Luminous Landscape Forum
Raw & Post Processing, Printing => Adobe Lightroom Q&A => Topic started by: davidgp on July 11, 2017, 02:27:04 pm
-
Hi,
Looks like Adobe it is focusing into try to make Lightroom quicker: https://blogs.adobe.com/lightroomjournal/2017/07/on-lightroom-performance.html
Tom Hogarty wrote:
I would like to address concerns recently voiced by our community of customers around Lightroom performance, as improving performance is our current top priority. We have a history, starting with our first public beta, of working with our customers to address workflow and feature needs, and we’d like to take that same approach regarding your performance concerns. We already understand many of the current pain points around GPU, import performance, certain editing tasks and review workflows and are investing heavily in improving those areas. Over the past year we’ve added numerous enhancements to address your performance concerns but we understand we will have a lot of work to do to meet your expectations.
They even have a survey about the topic: https://www.surveymonkey.com/r/LrDesktop_performance
Regards,
David
-
One thing they might focus on the GPU utilization and open CL as LR seems to not take much advantage of these compared to CC
Paul Caldwell
-
One thing they might focus on the GPU utilization and open CL as LR seems to not take much advantage of these compared to CC
Paul Caldwell
Lr doesn't even use multiple cores very well. I have 24 physical cores, and Lr can only use a handful of them. Matlab can use them all. So can AutoPano and PTGui.
Jim
-
Does Adobe's request from Tom Hogarty imply that the next version ( LR 7 ) is not imminent but quite a long way away ?
-
I saw this and filled out the survey. My LR performance has been terrible. On a new-ish, 27" (NON-5k) iMac, with 24GB RAM and smart previews already built, I still typically have to wait 4-5 seconds to scroll pictures in Library module as the dreaded "Loading..." appears.
One the one hand, I'm glad they are finally acknowledging the issue.
On the other, I can't believe it has taken so long. There have been speed complaints for years. I remember when they were flipping to a subscription model, lots of people said the sky was falling but I said no, give them a chance. Seems like the joke is on me. Near as I can tell, they have moved 80% of the engineering onto other things and the 20% that are left have been pointed mostly to LR Mobile, adding camera support, and the occasional random feature (dehaze: moderately useful; the abandoned "easy import" debacle: useless).
I do love the application and hope they turn it around but I'm finally at the point I'll look at alternatives.
-
My only concern about this announcement is why did it take so long for this issue to become a 'priority'?
Anyone who has frequented any of the Adobe hosted or Lightroom related user forums over the past several years would observe that these poor performance issues have been around for a significant number of users for quite some time. Not weeks. Not months, But for several years.
While I personally have not experienced slowness to the extent it is detrimental to my workflow ... yet ... I do see where there is much room for improvement in this aspect of Lightroom functionality.
I can recall a time when Adobe took great pride in being on the leading edge of such matters. It seems that since we entered into the CC subscription model, the emphasis has changed from what it once was when revenues were earned solely upon the merit system.
I hope this move is sincere (I currently have no reason to think otherwise) I also hope that the accountants and executives up the chain of power don't overrule any potential avenue Hogarty and his team choose to use for correcting the issue.
-
Another performance issue I'd like resolved is the slow speed of updates and improvements:
LR1: 1/2007
LR2: 7/2008 (18 months later)
LR3: 6/2010 (22 months)
LR4: 3/2012 (21 months)
LR5: 6/2013 (15 months)
LR6: 5/2015 (22 months)
Now: 27 months since LR6
Although there have been useful new features in LR5 and LR6/CC (and improved local editing), there have been no significant improvements in basic raw processing since "Process 2012" with LR4, five years ago. We used to complain that Nikon Capture NX2 got no improvements in raw processing for years and years...
The probably explanation is that Process 2012 is perfect, and no improvement in raw processing is possible.
-
It will be interesting to see if Adobe actually do anything above and beyond kludging together piecemeal attempts to "fix" the myriad issues - they could instill a new bolus of confidence in their user base if they innovated with a completely new product. Sort of like the thread Schewe started a while ago about "If Thomas designed a new Photoshop" or whatever it was called.
The market is practically saturated with all-in-one raw converter image processing applications, of various quality and price points, none of which require a subscription. Adobe might consider getting out of its own way and starting fresh, with a team granted full creative freedom to innovate and a quality assurance program to make it work the first time instead of charging people to beta test its software, and folks who respond to the user base in a timely manner when the product does not work as intended.
kirk
-
...
The probably explanation is that Process 2012 is perfect, and no improvement in raw processing is possible....
...
The probably explanation is that Process 2012 is perfect good enough, and no improvement in raw processing is possible needed since everybody is hooked on paying each month anyway...
-
Another performance issue I'd like resolved is the slow speed of updates and improvements:
LR1: 1/2007
LR2: 7/2008 (18 months later)
LR3: 6/2010 (22 months)
LR4: 3/2012 (21 months)
LR5: 6/2013 (15 months)
LR6: 5/2015 (22 months)
Now: 27 months since LR6
Although there have been useful new features in LR5 and LR6/CC (and improved local editing), there have been no significant improvements in basic raw processing since "Process 2012" with LR4, five years ago. We used to complain that Nikon Capture NX2 got no improvements in raw processing for years and years...
The probably explanation is that Process 2012 is perfect, and no improvement in raw processing is possible.
Or, when was the subscription model implemented?
It's also become pretty quite around Julieanne Kost's contributions to tutorials, save a few on Lightroom Mobile.
Cheers,
Bart
-
...
The probably explanation is that Process 2012 is perfect good enough, and no improvement in raw processing is possible needed since everybody is hooked on paying each month anyway...
Regardless of purchase model, 2012 is good enough.
They should add requested features and improve performance before worrying about appeasing the pixel peepers that stare at conversion issues.
-
At least Tom Hogarty seems to think addressing our desktop needs with regards to performance is a serious issue. Perhaps that means they'll pause a bit with assigning resources to mobile. I do use Lr Mobile and it's handy but I really need Lr desktop at its best.
I'll also only believe that Lr and ACR are top notch again when their demosaicing can rival Iridient Developer/X-Transformer with my Fuji RAF files.
-
I do use Lr Mobile and it's handy but I really need Lr desktop at its best.
Agreed. While I fully understand that resources are not unlimited, the desktop version should not suffer because they chose to pursue a mobile solution.
-
I'm totally locked into Lightroom for what I do, so I have to live with what I get ;) I can live with the performance level as it is now, but would like improvements for sure. I'm using a MacBook Pro 2016 with a LG 5K external display. The files are from Nikon D810 (36MP) and Canon 5DSR (50MP). One of the annoying thing is the time it takes to render a picture completely sharp ready for editing in the develop module. The library module is pretty fast since I always generate 1:1 previews, so there is never a delay in zooming in pictures (before I have edited one, of course).
I find a few things annoying and lacking:
- There are quite a number of global edit functions that cannot be done as local edits, e.g. HSL.
- Presets are only absolute values. I cannot make a preset that changes a slider value by a delta value relative what the value already is. This really limits the value of presets.
- When I do a HDR or Pano merge Lightroom does not keep track of the files that went into a merge.
- I would like a really good focus stacking merge that avoids blur haloes
- I can't see clipping in the library module. I have to look at the histogram in library and guess when there are clippings and then go to develop to see the actual clipping
- I can't see the real clipping in Lightroom even in the develop module. I have to go to Rawdigger to see the real clipping. I always bracket my shots and choose the most exposed without clipping and this is overly complicated in some cases when not seeing the real clipping.
There are probably more that I do not have in my mind right now, since I normally don't list these since I have long given up on getting these anyway....
-
- Presets are only absolute values. I cannot make a preset that changes a slider value by a delta value relative what the value already is. This really limits the value of presets.
I agree, the stamping of static values is getting really old.
-
I agree, the stamping of static values is getting really old.
It's not clear to the value of additive or substractive values. Static values give one a starting point, which is then adjusted to taste.
-
At least Tom Hogarty seems to think addressing our desktop needs with regards to performance is a serious issue.
speaks volumes about how they do QA in house... it takes users to complain and they can't detect issues themselves ? are they not testing their software on data/hw that their uses have ?
-
Regardless of purchase model, 2012 is good enough.
it is good enough also because of the purchase module - there is less pressure to innovate
-
it is good enough also because of the purchase module - there is less pressure to innovate
Agreed. Once you can persuade users to pay regularly just to get the same old same old, then no need to innovate. Cashflow sorted, and you can sit back and relax, and watch the money coming in. Of course, there are threats from competition, but corporates in that recurring-income situation are under enormous pressure to milk the revenue, and run the risk of not investing in developing the product until it's too late.
I'm quite sure Tom Hogarty and others at the coal face are aware of the risks, but I wonder if those that set the R&D budgets are aware?
-
Agreed. Once you can persuade users to pay regularly just to get the same old same old, then no need to innovate. Cashflow sorted, and you can sit back and relax, and watch the money coming in. Of course, there are threats from competition, but corporates in that recurring-income situation are under enormous pressure to milk the revenue, and run the risk of not investing in developing the product until it's too late.
I'm quite sure Tom Hogarty and others at the coal face are aware of the risks, but I wonder if those that set the R&D budgets are aware?
Exactly ... but as long as Adobe has enough users in this situation:
I'm totally locked into Lightroom for what I do, so I have to live with what I get ...
Then there really is no incentive to attack problems and push the envelope on development with as much enthusiasm as they once did when their revenue was gathered in a different manner.
Sure, $10 a month is nearly painless ... if enough end users truly consider CC a 'service' and not the purchase of a toolset, there is little we as users can do to pressure Adobe to be more diligent on such matters. From a workflow standpoint there are few options that are as efficient for a volume workflow.
I was concerned that complacency would eventually set in once Adobe went the CC route, I had hoped it wouldn't have arrived so quickly.
-
Does anybody here (who is permitted to speak freely about it) know what kind of software development resources are available to maintain and enhance Lightroom?
My impression from some of the things Jeff Schewe has posted in the past is that there is a small dedicated Camera Raw staff, but are there also any dedicated LR application developers?
In my experience, mature, niche software products—and I assume that's how Adobe views Lightroom in the context of its Creative Suite—often need to compete for staff time with some or many other products.
This survey "to address concerns recently voiced by our community of customers around Lightroom performance" strikes me as rather odd in a couple of respects. First, the wording of the announcement: "recently voiced?" Really? Second, the broad scope of the questionnaire, which might be summarized as "tell us everything that bothers you regarding Lightroom's performance, all the ways you use it, and the characteristics of your workload and computing environment." Somehow, this doesn't sound to me like a well-considered attempt to collect the kind of information from customers that would be useful to software developers, especially members of a pool who were called in from time-to-time for their expertise in performing a specific task.
Maybe I'm being too cynical, but I can't help wondering whether Adobe has finally figured out how to significantly enhance Lightroom's performance—perhaps primarily or even solely by improving the Camera Raw algorithms and code—and that this survey is essentially a marketing gimmick for the next major release: You told us you wanted Lightroom to be faster and we delivered!
-
No-one who knows could say anything of significance, Chris. You know that, don't you?
I don't think you should give too much credence to the idea that "Adobe has finally figured out how to significantly enhance Lightroom's performance".
Performance isn't a simple thing - it's multi-faceted. One thing I have advocated since day 1 was the ability to browse embedded previews, just like Photo Mechanic. That's one thing they could do, resolving problems in the culling stage of typical workflows. But it would only open up that one bottleneck. Others potential enhancements are specific to other aspects, for example rendering a raw image after lots of dust spot healing. Some are specific to certain cameras (eg I can export JPEGs from DNGs 35% faster than I can export them from the Fuji raw files used to create the DNGs so there must be an opportunity there). More might come from eliminating unnecessary re-rendering of images, so for instance syncing photos to Mobile might also generate smart previews (see below). It's likely to be a whole range of improvements, not one big thing, and targeting a number of bottlenecks.
I'd also question your "finally". If you read Hogarty's post, you'll see that he refers to things they have already been doing to address performance. I call these "what have the Romans ever done for us?". One was the ability to use smart previews in place of the originals. I just wonder how many take advantage of that. Another is GPU acceleration of certain tasks which, at least in my case, has not delivered any obvious value and just seems to expose LR to more driver-related problems (AMD's recent driver update crashed LR on Windows). The trick is probably to join up enough of these fixes to make general progress impossible to ignore. So imagine that syncing to LrMobile also generated those smart previews to speed up adjustment, for example.
The most surprising thing to me is that Hogarty chose to go public. Of course, they might learn something from the whinefest. Or maybe a bit like Trump's Twitter strategy, for a few days no-one will be talking about anything else....
John
-
I call these "what have the Romans ever done for us?". One was the ability to use smart previews in place of the originals.
Well I never knew the Romans invented smart previews!
I agree with you about the other enhancements. During the first 5 years of LR, we saw lots of improvement to basic raw processing. In the 5 years since then we've seen more local editing and other new features, some more successful than others (remember the attempt to change the import dialog?) For me, performance is OK on a fast machine, tediously slow on my old laptop.
The fact that there's been no improvement in basic raw processing in 5 years does makes me wonder if the others have now overtaken LR. Maybe it's time to look at Capture One again...
-
The fact that there's been no improvement in basic raw processing in 5 years does makes me wonder if the others have now overtaken LR.
I'm not so sure about that there's been no improvement in basic RAW processing. By that I think you may mean demosaicing algorithm. I can only say that I've seen improvements in demosaicing for my Fuji RAF files. They don't seem quite up to what I get from Iridient yet but they're closing the gap.
-
I just have one comment to all the inspired comments from lots of experienced users of "Photoshoppers" and ACR/Lightroom.
While the latest version of Lightroom was in Beta format awaiting release I was involved in testing the features of the application and thought that hare is a product that would eventually replace Adobe Camera Raw. However this did not materialize, and I believe this was because the core of existing Photoshop / ACR users were just not ready to accept change. Now Adobe has moved to Photoshop CC with the (photographers package) which includes Adobe Camera Raw, and Photoshop Lightroom at no additional cost to the subscriber.
Adobe is now saddled with managing two applications that provide, file management and processing of RAW image file from digital cameras! Why have they not found a way to incorporate both in a single application.
-
I just have one comment to all the inspired comments from lots of experienced users of "Photoshoppers" and ACR/Lightroom.
While the latest version of Lightroom was in Beta format awaiting release I was involved in testing the features of the application and thought that hare is a product that would eventually replace Adobe Camera Raw. However this did not materialize, and I believe this was because the core of existing Photoshop / ACR users were just not ready to accept change. Now Adobe has moved to Photoshop CC with the (photographers package) which includes Adobe Camera Raw, and Photoshop Lightroom at no additional cost to the subscriber.
He Adobe is now saddled with managing two applications that provide, file management and processing of RAW image file from digital cameras! Why have they not found a way to incorporate both in a single application.
Not exactly... Lightroom and Camera RAW shared the same RAW processing pipeline under the hood. The difference it is in the GUI.
Of course, then Lightroom has more modules that give it its uniqueness with respect Camera RAW.
-
The fact that there's been no improvement in basic raw processing in 5 years does makes me wonder if the others have now overtaken LR. Maybe it's time to look at Capture One again...
I don't see any significant difference between LR & CO. Any difference in results is mainly due to how the user drives the software. The defaults of either may suit an individual better, but once you know how to use both there'll be little difference. LR still leads with a more intuitive interface and workflow for most photographers.
For Bayer based sensors the differences between raw converters is pretty minimal, it's how they're bundled and presented that becomes important.
-
I'm totally locked into Lightroom for what I do, so I have to live with what I get ;) I can live with the performance level as it is now, but would like improvements for sure. I'm using a MacBook Pro 2016 with a LG 5K external display. The files are from Nikon D810 (36MP) and Canon 5DSR (50MP). One of the annoying thing is the time it takes to render a picture completely sharp ready for editing in the develop module. The library module is pretty fast since I always generate 1:1 previews, so there is never a delay in zooming in pictures (before I have edited one, of course).
I find a few things annoying and lacking:
- There are quite a number of global edit functions that cannot be done as local edits, e.g. HSL.
- Presets are only absolute values. I cannot make a preset that changes a slider value by a delta value relative what the value already is. This really limits the value of presets.
- When I do a HDR or Pano merge Lightroom does not keep track of the files that went into a merge.
- I would like a really good focus stacking merge that avoids blur haloes
- I can't see clipping in the library module. I have to look at the histogram in library and guess when there are clippings and then go to develop to see the actual clipping
- I can't see the real clipping in Lightroom even in the develop module. I have to go to Rawdigger to see the real clipping. I always bracket my shots and choose the most exposed without clipping and this is overly complicated in some cases when not seeing the real clipping.
There are probably more that I do not have in my mind right now, since I normally don't list these since I have long given up on getting these anyway....
Hello,
I fully agree to all Hans mentioned above, but I would like to add one thing :
The local adjustments should give the option to paint 'sharpness' using different settings like radius and details than the one chosen in the details panel.
Robert
-
- I can't see the real clipping in Lightroom even in the develop module. I have to go to Rawdigger to see the real clipping. I always bracket my shots and choose the most exposed without clipping and this is overly complicated in some cases when not seeing the real clipping.
why not use FRV (fastrawviewer) before importing to LR ... cull there (FRV like RD shows clipping in raw channels in place) and it can even generate XMP with certain adjustments that LR will pickup ...
-
why not use FRV (fastrawviewer) before importing to LR ... cull there (FRV like RD shows clipping in raw channels in place) and it can even generate XMP with certain adjustments that LR will pickup ...
I do that, and it works great. But it makes Lr seem even slower when I start using it.
-
But it makes Lr seem even slower when I start using it.
indeed, now poor Lr must also apply FRV corrections stored in .xmp sidecars ... what a sisyphean task ;D
-
indeed, now poor Lr must also apply FRV corrections stored in .xmp sidecars ... what a sisyphean task ;D
I don't use the .xmp sidecars. I just use FRV for culling. I'm talking about psychology.
Jim
-
I'm talking about psychology.
OK
-
I have 24 physical cores
Dual socket or single socket? Broadwell or an earlier generation? I'm just curious. Last year a local business was selling a used HP dual socket workstation with both sockets populated with 14 core Xeons (which of Haswell or Ivy Bridge, I forget). I'm sure it cost them a pretty penny when new. They were asking for about $2,500. I didn't get it, and ended up going with a new 4 core 6700K because (1) it was cheaper and (2) for what I do most, faster.
As an aside, I find Autopano can gobble up lots and lots of RAM / swap space when assembling large panos. And as another aside, I tend to find its stitching more reliable than PTGui. How about you? And finally, I'm relieved an update to Autopano was finally released after the sale to GoPro, although I'm still worried GoPro doesn't care much for it and it therefore won't receive much investment. Bummer.
-
And finally, I'm relieved an update to Autopano was finally released after the sale to GoPro, although I'm still worried GoPro doesn't care much for it and it therefore won't receive much investment. Bummer.
Something's going on with GoPro for sure. No one answered the phone at their San Mateo California office several months ago so that I could find out how to return a GoPro I found in working condition at the bottom of my local river. I posted this on their Facebook page months ago and still haven't gotten a reply.
When I called their customer service number as their website instructed for such matters, I couldn't understand the CSR over his thick Pakistani accent talking to me from the Philippines. He also didn't speak as if he knew what I was talking about. I gave up and am now a proud owner of a GoPro. Just need a battery charger.
-
https://www.bbb.org/greater-san-francisco/business-reviews/photographic-equipment-and-supplies-retail/gopro-in-san-mateo-ca-355874/reviews-and-complaints
-
https://www.bbb.org/greater-san-francisco/business-reviews/photographic-equipment-and-supplies-retail/gopro-in-san-mateo-ca-355874/reviews-and-complaints
Was that meant to confirm my suspicions, Ron?
Or are you telling me to file a complaint to BBB instead of posting here?
-
Sorry, Tim. Just meant to confirm/show that your complaint about support seems more or less universal with GoPro.
-
Not being able to drag curve points "off" the chart in order to delete them is very annoying.. why would they do away with that at this point??