Problem with an “echo” while playing a video

Hi all

In my experiment (Win10, version 2020.1.1) I play a series of video clips separated between blocks (super basic code), all mp4 format. There are 3 types of blocks: audio-only blocks (which are technically audiovisual but with a black screen I placed in the video-editing software to replace the talking head), video-only blocks, and regular audiovisual blocks.

The problem – sometimes an “echo” appears with the audiovisual clips: The clip plays normally and then, in the end, the last 2-3 words of the sentence will suddenly repeat. This repetition is in the auditory modality only! (the visual component will keep playing normally). This happens at random and not with specific clips (so it’s not a problem with the file) and does not happen every time. It only happens with the audiovisual clips but not with the auditory-only clips (which, as mentioned, are basically the same in every parameter I can think of – e.g., when exporting from the video editing program I export both audio and video).

I have absolutely no idea what the problem can be, don’t even know where to begin… :frowning:

I’d very much appreciate your help guys!
Thanks
Maya

2 Likes

This sounds a bit like two copies of the video are playing. Only one would be seen on screen but the second can be heard when the first one finishes. Is that possible?

Hi

I don’t think so because if that was the case - wouldn’t there always be we two copies being played, or at least at the same spots?

Hi! I have this exact same problem :sob: :sob:

Did you manage to solve it???

I have the same problem. Somebody else (Videos are getting cut off or audio is wrong at the very end of the presentation) suggested to switch to mp4-files, but that didn’t do the trick, either. Have you made any progress?

@Benedict and @Eugene_Hwang
What version of PsychoPy are you using?

Is this for an online study or local?

Please could you show the code or components you are using to play the movie files

Hey - thanks for picking up on this.
I’m running it locally and I’m using the builder, so I’m not sure whether the code is reproducible, but it should be the part below. It’s inside a loop called “trials1”, and following the videos there’s another routine with questions about each video (inside the loop). (There’s another loop (“showvideo”) just around the videos, but that one doesn’t do anything atm.
Let me know if you need any more information. Thanks!

------Prepare to start Routine “Video”-------

    continueRoutine = True
    # update component parameters for each repeat
    movie = visual.MovieStim3(
        win=win, name='movie',
        noAudio = False,
        filename=video,
        ori=0.0, pos=(0, 0), opacity=None,
        loop=False,
        size=(1200,900),
        depth=0.0,
        )
    # keep track of which components have finished
    VideoComponents = [movie]
    for thisComponent in VideoComponents:
        thisComponent.tStart = None
        thisComponent.tStop = None
        thisComponent.tStartRefresh = None
        thisComponent.tStopRefresh = None
        if hasattr(thisComponent, 'status'):
            thisComponent.status = NOT_STARTED
    # reset timers
    t = 0
    _timeToFirstFrame = win.getFutureFlipTime(clock="now")
    VideoClock.reset(-_timeToFirstFrame)  # t0 is time of first possible flip
    frameN = -1
    
    # -------Run Routine "Video"-------
    while continueRoutine:
        # get current time
        t = VideoClock.getTime()
        tThisFlip = win.getFutureFlipTime(clock=VideoClock)
        tThisFlipGlobal = win.getFutureFlipTime(clock=None)
        frameN = frameN + 1  # number of completed frames (so 0 is the first frame)
        # update/draw components on each frame
        
        # *movie* updates
        if movie.status == NOT_STARTED and tThisFlip >= 0.0-frameTolerance:
            # keep track of start time/frame for later
            movie.frameNStart = frameN  # exact frame index
            movie.tStart = t  # local t and not account for scr refresh
            movie.tStartRefresh = tThisFlipGlobal  # on global time
            win.timeOnFlip(movie, 'tStartRefresh')  # time at next scr refresh
            movie.setAutoDraw(True)
        
        # check for quit (typically the Esc key)
        if endExpNow or defaultKeyboard.getKeys(keyList=["escape"]):
            core.quit()
        
        # check if all components have finished
        if not continueRoutine:  # a component has requested a forced-end of Routine
            break
        continueRoutine = False  # will revert to True if at least one component still running
        for thisComponent in VideoComponents:
            if hasattr(thisComponent, "status") and thisComponent.status != FINISHED:
                continueRoutine = True
                break  # at least one component has not yet finished
        
        # refresh the screen
        if continueRoutine:  # don't flip if this routine is over or we'll get a blank screen
            win.flip()
    
    # -------Ending Routine "Video"-------
    for thisComponent in VideoComponents:
        if hasattr(thisComponent, "setAutoDraw"):
            thisComponent.setAutoDraw(False)
    movie.stop()
    # the Routine "Video" was not non-slip safe, so reset the non-slip timer
    routineTimer.reset()
    thisExp.nextEntry()
    
# completed showvideo repeats of 'showvideo'

Oh, and version 2021.2.3, sorry!

By code or component I actually meant a screenshot of your movie component (any tabs where you have made changes to the default) rather than code created automatically from it. It looks like you are setting a movie component called movie in a routine called Video as the contents of a variable called video every repeat (which I assume is a column in your spreadsheet). I also assume that your experiment is set to use pixels.

Is the issue that you are getting two copies of the audio track?

Do you have any code components within the loop, or any other components that could produce a sound other than movie?

Sorry, my bad! See below for the screenshots:

These are the settings for the movies:
settings1
settings2

So basically I’m loading a video from a spreadsheet and then asking the corresponding question.

Regarding my issue:
The video plays perfectly fine and the audio is correct, right until there is only one second remaining. For the last second, it glitches back: The video keeps running, but the audio jumps back a second. (For example, for a video that is 20 seconds long, everything is fine until second 19, but then for the last second, you hear the sound that belongs to second 18 while seeing the correct video).

Again, many thanks!

Regarding the general structure:

This is the loop:
Loop

There is an additional code element in the loop trials1, because for some videos, I want to ask two questions. In these trials, the variable video is called “LEER” (meaning empty in German), showvideo is set to 0 and no video is shown, but we jump directly to the question.

However, I don’t see how this might cause any interference with the audio…

Since you’re running this locally, please could you try different movie Backend settings, in case anything works better than moviepy for you?

I assume that you’ve checked the movies are fine when played outside PsychoPy.

The videos work fine when played in VLC, yes. I’ve tried converting them to different formats (like mp4), but the same problem occurs.

Regarding the backend setting, moviepy is the only one that works at all, the others let the experiment terminate before the first movie, giving me various error messages such as
“AttributeError: module ‘vlc’ has no attribute ‘CallbackDecorators’” (for VLC)
“AttributeError: module ‘vlc’ has no attribute ‘dll’” (for opencv)
" avbin.dll failed to load. Try importing psychopy.visual as the first library (before anything that uses scipy) or use a differentmovie backend (e.g. moviepy)." (vor avbin).

I’ll try to get avbin to run and retry then…

Well, I tried to get the other backend settings to run and failed miserably.

Right now, the only workaround I can think of is to add a black frame without sound to the end of the video…

1 Like

did this solve your problem?

Overall, yes. One of our RAs reported occasional troubles starting the experiment, but I don’t believe that was connected.

@nl3 In the latest version of psychopy, the ffpyplayer backend should provide the smoothest experience with videos

Adding a black video segment at the end worked for me (using PsychoPy v2021.2.3). Thanks for posting your solution.