-
|
Hey folks, Suppose I want to display some video in the background of a sketch... what do you suggest as a replacement of Processing Foundation's Video library? |
Beta Was this translation helpful? Give feedback.
Replies: 3 comments 10 replies
-
|
First, why replace the Processing Foundation's Video library? It works just fine with py5. You'll need to add the video jar to your classpath, which can be done by creating a "jars" subdirectory and putting it in there. Note that py5 supports a user-defined If you don't want to use Processing's Video library, you can also use opencv. Everything you know how to do with webcams would be applicable here. |
Beta Was this translation helpful? Give feedback.
-
|
Hmm, you have great points. I guess it still feels a bit awkward for me to move jars around, and, as I got burned by Processing Foundation's libraries not working on different platforms in Processing Python mode anymore (both Serial and Video had problems on Windows), so my gut instinct told me I should move away from them, but this is just prejudice, I should definitely try the Processing Video library! |
Beta Was this translation helpful? Give feedback.
-
If the Video library had problems on Windows, I doubt py5 will solve them. What were the problems? How does the Video library work for Processing in the PDE? I imagine the tool the Video library uses to decode videos, gstreamer, might have some install complications on Windows, but I would be surprised if it couldn't work at all. |
Beta Was this translation helpful? Give feedback.
-
|
It had problems only on Python mode... I'll see if I can make it work with py5... but it will take some time. UPDATE: Today I could confirm video is working on Java but not on Python Mode on Windows: Next step is to try with py5... I have copied all the necessary files, tomorrow at Sesc I'll try and test it... |
Beta Was this translation helpful? Give feedback.
-
Sounds good! Keep me in the loop as you make progress. |
Beta Was this translation helpful? Give feedback.
-
|
After downloading and installing the Processing Foundation's Video library from the Processing 3.5.4 IDE, I copied the contents of the The Video library for Procesing 3.54 (based on GStreamer 1.16.2) kind of worked, freezing after some time. The newest Video for Processing 4.X is a bit different and I couldn't make it work at all. import py5
from processing.video import *
def setup():
global movie
py5.size(560, 406)
py5.background(0)
# Load and play the video in a loop
this = py5.get_current_sketch()
movie = Movie(this, 'launch2.mp4')
movie.loop()
def draw():
if movie.available():
movie.read()
py5.image(movie, 0, 0, py5.width, py5.height)
py5.run_sketch()
Defining a Using the Video library for Processing 4.2 I got this: |
Beta Was this translation helpful? Give feedback.
-
|
I am on my phone right now and have a few questions. If you are on windows, can you do this in an anaconda environment and install gstreamer through that? I have only used the video library on Linux. I know that library changed slightly from 3 to 4 (one change was a PR from me). The movie event function is I believe movie_event() (it is mentioned on the Sketch class page in the reference docs). If that doesn't get called, it might be a bug. |
Beta Was this translation helpful? Give feedback.
-
|
Curious that you mention it! Now I remember I tried Microsoft managed to wreck my dual boot Windows, somehow it broke without me logging in for months, so I'll only have access to a computer with Windows next Tuesday at Sesc (the community center closes Mondays). I'll see if I can set a new env with anacaonda then. |
Beta Was this translation helpful? Give feedback.
-
|
No luck on Linux either :( |
Beta Was this translation helpful? Give feedback.
-
|
Maybe we should avoid Processing Video and use OpenCV instead... this works, adapted from my webcam capture code, with a few tweaks: import py5
import cv2
import numpy as np
from py5 import create_image_from_numpy
movie = cv2.VideoCapture('launch2.mp4')
movie_width = int(movie.get(cv2.CAP_PROP_FRAME_WIDTH))
movie_height = int(movie.get(cv2.CAP_PROP_FRAME_HEIGHT))
fps = movie.get(cv2.CAP_PROP_FPS)
py5_img = None
def setup():
py5.size(movie_width, movie_height)
py5.frame_rate(fps)
def draw():
global py5_img
success, frame = movie.read() # frame is a numpy array
if success:
# gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
# edges = cv2.Canny(gray, 100, 80)
# edges_rgb_npa = cv2.cvtColor(edges, cv2.COLOR_GRAY2BGR)
# blended_rgb_npa = cv2.addWeighted(frame, 0.6, edges_rgb_npa, 0.4, 0)
# py5_img = create_image_from_numpy(blended_rgb_npa, 'RGB', dst=py5_img)
py5_img = create_image_from_numpy(frame, 'RGB', dst=py5_img)
# display image
py5.image(py5_img, 0, 0)
else: # If it can't read frame, try starting again
movie.set(cv2.CAP_PROP_POS_FRAMES, 0)
if py5.frame_count % 30 == 0:
py5.window_title(f'FR: {py5.get_frame_rate():.1f}')
def exiting():
print('over and out')
movie.release()
py5.run_sketch()More ideas at: https://vuamitom.github.io/2019/12/13/fast-iterate-through-video-frames.html |
Beta Was this translation helpful? Give feedback.
-
|
Does Processing Video work at all on Windows when run through the PDE? I'd be surprised if it did not. However, we want to provide folks with a way that works on all systems and does not cause headaches to get set up. OpenCV seems to be that solution. |
Beta Was this translation helpful? Give feedback.
-
|
This is what I got last week using the Processing IDE:
Yeah... I'd like to collect some tips on how to do things I used to do with Processing Video, like changing the movie speed... Someone (probably not me or you) could create a really nice "py5 library" nicely encapsulating OpenCV video features :) |
Beta Was this translation helpful? Give feedback.
-
Is there a bit of a stutter when you loop around to the beginning? It's a hard thing to get rid of.
Yes, or to start, write some documentation for this! |
Beta Was this translation helpful? Give feedback.
Maybe we should avoid Processing Video and use OpenCV instead... this works, adapted from my webcam capture code, with a few tweaks: