Trying out the program as my old Photoshop 7 doesn't have the features I need these days. Anyway, I have noticed that once into shapening with a JPEG or DEF, as soon as I slide the sharpening mask slider even to a value of "1" the program crashes. Other sharpening effects seem to work fine. PC is an AMD Athlon XP 2800+ (yes, I read the threads regarding a glitch with this processor and Photoshop CS5, which was finally fixed with a patch - don't know if something similar exists in LR), 2.5GB RAM, XP sp3, all the usual robustness to allow the program to work. Is there a setting or control I am neglecting (the beauty of trial version is to test everything before buying!). I admit I do not know all the features of LR yet, but don't think it should crash if I move that slider??? Thoguhts appreciated.
Here's a bit of a follow up should anyone care to comment. And BTW...the video card is an NVIDA GeForce 6200; I just made sure it has the latest driver. Still will consistently crash the program (not the PC....it's just the LR app that suddenly closes) whenever I try to apply the sharpening mask. Now though I also saw an "unexpected close" error, and it referenced the CameraRaw.dll file. I'm wondering if like from what I read in CS5 bugs that the AMD processor is a bit touchy in this part of the LR application????
This may not have caused the initial crash but you can check the following. When LR opens a catalog it creates a file called "Lightroom 3 Catalog.lrcat.lock" along side the Catalog File.
This is to ensure that there is no other use of the catalog while it is open. When LR is closed in the normal way the .lock file is automatically deleted. If however LR closes unexpectedly the .lock file will remain in place preventing LR from starting again. See the screen capture of the relevant folder when LR is open.
With LR closed go to your My Pictures>Lightroom folder and check if there is a .lock file present. If there is one delete the file and then try starting LR.
Thanks for the thoughts. I checked the folder and did not find a xxx.lrcat.lock
file. And once the program unexpectantly quits, I can re-launch it with no problem, it's not as if a reboot is necessary to reload Lightroom.
I have also downloaded the trial of PS CS5 and just finished installing 12.0.1 (good thing I had read up on the Athlon processor fix in version 12.0.1!). Anyway, trying sharpening and masking there, it works fine. I do like LR's method of using a quick slider, however (unless I'm missing a trick in PS to do it the same). At any rate I have trial programs to test and compare, and will add PS Elements 9 to the beta test for the next 30 days...we'll see who wins! Though I'd love to have a definitive fix/answer to the sharpening mask in LR, not that I use that function to excess...
Lightroom might use some SSE2 or later instructions that your Athlon does not support. According to Wikipedia, Athlon XP 2800+ support only SSE.
Anyone know what SSE, if any, is Lightroom using?
I don't think that has anything to do with it, because sometimes adjusting the Masking slider does not make Lightroom crash. It can just happen, randomly I'm afraid.
I thought it happened mostly with high ISO shots but after some testing that is not exactly the case. Though it seems to happen more often with >400 ISO shots, it can also occur with <400 ISO ones.
About an hour ago I developed an ISO 200 image and everything was ok. Reading your reply made me start LR again and develop some pictures, everytime I invoked the Masking it made LR crash.
Then I reopend the shot that I developed an hour ago and readjusted the Masking slider (from 25 to 35). LR did not crash.
Developed some other ones, again LR didn't crash.
So it seems that every time LR is started it thinks by itself: What shall I do this time, behave myself or not?
Here's the info produced by a slider crash and a Microsoft reporting error:
AppName: lightroom.exe App/Ver: 3.2.0.10 ModName: cameraraw.dll
ModVer: 3.2.0.10 Offset: 002b729f
The details read as follows:
Exception Information
Code: 0xc0000005 Flags: 0x00000000
Record: 0x0000000000000000 Address: 0x00000000607b729f
System Information
Windows NT 5.1 Build: 2600
CPU Vendor Code: 68747541 - 69746E65 - 444D4163
CPU AMD Feature Code: C1C3FBFF
Module 1
lightroom.exe
Image Base: 0x00400000 Image Size: 0x00000000
Checksum: 0x001f155f Time stamp: 0x4c6901e3
Well, then it gets into a lot more data info that I don't have the time now to key in, I'm hoping this gives a good indication. The other affected modules listed are ntdll.dll, kernel32.dll, ui.dll, substrate.dll, AUTHZ.dll, ADVAPI32.dll, RPCRT4.dll
I know that PS CS5 had an issue with Athlon, and in fact I tested it when I downloaded the trial version. I used the program without the patch, and sure enough I got the application error that prevented the program from launching. When I applied 12.0.1 patch, everything comes up OK in my demo version.
I'm wondering if LR has a similar glitch with Athlon/AMD processors, and if so we can hear from Adobe about a patch to fix this. It's imporant while I demo to know which version of which program will not only meet my needs, but be stable as well.
Thanks for all the help and support....
PS....@harrieb: the interesting thing is that I get this slider crash whether it's a RAW or a JPEG....
Since clvrmnky (sorry I misspelled that name, but hey) made me aware of the Event Viewer and saw that there was a problem with cameraraw.dll I decided to start Bridge again (haven't used it since going from LR 2.x to LR 3).
Well, it produces the same error, and - again - seems to do so randomly. So if there is something wrong with the combination of AMD/Athlon and Cameraraw, it will be hard to pinpoint the problem.
0xc0000005 is an "access violation." Something tried to access a place in memory that was illegal or impossible. We don't know which operation failed because we don't have a symbolic crashdump (we just have an offset), but we know it was in CameraRaw.dll. This does not automatically mean that this DLL is suspect, only that this is where it was when it did something bad. It could have been given a bad address by something else.
CameraRaw.dll is a workhorse utility DLL that offers a lot of image services. In fact, I suspect it crashes in a different place for raw vs. JPEG images. If not, it is probably in the GDIplus stuff. There are no CPU or GPU specific functions in this DLL, and I'm pretty sure Lr has no dependencies on specific CPU or GPU calls -- it just makes generic calls to the OS for such resources. Ps had to do this in the past in order to squeeze more performance out of then-current workstations (and it is a 100% compiled native app) but generally this sort of trickery is no longer required. Doing stuff like this is always a trade-off between performance and testing/support nightmares, and you are at the mercy of the various CPU and GPU revisions out there.
Lr offloads a lot of its stuff to a Lua VM, which acts as a sort of glue between various image services and native UI services. The win here is that it is relatively easy to port Lr to various platforms, and you don't have to maintain such large multiple streams of code lines for each platform.
Report what you can to Adobe, and copy and paste as much of the crash log(s) you have. A .dmp file would be nice, but this takes a little tweaking to get right. If Adobe needs it, they can instruct you have to provide this. Make sure you know if you are running in 64-bit mode or not.
I sent as much info as the bug report page would permit. I will be interested in a response. Looking back on the PS CS5 Athlon problem it seemed that Adobe needed about a month to get the equipment and reproduce the error, and then come up with the patch. But they were responsive to the bug report. I hope so here as well. LR looks like a good choice for my general needs
.
And BTW...while others report the slider crash is random or occasional, it is 100% repeatable and consistent each time on my PC at this time.
Same thing for me too ![]()
Lightroom 3.2 crashes every time I am trying to use masking in the sharpening tool.
And no, this is not SSE related.
I have used the debugger and saw that it crashes because of use of uninitialized pointer.
I will test this issue on other machine tomorrow and will post the results.
Hope this will be fixed in the near future...
// Added
I have debugged LR a little bit more.
It crashes because it does not allocate memory for the image data, but tries to use it.
As a temporary workaround - you can move 'Clarity' slider on a needed photo, then reset it if you do not want to increase clarity.
This will allocate memory properly.
Now you can move Masking slider in the 'sharpening' tool without a crash.
And as another side effect of this bug - selecting any slider in the sharpening tool with pressed alt should show you the effect on the grayscaled image, but it works not every time.
And no, this is not SSE related.
I have used the debugger and saw that it crashes because of use of uninitialized pointer.
The fact that it makes an AV, does not necessarily mean it is not related to SSE. You don't know the code paths in ACR. It could go to another branch for non-SSE2 CPUs and access a bad pointer in a routine that is not used by newer processors. So, far everyone that reported this problem is running an Athlon XP (non-SSE2 CPU).
What is your CPU model?
Well, it looks like it is related to SSE, but not directly.
I mean that program crashes not because of use of unsupported SSE instruction, but because of use of uninitialized pointer (zero).
Checked on Pentium 4 machine - code path that crashes LR on Athlon XP does not executes at all.
My CPU is Athlon XP too.
On Athlon XP LR crashes in CameraRaw.dll at relative virtual address 0x2B729F (0x607B729F if it is loaded at default base address).
Interesting....I tried the workaround, moving the "clarity" slider on presence, then moving the sharpening mask. The program does not terminate at that point. I can move the clarity positive or negative, and even put it back to zero and continue to move the sharpening mask with no ill effects...
And now I must tell you what's wrong with this method...
Moving clarity slider will allow you to change masking ONLY in the current session.
I mean if you move clarity, then move masking slider and then restart LR - it will crash again when trying to display photo with sharpening masking settings.
Well, some hours spend in the debugger and I've made fix for this crashes.
You can download it from here: http://www.multiupload.com/99210FAR5H
Note: USE IT AT YOUR OWN RISK.
While it works just great for me, I can not promise anything.
And note that this is NOT official fix for the problem, it needs to be fixed by developers so it's better to send bug reports.
This solution can be used while there are no better solutions from Adobe.
This fix is only for 32-bit Lightroom 3.2
To install - just unpack this library to the folder where you have your LR installed.
For example C:\Program Files\Adobe\Adobe Photoshop Lightroom 3.2
It is better to save original library (to be able rollback all changes) and then overwrite it with this fix.
In fact this is original CameraRaw.dll, but just little patched.
Hope this will work for you too ![]()
Enjoy.
Yes, indeed. I sent a bug report to Adobe, although haven't recevied a reply, nor seen an official comment on this thread, so everyone else that has had this same issue please send a bug report so Adobe can determine it's not a one-off incident.
As I'm working LR as a trial, I'm finding it to be a very valuable program for my needs. I've seen some books in the store that go into depth about mastering it, something that would also be a valuable addition to the library.
All in all, LR and PE9 are proving to be a great combination for my ultimate needs,
Did someone tested this issue with new Adobe Photoshop Lightroom 3.3 Release Candidate?
It can be downloaded here: http://labs.adobe.com/downloads/lightroom3-3.html
Have this problem been fixed, or we need to report it once more?
I'll try to test it in a few days...
It's not neccessary to have those CPU's to test.
On Windows OS processor without SSE2 support can be faked by changing values returned by IsProcessorFeaturePresent().
And unfortunately the bug is still there ![]()
Now Lightroom crashes in CameraRaw.dll at 0x607B977F.
It looks like they are not going to fix it at all...
Well, stay tuned.
If the problem will persist in LR 3.3 final release, I'll fix it by myself as in LR 3.2
Keep sending bug reports...
I have made small investigation and found out that this bug is present in all Lightroom 3.x versions.
Well, for those who don't like to wait ![]()
CameraRaw.dll from Lightroom 3.3 RC1 fixed for non-SSE2 CPUs.
http://www.multiupload.com/PMHYLNYVGI
Usage instructions are same as with previous versions, just overwrite original library in the Lightroom application folder.
cppasm wrote:
It looks like they are not going to fix it at all...
According to the system requirements, Windows users need at least a Pentium 4.
http://www.adobe.com/products/photoshoplightroom/systemreqs/
So, I doubt this will ever be addressed.
clvrmnky wrote:
According to the system requirements, Windows users need at least a Pentium 4.
So what? You can not use it on any 32-bit AMD CPU?
Lightroom works fine on those CPUs, and this is NOT CPU-related problem, this is memory management problem.
Lightroom just have a number of implementations of the same things optimized for different processors, and memory management bug comes out only in implementation for CPU without SSE2.
So Lightroom supports AMD CPUs in its own, but implementation just have bug with memory allocation.
cppasm wrote:
clvrmnky wrote:
According to the system requirements, Windows users need at least a Pentium 4.
So what? You can not use it on any 32-bit AMD CPU?
Lightroom works fine on those CPUs, and this is NOT CPU-related problem, this is memory management problem.
Lightroom just have a number of implementations of the same things optimized for different processors, and memory management bug comes out only in implementation for CPU without SSE2.
So Lightroom supports AMD CPUs in its own, but implementation just have bug with memory allocation.
It is related to CPU when we are talking about which instructions sets are supported, or not.
You need a P4 "or equivalent" and that means you will have SSE2. Some AMD CPUs are not equivalent to P4. Therefore: not supported.
This is what "system requirements" mean. It may work in other cases until it fails, but that is not Adobe's problem because they have made their architecture requirements clear.
clvrmnky wrote:
It is related to CPU when we are talking about which instructions sets are supported, or not.
Lightroom crashes not because of use of unsupported instructions, but because it tries to use unallocated memory.
This is memory management issue, not CPU issue.
P4 users may have memory leaks due to this bug, Lightroom is not crashing - but this does not mean that everything is ok.
Not surprisingly, Adobe Camera Raw 6.3 have the same problem.
To completely fix this problem, you should also install patched Camera Raw plug-in.
Download link, 32-bit only: http://www.multiupload.com/SDPSHS6IK8
Just overwrite original plug-in, by default it is located in the: C:\Program Files\Common Files\Adobe\Plug-Ins\CS5\File Formats
This will fix the problem for all CS5 applications: Photoshop, Bridge, etc because they all use this plug-in.
Sorry, but I will not patch previous versions, so you will need to update Camera Raw to version 6.3 before using patched plug-in.
This problem is present in all Camera Raw 6.x and appears only on non-SSE2 CPUs.
North America
Europe, Middle East and Africa
Asia Pacific