clever FFmpeg-GUI, small FFmpeg UI, fully portable, no installation.

[videoAI]

@ProWo

Tested Ok.
Thank you, and keep up the great work.

I think it would make sense to get rid of the [DNxHR] popup:
and please replace the other [DNxHD] popup as well.

Done.

[videoAI]

v 3.4.3.37
Bug#9 [Hevc Crash] fixed.
Tested Ok.

The point of exposing this bug is to show that 'Clever' still has short comings when handling indexes in arrays.
More validation to make sure the passed index value does not exceed the bounds of the array is highly recommended.

Thank you for fixing the bug, and keep up the great work.

A sample hevc test file causes a crash when dropped into clever.

Exception fixed.
Your sample is a raw hevc stream, not muxed into mp4, but renamed only to mp4.
Other editors like Avidemux can't load this file.

[videoAI]

@ProWo
Please find my replies in brackets below ..

Your sample is a raw hevc stream [Agreed], not muxed into mp4, but renamed only to mp4
[needed to rename the extension to mp4 to play it in my MPC player]
Other editors like Avidemux can't load this file. [but, I bet Avidemux didn't crash in the process]
I would recommend to use real video files instead.
[Well, according to the website that has hundreds of test files for all video formats,
it is a real video file used specifically for testing and my MCP player agrees as it plays it fine]

[oniiz86]

@ProWo I just noticed a minor misspelling with the misspelling of the "Multiplex" button, it has 2 ii's on the "Select Source File" screen.


Also when I add a PGS subtitle stream that I want to "Scan for errors", it mentions Errors found! & when I hover over the "View Errors" button it has a wrong description for the dialog pop up "Edit the Chapters File"


When I click on "View Errors" it tells me that the "Text Editor" is missing, I don't believe this is correct as it's a PGS image-based subtitle stream, not an SRT text-based subtitle.


I was wondering if your program can determine if subtitle streams are corrupt with the "Scan for errors" button, it says "Errors found!" but it doesn't mention what kind of errors it discovered?

[ProWo]

@ProWo I just noticed a minor misspelling with the misspelling of the "Multiplex" button, it has 2 ii's on the "Select Source File" screen.

Thx for the hint. Will be fixed.

Also when I add a PGS subtitle stream that I want to "Scan for errors", it mentions Errors found! & when I hover over the "View Errors" button it has a wrong description for the dialog pop up "Edit the Chapters File"

Will be fixed too.

When I click on "View Errors" it tells me that the "Text Editor" is missing, I don't believe this is correct as it's a PGS image-based subtitle stream, not an SRT text-based subtitle.

If errors are found, they are written to a text file. To view this file, you need a text editor (e.g., Notepad or Notepad++).

I was wondering if your program can determine if subtitle streams are corrupt with the "Scan for errors" button, it says "Errors found!" but it doesn't mention what kind of errors it discovered?

The subtitle stream is a stream like others; if errors are found, you need the texteditor to see them.

[Findu]

@ ProWo

Thanks for changing the Multiplex tooltips text like "Clear grid after muxing" so that it does not delete files any longer, but it still delete related processed files.

---

vp9_qsv speed test:

-map 0:0 -c:v libvpx-vp9 -b:v 0 -deadline realtime -crf 31 1x
-map 0:0 -c:v vp9_qsv -b:v 0 -preset veryslow 10x
-map 0:0 -c:v vp9_qsv -b:v 0 -preset veryfast 19x


av1_qsv speed test:

-map 0:0 -c:v libaom-av1 -crf 35 -cpu-used 8 1x
-map 0:0 -c:v libsvtav1 -crf 35 -preset 10 14x
-map 0:0 -c:v librav1e -qp 35 -speed 8 1x
-map 0:0 -c:v av1_qsv -preset veryslow 10x
-map 0:0 -c:v av1_qsv -preset veryfast 24x

Adding these addtional av1_qsv parameters also work for me:
-extbrc 1 -look_ahead_depth 40 -b:v 1M -bufsize 2M -rc_init_occupancy 512K -low_power 0 -adaptive_i 1 -adaptive_b 1 -b_strategy 1 -bf 4

But it may not be so interesting to add them to the gui?

---

Two questions:
Did you try the vpp_qsv filter in combination with _qsv encoding?
ffmpeg -h filter=vpp_qsv

Does hevc_mf encoding work for you?

---

ADD: Could you please increase the window size in Scan for errors!

[Attachment 87984 - Click to enlarge]

[ProWo]

Thanks for changing the Multiplex tooltips text like "Clear grid after muxing" so that it does not delete files any longer, but it still delete related processed files.

Should not be. Will change this.

EDIT: Done.

Two questions:
Did you try the vpp_qsv filter in combination with _qsv encoding?
ffmpeg -h filter=vpp_qsv

Not tested.

Does hevc_mf encoding work for you?

Do you mean hevc_amf?
No not working, but this is aspected, I'm on Intel platform.

ADD: Could you please increase the window size in Scan for errors!

The messages window size is the same for the whole app.
This is the unique message with longer text and appears only with a missing text editor.
I'll leave it as is is.

[Findu]

I dont mean AMD hevc_amf,
it is hevc via Media Foundation hevc_mf,
ffmpeg -h encoder=hevc_mf

[ProWo]

I dont mean AMD hevc_amf,
it is hevc via Media Foundation hevc_mf,
ffmpeg -h encoder=hevc_mf

Interesting. Never seen it. I will test it.

EDIT: Tested, doesn't work for me.

[Findu]

hevc_mf does not work for me either.

EDIT: Not on my AMD and Intel systems.

---

Combining qsv Encoding plus qsv filter slow down the speed for me.

For example the scale down from 4k h264 to 800x450 h264.

13,1x
ffmpeg -i Input.mkv -vf scale=800:450 -map 0:0 -c:v h264_qsv -global_quality 25 Output.mkv

9,45x
ffmpeg -hwaccel qsv -hwaccel_output_format qsv -i Input.mkv -vf vpp_qsv=w=800,vpp_qsv=h=450 -map 0:0 -c:v h264_qsv -global_quality 25 Output.mkv

[Findu]

@ ProWo

Implementing vp9_qsv, av1_qsv and mp3_mf.

The code I use for vp9_qsv is 10-25 times faster than libvpx-vp9, depending on the chosen preset (veryslow - veryfast):

Video ---> vp9 - Stream ID 0 from?C:\Users\Findu\Downloads\input.mp4?2242625?-hide_banner -loglevel error -stats -y -fflags +genpts -i "C:\Users\Findu\Downloads\input.mp4" -vf fps=24000/1000,colorspace=all=bt709:iall=bt709 -map 0:0 -c:v vp9_qsv -b:v 0 -preset veryfast -aspect 1.78 -an -sn -dn -metadata:g encoding_tool="clever FFmpeg-GUI" ?C:\Users\Findu\Downloads\input_V0.webm?

---

The code I use for av1_qsv is up to two times faster than libsvtav1, depending on the chosen preset (veryslow - veryfast):

Video ---> av1 - Stream ID 0 from?C:\Users\Findu\Downloads\input.mp4?2242625?-hide_banner -loglevel error -stats -y -fflags +genpts -i "C:\Users\Findu\Downloads\input.mp4" -vf fps=24000/1000,colorspace=all=bt709:iall=bt709 -map 0:0 -c:v av1_qsv -preset veryfast -aspect 1.78 -an -sn -dn -metadata:g encoding_tool="clever FFmpeg-GUI" ?C:\Users\Findu\Downloads\input_V0.mkv?

---

mp3_mf (Media Foundation) is between 2 and 2,5 times faster than mp3 when using the exact same settings and producing the exact same output on my AMD and Intel systems. It might be useful to add it. In the code simply replace mp3 with mp3_mf.

[ProWo]

----------------------------------------------------------------------------------
July 28, 2025 release version 3.4.4.
----------------------------------------------------------------------------------
Added:
Batch scan for errors.
Input check for idx field in join grid (numbers only).
Stay open memorized for video encoding, audio encoding, subtitle conversion, muxing.
Hardware acceleration decoding for batch scan for errors.
Rows limit to 9999 in batch/mux/join grids.
Adding folders prevented for batch/mux/join grid.
Hardware acceleration setting memozired for batch scan for errors.
New button colors to show that underlying grids contains elements.
Missing frequencys for audio codes.
hev1 for HDR recognition.
Change colorspace for prores codec.
Video-Bitrate info (if any).
Audiostream selection in preview (in case of multiple audios).
Clear grid after muxing option in mux grid.

Fixed:
Bug with to batch button missing after action and stay open checked.
Bug with missing no_analysis reset.
Bug with normal scan for erros.
Bug with destination folder (if to source is selected) in join grid.
Bug with "To batch" button remaining visible after clicking change (for change video DAR).
Bug in powersaveoff with partial batch task executing.
Bug with unused files in join grid.
Bug with missing opus frequencys.
Bug with resizable columns in grids (now prevented).
Bug in IVTC filterchain.
Bug with missing multiplex button with file can not be analyzed in start screen.
Bug with missing simple remux checkbox in mux grid.
Bug with missing quit button in start form, if no file is loaded.
Bug with filename handling in mux grid.
Bug with first file name check in mux grid.
Bug with shortest filename and first file check in mux grid.
Bug in save grid settings exiting batch tasks grid.
Bug with raw video renamed to mp4 file in start grid.
Bug with scan type variable.
Typo and wrong tooltip.
Bug with success label not cleared in mux grid.

Changed:
Save various variables handling.
Output file name generation (now it takes name from 1st selected stream/file) in join/mux grid.
Mux grid: hidden streams remains hidden now.
Autosize title column to fit content in mux grid
Errors descriptions with chapters.
Output filename check.
Input file extension check in join grid.
Hide settings with update options visible., added hev1 for HDR recognition
clevernorm filter m parameter range 1-20 now.
Start position handling.
Stream selection handling in mux grid.
First file name recognition in mux grid.
Change framerate without re-encoding (should work with all codecs now)
DNxHD and DNxHR encoding behavior.
fixed: bug with scan_type variable
Output file naming in mux grid.

----------------------------------------------------------------------------------

[ProWo]

Implementing vp9_qsv, av1_qsv and mp3_mf.

Thx for the info, will implement it.

EDIT:
Done. The mp3_mf works for me too, so I've tested it and it works fine.
Please test the implemented vp9_qsv and av1_qsv encoding, maybe as short sample, with all available pixel formats.

[Findu]

Pixel format qsv is not working for vp9_qsv and av1_qsv:

Video ---> vp9_qsv - Stream ID 0 from?C:\Users\Findu\Downloads\cut1.mp4?19978?-hide_banner -loglevel error -stats -y -fflags +genpts -i "C:\Users\Findu\Downloads\cut1.mp4" -vf fps=24000/1001,colorspace=all=bt709:iall=bt709:range=tv,form at=qsv -map 0:0 -c:v vp9_qsv -preset fast -aspect 1.78 -an -sn -dn -metadata:g encoding_tool="clever FFmpeg-GUI" ?C:\Users\Findu\Downloads\cut1_V0_#5.mkv?

Video ---> av1_qsv - Stream ID 0 from?C:\Users\Findu\Downloads\cut1.mp4?19978?-hide_banner -loglevel error -stats -y -fflags +genpts -i "C:\Users\Findu\Downloads\cut1.mp4" -vf fps=24000/1001,colorspace=all=bt709:iall=bt709:range=tv,form at=qsv -map 0:0 -c:v av1_qsv -preset fast -aspect 1.78 -an -sn -dn -metadata:g encoding_tool="clever FFmpeg-GUI" ?C:\Users\Findu\Downloads\cut1_V0_#2.mkv?

GUI error message:
[vf#0:0 @ 000001db1a5f0800] Error reinitializing filters!
[vf#0:0 @ 000002405a270800] Error reinitializing filters!

ffmpeg error message:
Impossible to convert between the formats supported by the filter 'Parsed_colorspace_1' and the filter 'auto_scale_0'

EDIT: "Stay open" at the "Video Encoding" page does not retain ist aktivate setting.

EDIT 2:
If you click on "To batch" in "Audio Encoding" with "Stay open" enabled, the "Task successfully completed" message appears in the bottom right corner. However, the message remains permanently visible. But it should disappear if you change any setting on the page. Like on the Multiplex page. This is required for the next click on "To batch"...

Could this be standardized on all pages? "To batch" receives the message on the page (so you can quickly do the next Batch) and the notification window only appears for buttons/tasks that directly process something, like Convert/Encode/Extract except for muxing.

[videoAI]

@ProWo

I was researching ffprobe and found it to be faster than FFmpeg in analyzing files. I think 'Clever' may benefit from using ffprobe in two areas:
1- File analysis when opening / dragging a file in.
2- Batch scan for errors, it is probably much faster than FFmpeg.

I found this info especially interesting ...

FFmpeg counts more frames than FFprobe:

FFmpeg, by default, sets constant frame rate mode for MP4 output.
When the input stream is VFR, ffmpeg will duplicate or drop frames
to generate a CFR stream. In the output stats, to the right of
frame=605, you can see dup=46, which indicates that ffmpeg added
46 duplicated frames. The short version is this happens when two
input frames are further apart than 1/FPS seconds, where FPS
represents the output frame rate. The output frame rate is set to
the detected input framerate (the tbr value), if not expressly set by
the user.

Add:

 -vsync vfr

to prevent frame duplication.

FYI ..

ffprobe is used to gather information about multimedia streams and containers.
It can extract detailed metadata, such as codec, bitrate, duration, and other
technical details, from audio, video, and other media files.

To identify errors in multimedia files using ffprobe, you typically look for specific
flags and options that can help diagnose issues. Here are some common
ffprobe commands used for error detection:

1. Basic Information and Errors:

ffprobe -v error -show_entries stream=codec_type,codec_name,codec_tag_string,codec_tag,profile,width,height,bit_rate,avg_frame_rate,time_base,start_pts,start_time,duration,nb_frames,disposition,codec_type,codec_name,codec_tag_string,codec_tag,profile,width,height,bit_rate,avg_frame_rate,time_base,start_pts,start_time,duration,nb_frames,disposition -of default=noprint_wrappers=1 input_file

This command provides detailed information about the streams in the file, which can help identify issues.

2. Show All Stream Information:

ffprobe -v error -show_streams input_file

This command shows detailed information about all streams in the file, which can help identify issues with specific streams.

3. Show Format Information:

ffprobe -v error -show_format input_file

This command shows detailed information about the format of the file, which can help identify issues with the container format.

4. Show Frames Information:

ffprobe -v error -show_frames input_file

This command shows detailed information about each frame in the file, which can help identify issues with specific frames.

5. Show Packets Information:

ffprobe -v error -show_packets input_file

This command shows detailed information about each packet in the file, which can help identify issues with specific packets.

6. Show Program Information:

ffprobe -v error -show_programs input_file

This command shows detailed information about the programs in the file, which can help identify issues with program streams.

7. Show Chapters Information:

ffprobe -v error -show_chapters input_file

This command shows detailed information about the chapters in the file, which can help identify issues with chapter markers.

8. Show Data Information:

ffprobe -v error -show_data input_file

This command shows detailed information about the data in the file, which can help identify issues with data streams.

9. Show Error Logs:

ffprobe -v error -show_entries stream=error -of default=noprint_wrappers=1 input_file

This command specifically looks for error entries in the streams, which can help identify issues directly.

[ProWo]

Pixel format qsv is not working for vp9_qsv and av1_qsv

I've removed them.

EDIT: "Stay open" at the "Video Encoding" page does not retain ist aktivate setting.

Thx, is fixed now.
After updating, close the app, open the portable.config file and delete the whole <save_set>/save_set> row.
Then save the file and start the app.

EDIT 2:
If you click on "To batch" in "Audio Encoding" with "Stay open" enabled, the "Task successfully completed" message appears in the bottom right corner. However, the message remains permanently visible. But it should disappear if you change any setting on the page. Like on the Multiplex page. This is required for the next click on "To batch"...

The message disappears with a new encode/to batch action, but because to batch is fast, you see it again. Not a bug.

Could this be standardized on all pages? "To batch" receives the message on the page (so you can quickly do the next Batch) and the notification window only appears for buttons/tasks that directly process something, like Convert/Encode/Extract except for muxing.

Will recheck this notification behavior.

[ProWo]

I was researching ffprobe and found it to be faster than FFmpeg in analyzing files. I think 'Clever' may benefit from using ffprobe in two areas:
1- file Analysis when opening / dragging a file in.

Is already used for this.

2- Batch scan for errors, it is properly much faster than FFmpeg.

Scan for erros doesn't work with ffprobe.

[videoAI]

found this online .. mentioning 'Clever' ..
https://yuhanmedia.com/resources/how-to-change-the-aspect-ratio-dar-of-a-video-on-wind...ods.html#part2

[Findu]

@ ProWo

When you abort a normal Scan for errors task (for the source file), click Various, Scan for errors, Batch, then you can't drop files to the grid and the HWA selection is missing.
When you go back to the Main page click Various, Scan for errors, Batch, everything is OK.

[videoAI]

@ProWo

Industry estimates that:
• 80–90% of Windows users favor GUIs over command-line tools
• 10–20% feel comfortable or proficient with the Windows CLI (PowerShell/CMD, etc)

For 'Clever' to become widely used by a large number of users, it needs to stream line its
workflow to become simple, fast, efficient, and intuitive to use.

Today, I scanned my video collection [3000 mp4s, mostly standard videos with
one video and one audio stream each]. I created a script to find all mp4s where
width > 1280 and ended up with a list of about 50 files, which I moved into a single folder.

I wanted to use 'Clever' to batch resize all 50 files to width =1280 [with 1.78 DAR] but
quickly found out it was next to impossible to do. 'Clever' doesn't even load multiple files
[except for Join/multiplex], what was the reasoning behind that?

Even loading the files one by one, resizing the video stream for each and every file would
have required another tedious and totally unnecessary step to mux the audio in.

Why is this step even required as an additional step?

Would it not make more sense to add a container dropdown to the video encode screen,
similar to the audio encode screen, and an optional checkmark to copy the audio stream,
in one go. This change should not affect [to batch] and the target file [batched or not] will
be ready to use right away.

Most FFmpeg commands, found online, already include the copy audio switch, and terminate
into a container like this:

ffmpeg -i input_video.mp4 -vf "scale=width:height" -c:a copy output_video.mp4

Is there any reason why this minor change [to optionally choose a container and add the audio stream in]
can't be done?

BTW, I ended up using Super Simple Video Converter to batch resize all 50 files in just 2 steps.

[davidt1]

I love the app for what I use it for. The author is the most responsive I have seen. Keep up the good work.

[mannequin80]

i see that av1_qsv available but it doesn't work currently?

[Findu]

@mannequin80

av1_qsv encoding requires Intel Arc series and Core processors from the 11th generation onwards.

https://en.wikipedia.org/wiki/Intel_Quick_Sync_Video
https://www.intel.com/content/www/us/en/docs/onevpl/developer-reference-media-intel-ha.../overview.html

[Findu]

@ ProWo

Thanks for adding the three additional encoders. They work well.

---

Sorry. I just noticed that the "Task successfully completed" message on the Multiplex page only disappears when you change something in the grid's Set column or selecting Simple remux.

Idea:
When the success message appears, could the next click inside the app window remove it? So, regardless if a selection is changed or not. Maybe that's easier to trigger/implement?

Or the message disappears after 2 or 3 seconds.

Or you enter a counter "Task # successfully completed".

---

How about a text change for this dialog?

[Attachment 88023 - Click to enlarge]

You have selected more channels than the original source audio stream. These additional channels will remain empty without remapping. Do you want to use Remap?

And could the Remap window open when you click on Yes?

---

For better readability, could you please add a space before the brackets?

[Attachment 88024 - Click to enlarge]

[Attachment 88025 - Click to enlarge]

[mannequin80]

@mannequin80

av1_qsv encoding requires Intel Arc series and Core processors from the 11th generation onwards.

obviously, i have one of those. i gave it a quick try but getting errors all the time.

[ProWo]

When you abort a normal Scan for errors task (for the source file), click Various, Scan for errors, Batch, then you can't drop files to the grid and the HWA selection is missing.
When you go back to the Main page click Various, Scan for errors, Batch, everything is OK.

Thx for the hint, fixed now.

[Findu]

@mannequin80

I did the integration testing on Intel Arc graphics and all works well.

But ProWo need the error messages to investigate ...

[ProWo]

I wanted to use 'Clever' to batch resize all 50 files to width =1280 [with 1.78 DAR] but
quickly found out it was next to impossible to do.

Not true.
Create a batch task with the 1st video, click batch tasks, drag the other 49 video into the grid and click execute.


[Attachment 88029 - Click to enlarge]

[ProWo]

Idea:
When the success message appears, could the next click inside the app window remove it? So, regardless if a selection is changed or not. Maybe that's easier to trigger/implement?

Will implement a timer.

You have selected more channels than the original source audio stream. These additional channels will remain empty without remapping. Do you want to use Remap?

Will be done.

And could the Remap window open when you click on Yes?

Of course.

For better readability, could you please add a space before the brackets?

OK

EDIT:
All done.

[videoAI]

I wanted to use 'Clever' to batch resize all 50 files to width =1280 [with 1.78 DAR] but
quickly found out it was next to impossible to do.

Not true.
Create a batch task with the 1st video, click batch tasks, drag the other 49 video into the grid and click execute.

... and I'll end up with 50 files with no audio .. do you see my point?