Hi How many images can an aspect library cope with. I am trying to work with a library size of between 220,000 and 250,000 images is this possible?
The reason I ask is it is taking a very long time to create a new library with this number of images and the time remaining keeps getting higher.
Those numbers shouldn’t be a problem in general. We are regularly testing with a library that contains 1 million images, which works fine. At that size, there are still a few places left to optimize where things don’t feel quite fluent anymore, but otherwise everything works as it should.
When things take that long at this stage, it is nowadays usually because of video files, where just reading the metadata unfortunately can take up to a few seconds, depending on the operating system. How much time remaining does or did it display for you?
Hi, When I wrote the last message it said 2 hours remaining, currently it is saying three hours remaining. Would it have much to do with the fact that the raw files it is looking at are Leica DNG files. Currently they seem to be from the Leica M10.
Usually the metadata part for RAW files is relatively quick, but it is definitely possible that it is somewhat slower than JPEG files or other RAW formats (I’d have to benchmark that, though), which could explain the growing time estimate.
Overall though, the performance so seems to be okay. Lets say 5 hours for 250k images, that would make roughly 14 images per second. Anything below 5 images per second or so might indicate that something is wrong, although that of course also depends a lot on the drive’s overall performance.
Would definitely be interesting to know how long it took in the end, though. I can run also a test here as a comparison.
I am running the latest build, I started about 11:00 am this morning and it is now 10pm for me and it is still saying I have 3 hours remaining so it is taking some time. I am curious as to how big the library file will be when it is finished as this was the reason that I paused trying out the software after my initial go with it.
Okay, that’s indeed quite long - to get it into perspective, is the drive in question an SSD or HDD and how is it connected to the computer?
When I get some time, I will make a benchmark comparing the metadata load speed of M10 DNGs vs. JPEG and one or two other RAWs to see whether that could indeed be responsible for a considerable slowdown.
And the App crashed before it finished. I have both the Library all the images on the same attached USB-C Sandisk 18TB hard drive.
Now I am using setting menu to scan the folders and see if that makes a difference.
I work for DxO as their UK product Liaison and I am hoping to be able to use Aspect for my DAM and then process images through DxO PhotoLab and then print using Nik Output sharpener from Photoshop or ideally with Affinity Photo. I am trying ideally work out a photo workflow that does not use Adobe software.
This is unfortunate, I’ve seen the crash reports, but there is no information about the crash in the log file and macOS system crash logs do not contain a sufficient stack trace to get to the origin of the crash. I would suspect that this is caused by a particular image file. If you have the chance, could you try to look at the last image (or the last few images) that is being mentioned in the full log file (look for “Load image by path”)?
The log file can be accessed by opening Terminal and entering open $TMPDIR, which should bring up a Finder window with the temporary files folder open. The log file is named “aspect-log.sdl” and is a simple text file and can be opened in a normal text editor. To verify whether that last image is responsible, it should be sufficient to drag it onto the Aspect application bundle and see if that makes it crash.
Regarding performance, what I can say from my recent experience with shingled magnetic recording - which the 18 GB drive should be using , too - is that the performance is considerably worse than for traditional perpendicular recording. That would probably explain the performance, although I’ll have to do an actual benchmark to get a good baseline for comparison.