images the DxO artefacts aren't problematic; also different levels of adjustment would probably produce a less over-processed (in my view) looking image.

I agree the DxO processed image does look a little more "digital" close in. The noise reduction was set to the default 40 (in a 0-100 range) in this example, so could be dialled back a bit. The other problem with this image is it isn't sharp across the frame. The nose of the aircraft is suffering from motion blur, for instance. But the noise reduction, in my opinion, is removing more noise with less adverse effect than LR could do by quite a margin. At full frame I quite like the result.

At the end of the day... you can't get more out of the capture than the camera gave you, just different interpretations.

[MarkJerling]

Not at the cost of Mark's one!

Cheapest K-1 in these parts!

[MarkJerling]

I tried revisiting one of my iconic (to me) photos using DxO PhotoLab 3. Their PRIME noise reduction engine is rather impressive. I know LR hasn't been well regarded in the NR department for a while, but I think this is an interesting comparison nevertheless.

The shot was taken 10 minutes after sunset. I had my 18-135 lens on (unusual for the circumstances), the aperture open to f/4.5, had pushed the K-5's ISO up to 4,000, and still the shutter was marginal at 1/160. In other words, there wasn't a whole lotta light and as I couldn't afford a slow shutter I had to push the other sides of the exposure triangle.

Anyway, here's the result I originally published, which was processed in Lightroom. I remember pushing the noise reduction slider too far and then pulling it back to get this. I told myself the noise in the sky added character.



You can see the full sized (3200px) version on the Flickr page.

Here's what I was able to do (without a lot of effort) in DxO. I reprocessed the photo with no reference to my original effort (until it came to cropping for this comparison). I just went with what felt right this time around, including using DxO's Smart Lighting and PRIME noise reduction. I'm not sure how to (or even if you can) view full sized attached images here, so I have also included a 100% cropped sample area of each side by side (I didn't perfectly match the crop of the full shot).

Alistair, that DxO processed image looks real nice and clean to me.

[zkarj]

I've done a little more work to illustrate the noise reduction of DxO on that sample photo. It's a little hard to tell with the downsized version I can show here, but the full version is available here. It's a JPEG at 100% quality made from the original PNG screen grabs.

[Kiwizinho]

Do you mean you are a prime user? People tend to be one or the other.
I was tempted of course but I have an 85 1.8 takumar ($300) and I would probably turn the AF off on the FA anyway. Cos that is another thing - you are either a MF guy or an AF guy -- two different discliplines.

I love the results primes give, but they can be a little inconvenient if your subjects aren't cooperative. I've taken some really nice portraits with my FA-50 1.4 also my Tamron 90 2.8 macro using it as a portrait lens.

I think a 50-135 2.8 is probably a similar price to that 85 1.8 on Trademe, but with the flexibility of a zoom. Of course if I had a K-1 I might think differently as I think the FA* 50-135 is APS-C only, although I already have a Tamron 70-200 2.8 which is a full frame lens, and probably equivalent to 50-135 on APS-C. The Tamron has very good centre sharpness even wide open, although edge sharpness falls off quite a bit even stopped down, so it's a nice portrait lens, but not so good for landscapes.

If the FA edge sharpness is good, that would be a good case for it over the zoom I have.

I don't mind MF although I prefer auto-aperture.

---------- Post added 11-17-19 at 09:03 PM ----------

Here's what I was able to do (without a lot of effort) in DxO. I reprocessed the photo with no reference to my original effort (until it came to cropping for this comparison). I just went with what felt right this time around, including using DxO's Smart Lighting and PRIME noise reduction. I'm not sure how to (or even if you can) view full sized attached images here, so I have also included a 100% cropped sample area of each side by side (I didn't perfectly match the crop of the full shot).

Impressive result.

[richandfleur]

I tried revisiting one of my iconic (to me) photos using DxO PhotoLab 3. Their PRIME noise reduction engine is rather impressive. I know LR hasn't been well regarded in the NR department for a while, but I think this is an interesting comparison nevertheless.

Do you use the mask slider when reducing noise in LR? In your case above I suspect you'd be able to isolate just the sky quite easily, if that's what was bothering you.

---------- Post added 11-18-19 at 10:30 AM ----------

@richandfleur
I think you should forget about in-body stabilisation for video and embrace the idea of using a small gimbal like the "Weebill S".

Tools for tasks. If I'm shooting something that requires this then I will. Otherwise, I want my Pentax to perform as well as it can in times when I haven't prepared in advance, or can't.
A bit like those times when you are not lugging off camera flash, stands and reflectors to take still shots away from the studio.

I notice the over use of slow motion in that video, which aides in the appearance of stabilisation. You achieve this with higher frame rates, then slowed down to a normal frames per second value.
Again you can't do that very nicely on a Pentax. It's interlaced at anything above 30 fps.
Basically Pentax is just miles behind in video capabilities, and it's been interesting to read surprised new K-1 converts hit this realisation, learning that their new modern camera can't do video any better than something from 2012.

I don't want to keep going over this though, it is what it is. There's a lot riding on the capabilities of the coming new K-3 replacement for me.

[Class A]

A bit like those times when you are not lugging off camera flash, stands and reflectors to take still shots away from the studio.

Sure, I get that.

On the other hand, how much "in a pinch" compromise is one ready to accept?

As Joe McNally once said (I'm paraphrasing): "Direct flash = no picture".
Sure, you can take a shot with a pop-up flash on the camera. Will it result in a picture worth having? In almost all instances the answer is "no", if you want to maintain certain standards.

Same with audio, a built-in microphone could work in a pinch, but in how many videos would you really want to incorporate sound recorded in such a manner?

While in-body sensor stabilisation can achieve very good results, if the rest of your footage uses a gimbal, I'm pretty sure the difference will be noticeable.

I notice the over use of slow motion in that video, which aides in the appearance of stabilisation.

The regular speed footage was very stable as well, though. I don't think he wanted to oversell the gimbal. It's a comedy channel and I watch his videos mainly for the entertainment factor. He makes good points about technical details now and then but that's a secondary (and unreliable ) aspect of his videos.

Part of the motivation for my post was to get some of you to chuckle about this guy. While he recently turned more into a "review" channel because he has been discovered by manufacturers as a potential promotion vehicle and I feel that the entertainment factor has suffered from him trying to do the stretch between being funny and giving the product a fair deal at the same time, I think he is still very funny at times. His older videos are funnier, though, IMHO.

The other motivation was to point out that there appear to be pretty compact gimbal options out there that you may not have been aware about. I thought you might appreciate hearing about such options. Apparently not.

Again you can't do that very nicely on a Pentax.

So it seems that adding in-body stabilisation on its own wouldn't be really useful then.

Basically Pentax is just miles behind in video capabilities, and it's been interesting to read surprised new K-1 converts hit this realisation, learning that their new modern camera can't do video any better than something from 2012.

I have to say I lack empathy for those who buy a tool and expect it to be usable for something which it wasn't primarily designed for. If they don't check in advance of the purchase whether the tool is capable to meet their requirements for this secondary function then they have no right to be surprised/disappointed just because "It's <insert current year>".

DPReview and others are trying to sell us the idea that every camera for stills photography needs to be useful for video as well.
I dispute that assumption and decidedly reject the notion of compromising on stills photography just to optimise the secondary use case.

Video cameras have different requirements and hence require different decisions regarding ergonomics, extensibility, interfaces to accessories, etc. It is OK when a stills photography camera gives one some video as a by-product -- Pentax once said about the K-7 video capabilities that they essentially just made it possible to record the LiveView feed -- but it doesn't make sense to expect getting a great video camera as well as a great stills photography camera without having to pay for it somehow.

Nowadays when "good video" appears to imply a soul-destroying (slight exaggeration here ) viewfinder experience, striping, and banding, I'm not in favour of inflicting all these downsides on photographers, just because there are companies that see a revenue/profit optimisation opportunity in selling compromise tools, rather than speciality products.

Even just making it more expensive for stills photographers to buy equipment because the hardware is designed to handle video as well is objectionable to me.
I'm aware of the argument that supporting video as well is thought to drive up sales numbers which results in a better mass production price/cost ratio. However, that assumes the video part is competitive with alternative offerings. In case it isn't, it could be better to increase the attractiveness of a camera (in terms of reduced price) for people that don't care about video in the first place.

I'm sure there are other cameras for still photography, say from Phase One and Hasselblad that are not very good at doubling as video cameras.
If people don't take issue with these cameras "being incomplete" why would Pentax cameras be expected to follow Sony's lead that assumes that every photographer is also a videographer?

I get that you are a hybrid user and I think anyone in your category has every right to want good video performance as well as good stills photography performance. I just object to the idea purported by DPReview and others that it is implausible for any camera to largely neglect video as it suggests that there aren't customers who want a camera that provides uncompromised stills photography performance.

There's a lot riding on the capabilities of the coming new K-3 replacement for me.

While I hope that the K-3 replacement will be convincing in the video department, I also hope that it won't use OSPDAF.

I don't think Pentax cameras should attempt to beat Sony / Panasonic cameras at their game.
Ricoh wouldn't succeed in the first place and they would just destroy the unique selling proposition they have left at the moment.
In my view, as soon as Ricoh will turn Pentax cameras into "also-rans" competing with the hybrids from Sony/Panasonic, etc., it will be the start of the end.

I'm not saying that Ricoh should neglect video performance but that they should try to achieve it without compromising stills photography, e.g., by using hybrid viewfinders, instead of EVFs, or using a DFD AF approach, rather than OSPDAF, plus they shouldn't increase the price for a stills photography camera just because good video requires the use of more expensive hardware components or additional expenses for video codec licenses or whatever.

[NZ_Ross]

Tools for tasks.

Yup, agreed

[richandfleur]

Class A, mate that's a huge post and makes some really good points.

I had just let the video discussion rest, as I'm just not that into the on going discussion whilst there's nothing new from Pentax to discuss.

I would like the return of sensor stabilisation and I'd like a better overall image quality from Pentax. However, Pentax aren't great at releasing new firmware for older models, and don't have anything new to improve on, so there's really nothing new to discuss.

Yes, some sony cameras are experiencing banding issues in certain conditions.
Not entirely sure that's video related. I've never heard of any such issues from the likes of the Pentax K-70, which also features this.

I appreciate the thought and link, sorry if if came off blunt. Am pretty exhausted at the moment and not looking to get into keyboard bashing at this stage.
Yes, whole camera gimbals, in fact gimbals in general have come a long way. A family member has a new spoon that absorbs tremor and helps him eat. The tech is out there and real.

[microlight]

I don’t shoot video very much, either using my K-3II or iPhone 6+; guess I just haven’t got the video ‘bug’. When I do with the reflex though, it always strikes me that it’s not really the ideal ergonomic shape for video whereas it is for OVF. As a very rare LiveView user, my usual stance for stills is the film-SLR wide-legged elbows-tucked-in OVF one (I’m sure you all know what I mean!), so that it then feels a little unnatural and unstable to then have to hold the unit out in front for video (or LiveView), especially when also manoeuvring the controls in manual.

The video’s entertaining, thanks for that, but even the smaller gimbal is pretty chunky. I have a gimbal for the iPhone and I rarely use that! 😂

[Class A]

However, Pentax aren't great at releasing new firmware for older models, ...

True, the K-1 received a number of firmware updates that even introduced new features but that was at a time when it wasn't an "old model" yet.

...so there's really nothing new to discuss.

Yes, let's hope the K-3 successor will give us something to discuss.
It seems a foregone conclusion that the AF system has been considerably revised. At the very least, there will be a lot more frame coverage than it used to be the case for Pentax APS-C cameras.

Let's hope they improved video capabilities as well.
I'm not sure why they even bother to add microphone input and headphone output to their current cameras, given the otherwise mediocre support for video.

Yes, some sony cameras are experiencing banding issues in certain conditions.
Not entirely sure that's video related.

The connection to video is the desire to support AF during video recording. The latter only really works well with OSPDAF. At least some OSPDAF implementations give rise to both striping and banding. I'm not sure the K-70 actually avoids these issues, I rather suspect it hasn't come under enough scrutiny. I'm not sure, though.

Panasonic's DFD system avoids the OSPDAF limitations and seems to be improving but isn't as good yet and may never be.

I appreciate the thought and link, sorry if if came off blunt.

All good!

[NZ_Ross]

Panasonic's DFD system avoids the OSPDAF limitations and seems to be improving but isn't as good yet and may never be.

Again it is about tools, and use case - I haven't found any issues yet with the DFD on my G9 - but I am not shooting high speed sports or BIF with it.

[zkarj]

Do you use the mask slider when reducing noise in LR? In your case above I suspect you'd be able to isolate just the sky quite easily, if that's what was bothering you.

No. I didn't really understand what that slider did. Which is another tick against it I suppose. I did not have to deal with two sliders in DxO.

I don’t shoot video very much, either using my K-3II or iPhone 6+; guess I just haven’t got the video ‘bug’. When I do with the reflex though, it always strikes me that it’s not really the ideal ergonomic shape for video whereas it is for OVF. As a very rare LiveView user, my usual stance for stills is the film-SLR wide-legged elbows-tucked-in OVF one (I’m sure you all know what I mean!), so that it then feels a little unnatural and unstable to then have to hold the unit out in front for video (or LiveView), especially when also manoeuvring the controls in manual.

Agreed. When I'm tracking a 300mph WW2 fighter whizzing by (for my still shots), the DSLR "elbow stance" ergonomics are excellent. If I were wanting to shoot video I'd ideally like exactly the same experience. Back in 1987 or so I took an on-the-shoulder Panasonic Video Camera (it took full sized VHS tapes) that had a EVF in an eye-piece to an air show at Ohakea and I am still gutted to this day that I mislaid or recorded over the tapes because, IQ aside, they were epic. Importantly, there are now professional outfits who SELL footage of air shows who in 2019 still cannot track their subjects as well as I managed back in '87. I'm pretty sure their problem is exactly the issue of ergonomics.

I did use LV for some stills just yesterday afternoon when I wanted a portrait of someone who would not smile if I was not smiling back at her, so being buried behind the camera was not an option.