This is kind of post that I'm still thinking why I did that 🤔. But remember that not every line of code should be serious. Keep it fun! 🎙
Let's make Rick Astley sing "Never Gonna Give You Up" in console.
I assume that you already have some .mp4 file. If not, you can use
pytube. Check out this snippet:
import pytube class YT_video(): def __init__(self, url): self.url = url youtube = pytube.YouTube(self.url) self.video = youtube.streams.get_highest_resolution() def download(self, dest_folder = 'downloads'): self.video.download(dest_folder)
import cv2 video = cv2.VideoCapture(video_path) while True: _, frame = video.read() cv2.imshow("video", frame) if cv2.waitKey(1) & 0xFF == ord("q"): break video.release() cv2.destroyAllWindows()
Ok now we have our core. We play video, but there is no sound, it is playing to fast and our goal is to have it in console.
Let's solve those problems one by one.
You can use
ffpyplayer package. I added to our code next part for handling sound. It's only music player that will start song in the background.
import cv2 video = cv2.VideoCapture(video_path) music_player = MediaPlayer(video_path) while True: _, frame = video.read() cv2.imshow("video", frame) if cv2.waitKey(1) & 0xFF == ord("q"): break video.release() cv2.destroyAllWindows()
But wait, Rick is dancing twice as fast as sings. That is not spreading joy 🤔.
To know FPS (Frames Per Second) we will use OpenCV. First we calculate how long should one frame last. Then we will wait a bit before displaying next one.
fps = video.get(cv2.CAP_PROP_FPS) seconds_per_frame = 1 / fps
Now add small wait in the end of out While.
import cv2 video = cv2.VideoCapture(video_path) music_player = MediaPlayer(video_path) while True: frame_t_start = time.time() _, frame = video.read() cv2.imshow("video", frame) if cv2.waitKey(1) & 0xFF == ord("q"): break while ( time.time() - frame_t_start ) < seconds_per_frame: pass video.release() cv2.destroyAllWindows()
Before we print it to console, we need to make our video grayscale. Use below to change frame. I've added some threshold to make it more 'binary' for console. You can play with values
treshold_type. Check docs of OpenCV to read more.
I set my
treshold_type for 3 and
treshold around 120.
frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) _, frame = cv2.threshold(frame, treshold, 255, treshold_type )
Now our image is a 2D matrix of numbers. We will replace every number with some character. Assuming that our grayscale is from 0 to 255, find a char that suits perfect.
Assume that our grayscale is such string:
GRAY_SCALE = "@$#*!=;:~-,. "
Method for mapping number to char will be:
def grayScaleNumber(num): scale_size = len(GRAY_SCALE) index = int((num / 255) * scale_size) return GRAY_SCALE[index]
Let's add simple print method instead of
One more thing here is to resize the video. OpenCV can handle it for us.
def printFrameInConsole(frame, height, width): console_out_dim = ( int(width / SCALE),int(height / SCALE)) frame = cv2.resize(frame, console_out_dim, interpolation = cv2.INTER_AREA) to_print = '' for row in frame: to_print += ' '.join([ grayScaleNumber(num) for num in row.tolist()]) + "\n" sys.stdout.write(to_print) sys.stdout.flush()
Remember to play with it a bit. Check out some different thresholds and threshold types. Find the best Rick Astley for yourself.
I will upload some video as soon as I can. Now admire console Rick 😅
Inspired by PLED.