search

mvcisback / SSLVC

Sound Source Localization using Visual Cues
4 stars 1 forks source link

Video 1 Re-sized #13

Closed ghost closed 9 years ago

ghost commented 10 years ago

Feel free to move it where it should be. I'm going to upload a bunch of stuff to this link later today.

https://www.dropbox.com/sh/erms26lcem5hv0o/AABYGDXmO5c9g-D-p5L4rCc5a?dl=0

mvcisback commented 10 years ago

Is this just going to be resized videos that are easier to work with?

ghost commented 10 years ago

yeah, I think I will also put up a .mat file that has all the frames and associated audios for different colormaps and you can just pick one and play with that.

ghost commented 10 years ago

@mvcisback and @ffaghri1 okay, so here are some information you need if you use the mat files. I haven't verify them yet. Size of the original video is 200(w)_300(h)_3(colors)_472(frames). size of the video is divided to 472_1473 samples, at sampling rate 44100. If you use YUV,HSV or RGB then all channels are put below each other and vectorized for each frame, that will give you a 180000_472(frames) for all frames. Putting audio below this, which is a mat file I put in the dropbox, gives you 181473_472(frame). So, the first 180000 rows are video pixels and you can reshape it to 200_300_3 to see it and the rest of that is the audio which you can hear it with sampling rate 44100.

ghost commented 10 years ago

oh and gray is obviously 60000_472 video and 472_1473 audio, since there is only one channel.

ffaghri1 commented 10 years ago

@ramili could you tell us how to access the data. I loaded one of the .mat files as video_sound variable with size 472x61473. How do I look into a particular frame or audio part? how is it shaped?

ghost commented 10 years ago

So, there are 472 frames. Each are 200_300_1 gray color. And each audio is 1473 samples per frame, they're energy of the audio signals. So to reshape you say, Reshape(video_sound(i,1:200_300),200,300); %video Plot(video_sound(i, 60000+1:end));%audio If you use the three channels, then it's 200_300*3.

On Wednesday, October 29, 2014, Faraz Faghri notifications@github.com wrote:

@ramili https://github.com/ramili could you tell us how to access the data. I loaded one of the .mat files as video_sound variable with size 472x61473. How do I look into a particular frame or audio part? how is it shaped?

— Reply to this email directly or view it on GitHub https://github.com/mvcisback/SSLVC/issues/13#issuecomment-61042693.

Thanks, Best Regards, Ramin

mvcisback commented 10 years ago

We're going to redo this for new videos.