I’m evaluating 9.8.0 / build 671 on a Windows 11 Home four year old laptop with Ryzen 7 5700U CPU, a cheesy little embedded GPU and 15 GB RAM. Yeah, it’s a low end machine by photo editing standards, but it works well enough for my occasional editing. Slow-ish at times, but works.
I exported a handful of test images this afternoon but suddenly a few hours later it’s stopped exporting (even after an app restart). When I click Export the app briefly consumes 2% - 4% CPU then just sits there idling. No TIFF ever shows up. I’ve tried the custom profile below, the stock TIFF profile and the stock JPG profile. Nothing. It worked before, but no longer.
Well, I can maybe shed light on my own question. This sub-process appeared to be possibly hung (?) so I shut down PL and two other reasonably large apps (Firefox, Joplin) then restarted PL. It then finally exported.
Maybe a memory issue? But I had 6GB+ available at the time and it’s a cropped 24 mp image, so it’s not large. The final TIFF was only 91 MB. No AI masks were used, very simple edit.
??
If I have to shut down other apps, then that’s a hard stop showstopper. I realize 16 GB RAM isn’t huge by current standards, but NO OTHER raw processor I’ve used every had a memory shortage issue.
Unfortunately, I’d be pretty sure that your experience is related to configuration of your “low end machine” (as you describe it yourself).
PLv9 has a number of new features that are now really pushing PC resource requirements - and it’s not unusual for users to report experiences such as yours.
Fair enuff - - One other thing to consider, tho, is that PhotoLab provides the best RAW-processing and noise-reduction engine going around (IMnsHO).
Probably true, although darktable and ART / RawTherapee offer some real power and pretty deep options in that regard, at least for demosaicing and editing (albeit currently AI-free).
My initial impression is PL may be too automatic and lacking in operator control for my tastes. I rarely like the machine making decisions for me, in general. There’s no way a developer knows what I want to do. LOL I enjoy editing, learning techniques and manually enhancing images, in my modest way. Being a landscape photographer here in the landscape-poor deep south US (Louisiana) every shot pretty much HAS to be ‘pushed’ to make it even worth viewing.
But I need to spend more time with PL before deciding one way or the other.
That’s the separated PL export “thread” - usually called “DopCor” in this forums. You can “end task” (kill) the process - PL automatically restarted it. After export, this DopCor “stay” - and after (2 hour default value) a while, this DopCor automatically unloaded (free up memory).
I shut down PL and two other reasonably large apps (Firefox, Joplin) then restarted PL. It then finally exported.
Good the hear that. PL restart can help (and as previously i write: DopCor “kill” may also do the thing). With 16GB of system memory → the integrated Radeon iGPU also can share some (like 1GB, i guess that’s why you write 15GB).
The “denoising” (NR) what’s really use a memory in amount. DP XD3 use more than DP3, etc.
"NO OTHER raw processor I’ve used every had a memory shortage issue.
Apples to Oranges. RAW processors not the same, not do things in the same way, noise reduction (quality) not the same, etc.
You find out the way. Once happen, not means always happen.
Computer memory isn’t just RAM. How big is your pagefile (swap file)? An old rule of thumb is to make sure it can grow to 3x the size of RAM. Do you have a lot of free space on your C: drive or whichever drive is being used for swap?
It auto-managed by Windows. I used to chase pagefile size but after two decades in IT I finally gave up and on balance it’s worked out. I’ve got ~170 GB free on the SSD so it’s good.
Good to know.* What made me think that was this statement from the users guide:
“We think DxO PhotoLab, and not you, should analyze each image to determine whether corrections are necessary and apply them.”
IMO that’s a creative decision, not a technical one. I would rather start from neutral – Or a custom defined starting point instead of from an image that’s already been processed. I realize demosaicing and such require some initial judgement calls but I like them to be as minimal as possible.
Not to totally derail this topic, but a few things that would be great…
Parametric masking, in particular selections defined by manual choice of chroma, hue, saturation and luminosity control (i.e., color similarity)
Boolean mask interop
Reusable masks (linked, saved to disk, etc.)
Local contrast curves (i.e., small / mid / large detail enhancement). I realize LC can be approximated with large radius USM, but it’s not ideal.
Metadata batch ops
Ability to click on a point in an image and assign that color / luminosity / saturation, etc. as a reference point for editing
Selection of highlight reconstruction algorithms
Not strictly necessary but a choice of tone mappers would be nice
Tone equalizer
Macros
Unbounded, scene-referred editing
Selection of sharpening algorithms (USM, RL Deconvolution, hi-pass, etc.). Hi-pass would infer layer blend modes, though.
L*a*b* vs RGB operations
Flat field
Dead / hot pixel filter
Ability to set a global, automatic color space
Some of these may be there, I’ve just not had time to dive deeply yet.
* And good to be able to post as well. Such a low daily limit on posts for new users seems counterproductive.
I don’t know the context of that quote, but I’m pretty sure it would have been in relation to lens sharpness optimisations - whereby PL applies corrections automatically, using an “Optics Module” that’s specific to the body+lens combo used to capture the image, and specific to the focal length used, and specific to certain characteristics of each individual image … Not something you could readily achieve manually.
Essentially, there’s nothing that PL does that you cannot override or take control of for yourself - according to whatever preferences you may have.
I won’t go thru your list item-by-item, but - many of them (admittedly, not all tho) are already covered by PL … just not necessarily using the same terminology or exactly the same approach that you have described.
Your best approach (as I understand you to be doing) is to make use of the trial period to test PL out for yourself - Only you can decide if it’s the tool for you.
Yes, no doubt annoying - - but there was a spate of spam attacks a while ago that caused a lot of angst … I guess this is a way of slowing it down.
For most photo software developers, the answer is simple: they will provide you with a number of correction tools and then leave you to fix each of your photos one at a time. That is not our approach. We think DxO PhotoLab, and not you, should analyze each image to determine whether corrections are necessary and apply them. This leaves you to judge the resulting output and then adjust any of the applied corrections according to your own tastes if you wish.
One of the things I like in general is parametric masking – The ability to define and tune masks by incrementally tweaking parameters, observing the results.
For example, here’s (part of) how a color similarity mask is adjusted in another editor:
Each LCH slider is a pair: The top one sets the parameter value while the bottom one sets the amount of consideration that parameter gets during mask calculation. It’s possible to very finely tune masks that way (you can also combine it with other masks, possibly derived through other methods).
Please note, I don’t expect PL to do it the same way, it’s just an example I think is powerful.
But at any rate, I’ll see what I think… Just as soon as I can get some spare time!
The full quote from the User Guide explains this a bit better, Len;
========================================================
Quoting from the UG;
… … We think DxO PhotoLab, and not you, should analyze each image to determine whether corrections are necessary and apply them. This leaves you to judge the resulting output and then adjust any of the applied corrections according to your own tastes if you wish.
What DxO PhotoLab does to your images
If your original photo is in RAW format, DxO PhotoLab deals with it using the latest and best RAW converter. During the “demosaicing” process (which converts raw data from the sensor into red, green, and blue pixels), this converter produces an image that is free of artifacts. – Not something one can do oneself.
Without any user intervention required, DxO PhotoLab corrects four optical defects: distortion, vignetting, lateral and longitudinal chromatic aberrations, and lens softness. It does this thanks to a unique database containing years of information on the optical defects of photographic equipment (lenses and cameras) tested in our laboratories. – As I was explaining earlier
DxO PhotoLab subtly retains highlights, while preventing clipping, and even recovers certain details from areas that have already been clipped. We advise you to work in RAW to take advantage of the greater latitude in processing. – You can switch this off, if you want
Thanks to our laboratory measurements, DxO PhotoLab includes DeepPRIME, the unrivalled processing and denoising technology. – Yep !
DxO PhotoLab offers an exclusive correction to selectively improve contrast and brightness in the shadows. This correction’s effect is similar to fill-in flash, radically transforming underexposed, backlit, or extremely high contrast images. – You don’t have to use this - but it’s a powerful feature … See Smart Lighting
DxO PhotoLab is (s/be “has”) a unique tool for modifying the color rendering of an image to make it look as though you took the shot with a different camera, or were using negative film. – That’s an option … You have to actually choose to do this. Otherwise, PL renders for the characteristics of the camera that captured the image.
Essentially, the software automatically performs nearly all the corrections mentioned above, leaving you to just fine-tune or adapt the results to your taste. – Yep !
Here’s the thing… It’s not that I can’t nor won’t update this four year old laptop, it’s just a question of justification. I’m a hobbyist landscape photographer and due to where I live (downtown butt-ugly) I basically can’t get even semi-good subject matter unless I travel a few hundred miles at a bare minimum. So I end up with maybe two or three batches of photos a year, other than the odd desperation shots I take locally. That being said, it’s difficult to justify staying on the bleeding edge given that for every other purpose this laptop is functionally fine.
I’ve not tried today but later yesterday it was working.
It was at 2 as recommended but is now 1.
I already had DeepPrime disabled as I don’t need it. I compared High Quality off and on and the difference is so slight I disabled it.
Hopefully I won’t need any denoising (I rarely do). I tend to shoot the vast majority of the time at base ISO so it’s clean. If needed I can do it later in Affinity.
It was that way initially but even slower, so I changed it shortly after installing.
Just tried to apply an AI Sky mask and it borked: 0x00000050 bugcheck (i.e., BSOD), PAGE_FAULT_IN_NONPAGED_AREA
Driver? Memory? GPU? Most likely the driver and Windows given no errors otherwise with any other software. Yeah, I know AMD drivers aren’t the best and my versions are old but nothing I find reveals newer drivers. Given I’ve had no issues at all I’m not eager to potentially screw things up by updating even if I could.
At general, in “CPU only” everything need to be works - of course, in CPU performance as-it-is.
So, i guess, now you AI Acceleration set to something like: “AMD Radeon Vega”?
And what is “even slower” exactly?
May also worth to try the “Compatibility” vs “max performance” modes to see, what’s works better.
Not in the DxO.
I suggest to be more familiar you notebook and iGPU (integrated GPU in the CPU). How its behave, how you can change (if possible), etc. As iGPU use memory for VRAM purposes from your notebook 16GB memory.
Im not familiar for a while about iGPU stuffs - especially in notebooks, but i try my best to describe (please check notebook user manual, try google it, ChatGPT and similar)
Some case you find in you PC (notebook) somewhere in the UEFI/BIOS settings, usually under the ‘Advanced’ menu to change it. At least in Desktop CPU’s in the past.
In notebooks in this days its may does some automatic “sizing” thing, and may you not able to change manually. May some notebook use ‘fixed’ ratio. May some cases start form the ‘smallest block’ like 512MB. Some notebook (iGPU) manual settings is ‘locked’ from the user (however, may some hack works).
So, i suggest to dig in how you notebook iGPU works/behave (VRAM handling/allocation). And im pretty sure, its can explain a few things for you →
Example: if you use GPU (you iGPU) intensive applications, the memory usage may happen like: from your notebook overall 16GB → 8GB goes to iGPU, and the remaining 8GB use for system memory → and the remained 8GB may not so much.
That in the nutshell the iGPU stuff → not expensive, good generic → but has its trade-off’s, and sometimes not fully straightforward to understand how its works, how its behave.
May also worth to check the notebook memory expansion possibilities. May possible to expand it to 24GB memory → not uncommon, as usually budget notebooks has some soldered 8GB at base, and the additional amount is added as separated (replaceable) standard (notebook) memory module.
I hope its helps.
No worries. Usually the the export 2->1 change solve things like that.
Yes, mine had a 24 GB option when I bought it, but I see this as an opportunity to upgrade, particularly since my wife has already said she can use the old laptop for other things. As the old saying goes, “strike while the iron is hot.”
