[vlsi-nastava:2004] - Re: Pixel...za Gvozdena, a i ostale.... by "Aleksandar Milutinovic" <aca.mil@galeb.etf.bg.ac.yu>

Za PAL standard:

Na netu sam nasao neko objanjenje da se svaki drugi piksel na ekranu boji, a
ne svaki. To je prihvaceno usled smanjene osetljivosti oka na boje
(prostorna lokalnost na slici...).  U nastavku je deo jednog .pdf koji mozes
naci na donjoj adresi. Po ovome mi lici da je nemoguce svaki piksel vratiti
u RGB format. Koliko sam ja razumeo, nas se to ne tice, zar ne?

Mene muci druga stvar: kako se vrsi sinhronizacija na pocetak slike? Ja sam
shvatio da se oni semplovi salju jedan za drugim, i da mi moramo da ih
pratimo i da reagujemo kada dodje do odredjenih tacaka. Jesam li u pravu?

Pozdrav,
 Sale

... This sub-sampling is possible since the human eye is less

sensitive to color spatial resolution then to intensity spatial resolution.
The direct benefit is that a

4:2:2 data stream is 2/3 the overall data size as compared to a
non-decimated image.

Figure 4 illustrates the sampling process. As shown, every one of the 720
pixel positions sampled

has a Y, Cb, Cr value, except that every 2nd Cb and Cr components are
dropped (decimated). The

resulting BT656 4:2:2 data stream (gray box) is a simple concatenation of
these 8-bit luma and

chroma samples. The aggregate BT656 data stream rate is of 13.5 MHz + 2 x
6.75 MHz = 27 MHz.

4 © Digital Creation Labs Incorporated 2004 Rev 1.0 April 2004

Color Space Conversions

So far, we have explored the fundamentals of the analog composite video
standard as well as digital

video encoding techniques. As described previously, the color information
contained in a composite

video signal is decoded into the Y, Cb, Cr color space as per the ITU-R
BT656 standard. This color

space is very convenient for many types of image processing where operations
are mainly performed

on the black & white video component (in that case, the Y luma component can
be used directly). To

display color images on a PC, however, the more familiar RGB color space is
used. In this color

space, each pixel is described by an 8-bit value of its red, green and blue
components. In order to

perform a conversion from the Y, Cb, Cr color space to the RGB color space,
a simple transformation

can be applied:

R = 1.164 x (Y-16) + 1.596 x (Cr-128)

G = 1.164 x (Y-16) - 0.392 x (Cb-128) - 0.813 x (Cr-128)

B = 1.164 x (Y-16) + 2.017 x (Cb-128)

(R, G, B are gamma-corrected and ranging between [0-255])

Equation 1: (Y, Cb, Cr) to (R, G, B) color space transformation equations

----- Original Message -----
From: "Goran Lukic" <glukic@galeb.etf.bg.ac.yu>
To: <vlsi-nastava@titan.etf.bg.ac.yu>
Sent: Monday, December 27, 2004 5:19 PM
Subject: [vlsi-nastava] Pixel...za Gvozdena, a i ostale....


> 10-bitni izlazni signal Pixel treba da pokazuje tekuci pixel....
> Ako imamo Cb1 Y1 Cr1 Y2 Cb2 Y3 Cr3...  onda bi pixeli trebalo da budu:
>
> 1.  Y1 Cb1 Cr1
> 2.  Y2 Cb1 Cr1
> 3.  Y3 Cb2 Cr2
> 4.  Y4 Cb2 Cr2
> 5.  .........
>
> ...i tako dalje....kako bismo ih mogli konvertovati u odgovarajucu RGB
> strukturu....
>