[M4IF Technotes] layout of YUV files

[Herbert Thoma]

Hi Stefan,
"Stefan A. Goor" schrieb:
> 
> Hi Everyone,
> I have a problem that's been baffling me for a few days.
> I'm trying read the YUV reference clips such as akiyo_qcif.yuv etc in terms
> of Y U and V components (or more accuractely Y Cr Cb, so I'm told) and
> convert them to RGB format.  I know there are tools available to do this but
> I am specifically coding in java and would like to be able to control
> exactly what is going on in terms of pixels.
> Basically the problem occurs when I do the conversion.  My understanding is
> that these files are planar YUV format (i.e. YYYYYYY....UUUU....VVVV) and
> not interlaced (i.e. YUYVYU...... etc).  Taking into account that all the
> frames are QCIF, the format is 4:2:0 and each Y U or V component is 1 byte,
> I assume that reading the first 25344 (176x144) bytes is the Y component,
> the next 6776 (88x77) bytes are U and the next 6776 bytes are the V
> component and so on and so forth for each frame.

144 / 2 = 72, therefore the U and V components are 88 * 72 = 6336 bytes.
Your conversion might work, if you use these numbers.
 Herbert.
> I separated the components out into individual files and used Paintshop to
> display these as raw greyscale images and it seems to coincide with what I
> understood but as you can see from the pictures on
> http://www.netsoc.ucd.ie/~stefg/Qs/Question.html.  The U and V files are the
> same size as the Y component because I scaled them up a factor of 4 to
> convert them to QCIF size i.e.
> 
> a b ..   =              a a b b ..
> c d ..          a a b b ..
> . .                     c c d d ..
> . .                     c c d d ..
>                         . . . .
>                         . . . .
> 
> Having this data, I tried to convert it to RGB using the following equations
> that I found on many websites so I'm guessing they're correct:
> 
> int R_value = (int)((1.164 * (Y_value - 16)) + (1.596 * (V_value - 128)));
> int G_value = (int)((1.164 * (Y_value - 16)) - (0.813 * (V_value - 128)) -
> (0.391 * (U_value - 128)));
> int B_value = (int)((1.164 * (Y_value - 16)) + (2.018 * (U_value - 128)));
> 
> and after each conversion any errors must be corrected using code such as:
> 
> if(value<0)
>         value = 0;
> else if (value>255)
>         value = 255;
> 
> But as you can the result (http://www.netsoc.ucd.ie/~stefg/Qs/Question.html)
> that I got from the conversion was just random garbage.
> As a result, I'm wondering is yuv data layed out in files with the first
> byte representing the top left corner pixel and the next byte the next pixel
> to the right etc?
> Does any one have any ideas, what went wrong? Are any of my assumption
> wrong?
> Sorry about the long-winded email, but I just wanted to state the problem
> clearly,
> Thanks for you help,
> Stefan
> 
> _______________________________________________
> Technotes mailing list
> Technotes   lists.m4if.org
> http://lists.m4if.org/mailman/listinfo/technotes

-- 
Herbert Thoma
FhG-IIS A, Studio Department
Am Weichselgarten 3, 91058 Erlangen, Germany
Phone: +49-9131-776-323
Fax:   +49-9131-776-399
email: tma   iis.fhg.de
www: http://www.iis.fhg.de/