VSE: Adding packed sound via python API produces strip with incorrect length #103662

New Issue

Antonio Vazquez · 2023-01-05T15:55:33+01:00

Antonio Vazquez commented

2023-01-05 15:55:33 +01:00

Broken: Blender 3.4.1

Steps to reproduce
#103662.blend

Open file
Run script

Look in console

It prints Strip duration is 1 which is incorrect.
This is because BKE_sound_info_get() does not work correctly with packed sound.

Broken: Blender 3.4.1 ![image.png](https://archive.blender.org/developer/F14115563/image.png) **Steps to reproduce** [#103662.blend](https://archive.blender.org/developer/F14136196/T103662.blend) - Open file - Run script # Look in console It prints `Strip duration is 1` which is incorrect. This is because `BKE_sound_info_get()` does not work correctly with packed sound.

Antonio Vazquez commented

2023-01-05 15:55:33 +01:00

Added subscriber: @antoniov

Antonio Vazquez commented

2023-01-05 16:08:41 +01:00

Added subscribers: @mendio, @pepe-school-land

Pratik Borhade commented

2023-01-06 07:02:12 +01:00

Added subscriber: @PratikPB2123

Pratik Borhade commented

2023-01-06 07:02:12 +01:00

Changed status from 'Needs Triage' to: 'Needs Developer To Reproduce'

Pratik Borhade commented

2023-01-06 07:02:12 +01:00

Hi, can confirm in uploaded file for versions 3.2 and later.
But I'm not able to replicate this from scratch. @antoniov , can you share the python code/command?

Hi, can confirm in uploaded file for versions 3.2 and later. But I'm not able to replicate this from scratch. @antoniov , can you share the python code/command?

Pratik Borhade commented

2023-01-06 07:25:48 +01:00

Added subscriber: @iss

Pratik Borhade commented

2023-01-06 07:25:48 +01:00

git bisect points to: 302b04a5a3
@iss ^

git bisect points to: 302b04a5a3 @iss ^

Pratik Borhade changed title from ~~VSE: Strip bar is displayed before Offset position~~ to Regression: VSE: Strip bar is displayed before Offset position

2023-01-06 07:28:18 +01:00

Antonio Vazquez commented

2023-01-06 11:26:47 +01:00

@PratikPB2123 The script is complex and it's part of the new version of the Storypencil add-on.

here an extract

def send_sound_strip(s, dest_scn):
    '''recreate sound strip in another scene
    :dest_scn: scene destination
    :return: newly create sound strip
    '''

    if s.type != 'SOUND':
        return
    vse = dest_scn.sequence_editor
    ns = vse.sequences.new_sound(name=s.name, filepath=s.sound.filepath, channel=s.channel, frame_start=int(s.frame_start))
    ns.sound = s.sound # reget the same sound source

    for attr in ('pitch',
                 'pan',
                 'show_waveform',
                 'speed_factor',
                 'volume',
                 'mute'):
        if hasattr(s, attr):
            setattr(ns, attr, getattr(s, attr))
    if ns.volume == 0:
        ns.volume = 1
    return ns


- s must be a sound type strip
- tgt_scene is the target scene
ns = send_sound_strip(s, tgt_scene)
if ns:
    ns.frame_start = new_start
    ns.frame_offset_start = s.frame_offset_start
    ns.frame_final_duration = s.frame_final_duration

The script is adding a sound strip to a different scene, I mean, there is a VSE scene where you are editing the storyboard and when you switch to the Strip scene (you have one strip for each scene in the VSE), the audio is copied to the Edit scene.

If you need the full script I can provide you, but sure you will need some "training" to use it. Let me know if you need it and I can help you with the test.

@PratikPB2123 The script is complex and it's part of the new version of the `Storypencil` add-on. here an extract ``` def send_sound_strip(s, dest_scn): '''recreate sound strip in another scene :dest_scn: scene destination :return: newly create sound strip ''' if s.type != 'SOUND': return vse = dest_scn.sequence_editor ns = vse.sequences.new_sound(name=s.name, filepath=s.sound.filepath, channel=s.channel, frame_start=int(s.frame_start)) ns.sound = s.sound # reget the same sound source for attr in ('pitch', 'pan', 'show_waveform', 'speed_factor', 'volume', 'mute'): if hasattr(s, attr): setattr(ns, attr, getattr(s, attr)) if ns.volume == 0: ns.volume = 1 return ns - s must be a sound type strip - tgt_scene is the target scene ns = send_sound_strip(s, tgt_scene) if ns: ns.frame_start = new_start ns.frame_offset_start = s.frame_offset_start ns.frame_final_duration = s.frame_final_duration ``` The script is adding a sound strip to a different scene, I mean, there is a VSE scene where you are editing the storyboard and when you switch to the Strip scene (you have one strip for each scene in the VSE), the audio is copied to the Edit scene. If you need the full script I can provide you, but sure you will need some "training" to use it. Let me know if you need it and I can help you with the test.

Richard Antalik commented

2023-01-06 12:35:25 +01:00

Changed status from 'Needs Developer To Reproduce' to: 'Needs User Info'

Richard Antalik commented

2023-01-06 12:35:25 +01:00

I can confirm there is difference in strip appearance between versions, but I don't think this is actual bug.

Applying following code produced correct result:

    ns.frame_start = new_start
    ns.frame_offset_start = s.frame_offset_start
    ns.frame_final_duration = s.frame_final_duration

One thing that is odd is that in sample file seq->len is 1, which seems to produce incorrect end offset and strip ends up inverted. This can happen if filepath argument of new_sound() is not valid.

Duration of sound strip can be fixed by triggering update via scene FPS property. But the strip is then even more broken than before, but this is bit different issue, see D9997.

@antoniov Can you check if sound file is correct? Or perhaps look at strip.frame_duration property during strip execution - it should not be 1.

I can confirm there is difference in strip appearance between versions, but I don't think this is actual bug. Applying following code produced correct result: ``` ns.frame_start = new_start ns.frame_offset_start = s.frame_offset_start ns.frame_final_duration = s.frame_final_duration ``` One thing that is odd is that in sample file `seq->len` is 1, which seems to produce incorrect end offset and strip ends up inverted. This can happen if `filepath` argument of `new_sound()` is not valid. Duration of sound strip can be fixed by triggering update via scene FPS property. But the strip is then even more broken than before, but this is bit different issue, see [D9997](https://archive.blender.org/developer/D9997). @antoniov Can you check if sound file is correct? Or perhaps look at `strip.frame_duration` property during strip execution - it should not be 1.

Antonio Vazquez commented

2023-01-06 16:15:04 +01:00

@iss The sound strip looks right in the target scene. If you play it, you get the right sound in the right place. About duration, in the UI I can see it's 42, the correct number.

Maybe I need to run some type of update to refresh Strip information.... any idea about this? You say FPS update the strips... what is your advice, change FPS to something and then back to the right value?

EDIT: It looks reload strips fix the issue.

@iss The sound strip looks right in the target scene. If you play it, you get the right sound in the right place. About `duration`, in the UI I can see it's 42, the correct number. Maybe I need to run some type of update to refresh Strip information.... any idea about this? You say `FPS` update the strips... what is your advice, change FPS to something and then back to the right value? EDIT: It looks reload strips fix the issue.

Richard Antalik commented

2023-01-08 18:42:02 +01:00

In #103662#1470250, @antoniov wrote:
@iss The sound strip looks right in the target scene. If you play it, you get the right sound in the right place.

Ah did not realize there are more scenes... So I will try to run the script to reproduce. The issue could be, that the sound is packed.

About duration, in the UI I can see it's 42, the correct number.

Correct, but duration in UI is calculated from whole strip duration minus offsets. Internally the strip had duration of 1 frame.

It looks reload strips fix the issue.

Never used that operator so did not come to my mind, but glad there is solution to fix existing file. Still would like to prevent such issue.

> In #103662#1470250, @antoniov wrote: > @iss The sound strip looks right in the target scene. If you play it, you get the right sound in the right place. Ah did not realize there are more scenes... So I will try to run the script to reproduce. The issue could be, that the sound is packed. > About `duration`, in the UI I can see it's 42, the correct number. Correct, but duration in UI is calculated from whole strip duration minus offsets. Internally the strip had duration of 1 frame. > It looks reload strips fix the issue. Never used that operator so did not come to my mind, but glad there is solution to fix existing file. Still would like to prevent such issue.

Richard Antalik changed title from ~~Regression: VSE: Strip bar is displayed before Offset position~~ to VSE: Adding packed sound via python API produces strip with incorrect length

2023-01-09 18:56:45 +01:00

Richard Antalik commented

2023-01-09 18:56:45 +01:00

Changed status from 'Needs User Info' to: 'Confirmed'

Bastien Montagne added this to the Core project 2023-02-09 15:43:54 +01:00

Bastien Montagne removed this from the Core project 2023-02-09 18:20:29 +01:00

Philipp Oeser removed the

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

VSE: Adding packed sound via python API produces strip with incorrect length #103662

Look in console