Release Candidate version 23.04.20

You can try out the release candidate for next big release including: Qt 6, GPU Effects, Motion Tracker, Glaxnimate 0.5.3, Speed filters, filter sets, Mask: Simple Shape VUI! Shotcut - Beta Version 23.04

Over the previous beta version, this is mostly just bug fixes except motion tracking was integrated into more filters such as Crop: Rectangle, Spot Remover, and all text filters.

The next phase will only be addressing only major issues and not include general fixes, changes, or enhancements. This is expected to last through the end of the month.

6 Likes

I have problems with this release. If you add ā€œcountā€ to the timeline, the render will always fail, even if you try to export without a hardware encoder.

mlt_repository_init: failed to dlopen /home/docent27/Š”Š¾ŠŗуŠ¼ŠµŠ½Ń‚Ń‹/Š”ŠžŠ¤Š¢/Shotcut/Shotcut.app/lib/mlt-7/libmltsox.so

(libsox.so.3: cannot open shared object file: No such file or directory)

[libx264 @ 0x7ffb180021c0] using SAR=1/1

[libx264 @ 0x7ffb180021c0] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX FMA3 BMI2 AVX2

[libx264 @ 0x7ffb180021c0] profile High, level 4.2, 4:2:0, 8-bit

[libx264 @ 0x7ffb180021c0] 264 - core 164 r3095 baee400 - H.264/MPEG-4 AVC codec - Copyleft 2003-2022 - x264, the best H.264/AVC encoder - VideoLAN - options: cabac=1 ref=2 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=6 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=12 lookahead_threads=2 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1 b_bias=0 direct=1 weightb=1 open_gop=0 weightp=1 keyint=300 keyint_min=30 scenecut=40 intra_refresh=0 rc_lookahead=30 rc=crf mbtree=1 crf=23.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00

Failed with exit code 11

Failed with exit code 11


The appimage version has the same error.

Same on Windows 10

For some reason in glaxnimate I canā€™t download emote packs. I understand that most likely the developer of shotcut has nothing to do with the development of glaxnimate, but I felt it necessary to notify about this bug, which was not in the previous stable version of shotcut + glaxnimate.

That is a OpenSSL compatibility issue on Linux that will likely not be addressed. Basically, we do not bundle OpenSSL libs on Linux because it is better to get those from the host OS for security updates (following AppImage guidelines). But OpenSSL broke backwards compatibility with version 3. Qt is built against older OpenSSL that appears in Ubuntu 20.04 LTS /Debian 11. So, it can work on an older OS with OpenSSL 1.1 but not a newer one with OpenSSL3. The Flatpak runtimes build a custom Qt against a compatible bundled OpenSSL since these are maintained aggressively similar to a distribution.

Add a motion Tracker (CarTracker) - previews fine, analysis is also fine
Add a 2nd Motion Tracker (CornerPinTracker) - previews fine, analysis of this tracker is copied to the analysis output of 1st Motion Tracker (Car Tracker)
Can be seen in .mlt

If it cannot be replicated, try it with filter trimming (as mentioned in the recent motion tracking thread)

fixed

2 Likes

I reproduced something a little different but basically the same. As a workaround for now, you can limit to one tracker filter per clip, and simply add a copy of a clip to the playlist for each different tracker. This is also an alternative to using filter trimming: instead, trim the playlist clip.

Update: this is fixed now, but one also should take into consideration the performance overhead of multiple trackers when doing analysis. Motion tracking can be intentionally affected by other filters that come before it such as typically Size & Position, or Crop: Source. So, running analysis does not automatically turn off all filters including other trackers. That means it may be still be running the first tracker when doing analysis on the second tracker and so on. A simple way to deal with this is to disable all Motion Trackers on a clip when running analysis. (Running analysis on a filter does re-enable that filter instance for the job.) Then, re-enable a tracker to preview its result. If that is bothersome to remember, you can overlook the performance overhead, or continue to use the workaround above.

1 Like

Just as a side note. I did disable the first tracker prior to creating the second but I will get the latest build and try again. The secondary tracker was specifically for testing purposes as the other thread piqued my interest.

I noticed that in the new Shotcut version both saved presets and regular text in Text:simple filter (maybe others too) changed from bold to regular. Has something changed with the weight property? This is true also when going back from a project saved in the new version to the older one (but this doesnā€™t matter).

This is the diff between how the old one saved bold text (left) and same project edited to show bold in the new version (right):

Another note: I see the maximum zoom of the timeline is now way limited, I assume this is related to the guy that couldnā€™t open his project due to RAM usage but I have a few small/short timelapse projects in which I individually add pictures with 1 frame duration and this limit is making it very hard to manipulate stuff now even at max zoom.

I donā€™t know which of our 2 cases is more common but maybe there can be a dynamic limit to zoom depending on timeline length or number of added clips?

This is at max zoom, I need 2-4x more to properly move/resize stuff/see at least the date in filename:

image

1 Like

Just a side note:
if you have a number of single frame pictures, you would like to show as a video and they are named like

name_0000
name_0001
etc.
(number of digit dont mater)
Just import the first picture and check the image sequence in properties
image
Then shotcut will work with all the images as a video clip

Much easier to work with, than having every single picture on the timeline.

2 Likes

Problem is I need to re-position each picture individually to allign so canā€™t use this trick for these particular projects.

a frame is smallest unit you can have, you canā€™t spilt it, so it donā€™t make much sense to be able to zoom in so one single frame fills so much on the timeline.

In you case i would use the playlist and add all the picture to it and reorder picture with Ctrl-Shift-Up/Down and the render the playlist and use this video to add extra stuff in another project.

Of course, but sometimes you need precision in moving or cutting clips and with the new limit thereā€™s not much place to do 1 frame exact edge of clip drag&drop.

I didnā€™t use the right word here, by reposition I mean use the SPR filter on each photo, this is why they need to remain individually in the timeline as I keep refining them (+adding/deleting) depending on the overall movement.

Qt major version changed. Not major; will not be addressed.

This is at max zoom, I need 2-4x more

Not major, intentional, and will not be changed. Learn to deal with the new limits. Each clip has a tooltip.

This change is being reverted because it is often causing a crash. We can try upgrading the SDL2.dll in the version after next, which you can do manually for now by downloading from libsdl.org if you choose.

This seems to work as expected for me.

You can add filter to media in the playlist, the playlist is like a track, just displayed a little different
You can make your whole video on the playlist, if only one track is needed.