/t5/lightroom-classic-discussions/lightroom-3-2-release-candidate-available-on-adobe-labs/m-p/2857964#M3389Aug 12, 2010
Aug 12, 2010
Copy link to clipboard
Copied
Error when adding Date/Time to filename on import: This is a frequently reported bug in LR3 but is not mentioned as a fix in 3.2. When will this be addressed?? I sent a report several weeks ago and I know others have as well. When renaming on import, the time in the filename is shifted back by 4 hours. Problem occurs on both the desktop and laptop installations. I now import without renaming and need to remember to rename all files AFTER import to get an accurate time in the filename. My son, also on 2 computers, on a separate license in a distant city has the same bug. Others have complained as well. This is a big pain in the neck and seems like it should be a pretty simple fix. Please help us out!
Error when adding Date/Time to filename on import: This is a frequently reported bug in LR3 but is not mentioned as a fix in 3.2. When will this be addressed?? I sent a report several weeks ago and I know others have as well. When renaming on import, the time in the filename is shifted back by 4 hours. Problem occurs on both the desktop and laptop installations. I now import without renaming and need to remember to rename all files AFTER import to get an accurate time in the filename. My son, also on 2 computers, on a separate license in a distant city has the same bug. Others have complained as well. This is a big pain in the neck and seems like it should be a pretty simple fix. Please help us out!
I've come across this bug too. For me it was only one hours difference. Someone told me it has to do with which time zone you're in.
Now I've rethought my file naming scheme. I just do a letter (designating the camera used) and a serial number.
I noted no "export" related fixes listed in Tom's note references above. I am still experiencing a signficant difference in export times between LR 2.7 and LR 3.x. Timings are consistently double for the exact same images, exported exactly the same way, to exactly the same location. I have attempted to match all settings as much as possible (given 2003 vs. 2010 process model). I have not tried an export of the same images under 2003 un LR 3.x yet. Is anyone else seeing major differences in export and have you reported it? I just did, but just checking with the group.
Jay S.
Just a postscript to the above. Reverting to 2003 Process model and unchecking Lens correction brought export times "closer" to 2.7. The bigger offender (at least here) was the Lens Correction module. Process 2010 added only about 2 seconds to export. The bulk came from the lens correction added to the images to be exported. Would be curious if others see similar results.
Jay S.
Although I haven't specifically tested Lens Correction's effect on Export times, I have certainly become aware of the time it takes for Lightroom to compute the profile-based lens corrections (multiple seconds). It makes sense this would always be added to export times since exports always illicit a re-rendering from scratch.
I would like to see the option to declare a photo 'Done' in Lightroom and have it be locked for editing, and at that time, re-render or save the final RGB image so that exports and viewing are forever speedy for "Done" photos.
I noted no "export" related fixes listed in Tom's note references above. ...
Yes, JayS, I'm a sad puppy, too...though for a different reason than you. LR 3.2RC hasn't fixed uploading Web collections via sftp. I'm back to exporting and uploading manually. At exactly 50%, LR crashes and burns:
I noted no "export" related fixes listed in Tom's note references above. ...
Yes, JayS, I'm a sad puppy, too...though for a different reason than you. LR 3.2RC hasn't fixed uploading Web collections via sftp. I'm back to exporting and uploading manually. At exactly 50%, LR crashes and burns:
I am still experiencing a signficant difference in export times between LR 2.7 and LR 3.x. Timings are consistently double for the exact same images, exported exactly the same way, to exactly the same location. I have attempted to match all settings as much as possible (given 2003 vs. 2010 process model).
The new sharpening and NR are the likely cause. Better = more CPU time.
I am still experiencing a signficant difference in export times between LR 2.7 and LR 3.x. Timings are consistently double for the exact same images, exported exactly the same way, to exactly the same location. I have attempted to match all settings as much as possible (given 2003 vs. 2010 process model).
The new sharpening and NR are the likely cause. Better = more CPU time.
Lee Jay,
Not double though (at least according to Adobe)... As I said in the post, it isn't as much of the extra items in 2010.. The bigger impact came from the Lens Correction being on or off.
/t5/lightroom-classic-discussions/lightroom-3-2-release-candidate-available-on-adobe-labs/m-p/2858002#M3427Aug 13, 2010
Aug 13, 2010
Copy link to clipboard
Copied
In general I'm so pleased with 3.2 RC (64 Bit on Windows 7 Ultimate) that I use it for my production work. One issue very annoying for me hasn't been resolved yet: Like many I'm using keywords to organize my photos. Normally I work on a collection of photos with a given keyword like "event xyz" (so I filter my photos by this keyword). When I edit an image in Photoshop CS5 (64 Bit), save it and return to Lightroom the edited photo isn't shown (in Lightroom 2.x it was).
Two "workarounds": Exit Lightroom, Reopen it, Photo is there! Alternatively, deactivate Filter, locate edited image, remove keyword from image, add keyword to image ...
/t5/lightroom-classic-discussions/lightroom-3-2-release-candidate-available-on-adobe-labs/m-p/2858035#M3460Aug 16, 2010
Aug 16, 2010
Copy link to clipboard
Copied
I am puzzled about what the ACR cache really does, quite frankly. I get pretty much the same delay loading an image into the Develop module whether it is in the cache or not. The only thing which actually seems to affect the load time is how many edits have been made to the image, specifically perspective correction and local brush edits. They really slow things down. But I can purge the ACR Cache and afterwards I see no real difference in performance, either faster or slower. Mind you, I have probably have the worst-case scenario (or close to it) with 39MP files from a Hasselblad back, which expand to a 230MB 16-bit rendered image.
I had a problem with identical symptoms when I first started using 3.2RC, but now I don't have the problem anymore (I said more about that in a previous post in this thread).
It sounds like Lr is rendering your photos from scratch each time and not using the ACR cache (properly).
Try this:
- Open the cache folder in Explorer or Finder and delete everything in it.
- In develop mode, switch back and forth between the same two photos a few times.
How many entries are in your cache now?
If more than two then you have the same problem I had. If exactly two then I don't have a clue (how long is your render time? how big are the cache files?). If less than two then that explains your problem but its different than what mine was.
I don't *think* there was anything magic about the location - I theorize that somehow moving it forced Lightroom to re-initialize something. Anyway, I'm anxious to hear about your results...
It sounds like Lr is rendering your photos from scratch each time and not using the ACR cache (properly).
Rob, as per Ian Lyons' info, LR always renders photos (almost) from scratch.
The "almost" qualifier is there because if the image has been cached in the ACR cache, you save the (partial) demosaicing stage whose result is stored in the cache. Everything after that is re-rendered on the fly each time, hence one must not expect quick switching times if images have had considerable edits applied to them.
My switching (loading) time was around 3 seconds for cached, and around 10 seconds for non-cached. So, caching is a great savings for me. But, it doesn't sound like John is having any noticeable speedup whatsoever (cached vs non-cached), and if thats true then something's broken...
My switching (loading) time was around 3 seconds for cached, and around 10 seconds for non-cached. So, caching is a great savings for me.
Do you still see 10 seconds loading time for non-cached images? It never takes that long for my uncached images. No doubt my images are smaller than yours but 10 seconds seems long (for an unedit image with no expensive default operations applied).
My machine is about 4 years old now, 2GHz dual core / 333MHz bus (windows rating = 4.2).
12MP NEFs from D300. Download one of my raws (http://www.robcole.com/_temp/NIK3_29418.zip) if you want to compare apples to apples. Send me one of yours and we can compare oranges to oranges too.
I don't know what exactly I should expect because people often chime in with numbers that don't seem reasonable to me so I'm never sure what to believe.
I, like some others, wish Adobe would publish some performance guidelines, but I'm not holding my breath.
This is with all default develop settings explicitly turned off (using the panel activation switches).
With my default settings (which include sharpening) the numbers are:
Non-cached: ~5.3sec
Cached: 4.3 sec
So you see the differences aren't huge. Since my files are smaller, the difference is even lower, i.e.,
Non-cached: ~3.2sec
Cached: ~2.8 sec
Measurements varied (not only due to timing inaccuries), so take these numbers with a grain of salt. However, I did multiple measurements to make sure I don't report outliers.
As you can see, the ACR cache doesn't do wonders on my system with my image size.
Note that the display size doesn't come into these measurements as the rendering produces 1:1 images.
Here are my system specs:
Operating system: Microsoft Windows XP Professional Service Pack 3 (Build 2600) Version: 5.1 [2600] Application architecture: x86 System architecture: x86 Physical processor count: 2 Processor speed: 2.3 GHz Built-in memory: 1999.8 MB Real memory available to Lightroom: 716.8 MB Real memory used by Lightroom: 274.9 MB (38.3%) Virtual memory used by Lightroom: 267.9 MB Memory cache size: 1.6 MB System DPI setting: 110 DPI Displays: 1440x900
You are able to render a non-cached version twice as fast as me, yet our machines are comparable. And your cached time, as you've been saying, isn't much of an improvement. The relative values I come up with seem more reasonable - i.e. triple fast when going through the cache makes more sense than the 20%-ish savings you are getting (can anybody confirm - should we expect 20% or 300%?). I'm beginning to think your cached rendering is too slow and my non-cached rendering is too slow. - Not sure where to go with this...
As a reference point: It takes NX2 about 3 seconds for a default rendering of my raws.
Also, I tested ACR 6.2 independently of Lightroom by loading raws direct into Photoshop - timing is same as Lightroom for non-cached version (10 seconds), but a full 8 seconds for the cached version to load - more like your 20% savings.
I can not recall if you're running mac or win (or if this has already been mentioned), but if you are running windows have you tried to disable virus scanning for raw-files, LR catalog and LR-cache files? Depending on the anti-virus software of your choice you maybe can do the exclusion based on the file-extensions instead of folders. In my other system (dual-core) excluding those files from scanning makes a difference on loading times.
However, your experience with NX2 refers that you have either already done that for raw files but not for the cache files, or then the slowness is due to some ill interaction between LR and the system.
I am puzzled about what the ACR cache really does, quite frankly. I get pretty much the same delay loading an image into the Develop module whether it is in the cache or not.
John, I see the same. The ACR cache only provides a very minor speed up that pales in comparison to the delay that edited images cause.
I thought that maybe I'm not appreciating the ACR cache because my RAW files are relatively small (6 MP) so the time saved for demosaicing them isn't gigantic.
With your 39MP files, you should definitely be able to notice a speed up. Demosaicing such large files should take considerably longer and (provided disk access isn't very slow) saving that time should be noticable. The best way to test that is to switch between images without edits (and making sure you don't have computationally expensive default develop settings). Once the images are in the cache (purge it before you start testing) the switching between the images should be a quicker. Once edits such as adjustment brushes come into the mix, they are likely to dominate the rendering time and the speedup provided by the ACR cache might go unnoticed unless you do some precise timing.
I think you agree with me that it would be nice if one could opt in to an extra version in the cache that contains all edits instead of just the (even only partially?) demosaiced version of the original.
200
Replies
AdChoices