Hi, I use a 256 GB SDCard, have ~ 1,700 files on it and still over 200 GB free. Finder shows its content in less than 2 s. Using Aspect’s import feature, it takes > 5 min to show the pictures and being able to copy them to the Aspect library. Additionally, no duplicates are found are somehow marked. Therefore, I have to remember the last imported one and select new files manually. What is Aspect doing the whole time “generating metadata”? It does not seem to do anything usefull 
Is there any chance to speed this up? If not, the import feature for external SDCard is practically unusable 
.
While we do have an open ticket for potentially falling back to sorting by modification time until the full metadata of all files is available, right now the default sorting mode is by capture time and we need to read all metadata up-front. On the other hand, after this process is done, the metadata should be cached and opening the folder next time (or importing from it) should work without any delay.
Regarding the duplicates, we do not currently have an individual indicator for duplicates (which would definitely be a nice addition, though). Does the “Include duplicates” checkbox in the import sheet mention any duplicates for you? It should say something like “Found 100 duplicates”.
Regarding imports in general, the most time-saving way to do those is to enable the “Activate Auto Import” switch and let the whole import run automatically in the background (without having to wait for metadata to be read first).
Having said that, the process overall seems suspiciously slow here. I just made a test with 200 RAW+JPEG pairs on an SD card. Reading those took less than two seconds, which would amount to about 15 seconds for 1700 images. Is there possibly a larger number of video files on the card, as reading the metadata in those can take a lot longer?
1 Like
Thanks for letting me know how it should work. Unfortunately, on my machine it reads the meta data every time I “open” the folder on the SDCard. Therefore, I have to wait for several minutes, every single time I try to import a few pictures from my camera. And it never indicates duplicates in the “import sheet checkbox”, never ever. That’s why I am afraid to run it in the background because I do not want to handle thousands of duplicates in my Aspect library.
Regarding the slow meta data processing: I don’t shoot video with my camera. BUT I had a problem with another app that used Apple’s HEIF framework to generate thumbnails. There seems to be a macOS framework bug related to Fujifilm 40 MP pictures:
We determined that a Mac Studio takes 3.5 seconds to decode the HEIF file, and instrumentation shows that all the time is spent in Apple’s HEIF decoder.
There might be still some HEIF images on the SDCard and maybe that’s somehow related? Let me know if you want me to double check or send an example file to you.
It looks like you hit the nail with HEIF. We got some HEIFs from a Fuji X100 VI and while they are not actually getting decoded when reading just metadata, Apple’s HEIF implementation always reads the whole file, regardless of what parts are actually needed. For ~20 MB sized files that is enough to seriously slow down the import.
At the same time, the Windows HEIF decoder is not able to read any metadata from those files — so we got a problem there, too. The only solution seems to be to integrate another HEIF reader for metadata handling and rely on the OS supplied implementation only for the actual decoding (which is the only viable option due to the licensing/patent situation).
What is still strange is that the metadata doesn’t appear to get cached. That part worked flawlessly in our tests yesterday and even when I tried to deliberately damage the metadata cache, it always recovered as intended.
If you have the chance, can you send me a log file where you just open the folder that contains the on the SD card through the “File Browser” and then just quit the application after all metadata has been read? Hopefully there should be some kind of error that gives a hint where this goes wrong.
1 Like