This bug report is rejected as file size is not an absolute measure of lossless. The lossless H.264 preset was changed to not be intra-only (only I frames), and now has a GOP size of 25 frames. This allows creating much smaller files while still being lossless. You need to look at the resulting file with MediaInfo and ffprobe and look for the H.264 profile is “High 4:4:4 Predictive”. I just ran a test, and I see it. It also plays fine. Another thing you can do to verify is right-click the completed export job and choose Measure Video Quality, wait for that job to finish, right-click it, and view the report. You should observe 1.0 values for the ssim[Y] column
If you want intra only, change the GOP to 1. If you want a large file size (difficult with only 10 seconds as your video is “Length 00:10:02”) use Ut Video.
Does length of 00:10:02 mean ten minutes or ten seconds? I tried a one minute export of a PowerPoint slideshow (meaning GOP 25 will have a lot of compression gains) and the result was 168 MB. So if you had a ten second video that was 100-ish MB, it would be within reason. The real check is the codec profile as Dan mentioned.
Those export settings screenshots look fine. How does the exported video look if brought back into Shotcut? Is there a noticeable problem? File size alone doesn’t tell us anything since the GOP was changed in the export preset compared to previous versions.
The application log is mostly for Shotcut startup. We would need the export job log to detect any problems. Look on the Jobs panel and right-click the finished export to find the job log.
QP (at the end) is zero, meaning lossless. The profile is High 4:4:4 Predictive which is used by lossless. Then we see the statistics when encoding is done:
The average QP is zero, meaning lossless. Having 14,467 P-frames means this video is getting extremely good compression.
I don’t see anything wrong. The exported file looks legit. Unless there is a visible problem during playback that says otherwise, this file looks perfectly fine.