Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm using ffmpeg to take a mjpeg feed from an IP camera and convert it into a series of still jpeg images. Here's the command I'm using:

ffmpeg -i http://xxx:xxx@xx.xx.xx.xx/vid.mjpg -f image2 -vcodec mjpeg %05d.jpg

Does anyone know a way to specify the level of jpg compression when using ffmpeg to create stills?

share|improve this question

migrated from stackoverflow.com Aug 4 '11 at 4:31

This question came from our site for professional and enthusiast programmers.

up vote 26 down vote accepted

Here's what ended up doing the trick for me:

ffmpeg -i http://xx.xx/mjpg/video.mjpg -q:v 1 %05d.jpg

The -q:v 1 was the ticket. It sets the quality of the video encoder, where the -q means quality and the :v means video. Values range from 1 to 31, where lower means better.

share|improve this answer

$ ffmpeg -r 1/4 -i %03d.jpg -b 5000 -vcodec mjpeg -qscale 1 5000.avi

OR

$ ffmpeg -r 1/4 -i %03d.jpg -vcodec copy -qscale 1 copy.avi

Reference:

r       delay between next jpg, results 4 second
b       bitrate
vcodec  use jpeg encode
qscale  quality ratio
share|improve this answer
    
does qscale "compete" with the bitrate specifier, I wonder? – rogerdpack Jan 19 '13 at 17:45

You could try to export into PPM and use some other tool to convert into JPEG.

I looked into ffmpeg/libavcodec/mjpeg.c. I believe the quality is set to a fixed value.

Also you seem to convert a MJPEG video into JPEG still frames. I think in this case the code in ffmpeg/libavcodec/mjpeg2jpeg_bsf.c runs and the data isn't recoded. So the image quality wouldn't improve anyway.

This is the quantization table definition, I didn't see any reference to *val_?c where the values were scaled before use.

/* Set up the standard Huffman tables (cf. JPEG standard section K.3) */
/* IMPORTANT: these are only valid for 8-bit data precision! */
const uint8_t ff_mjpeg_bits_dc_luminance[17] =
{ /* 0-base */ 0, 0, 1, 5, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 };
const uint8_t ff_mjpeg_val_dc[12] =
{ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 };

const uint8_t ff_mjpeg_bits_dc_chrominance[17] =
{ /* 0-base */ 0, 0, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0 };

const uint8_t ff_mjpeg_bits_ac_luminance[17] =
{ /* 0-base */ 0, 0, 2, 1, 3, 3, 2, 4, 3, 5, 5, 4, 4, 0, 0, 1, 0x7d };
const uint8_t ff_mjpeg_val_ac_luminance[] =
{ 0x01, 0x02, 0x03, 0x00, 0x04, 0x11, 0x05, 0x12,
share|improve this answer
    
I'm not familiar with PPM, what is that? – weotch Aug 3 '11 at 22:28
    
Also, the images look better when I access the camera from the browser compared to the jpgs I'm creating. That's why I was thinking some compression was occurring. – weotch Aug 3 '11 at 22:31
    
PPM is simplest color image format one can imagine. Its just a header like "P6 640 480 255\n" and then the raw data. I just looked. Maybe it actually can't output this format. The only other option then would be -vc rawvideo. Did you try -quality? – whoplisp Aug 3 '11 at 22:41
1  
-vcodec ppm is needed to export ppm – OrangeDog Jan 10 '12 at 14:08

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .