Infoseite // YUV Farbsampling



Frage von LuFaHa:


Dear Slashcam community
in my following questions are the professionals in the video - and TV technology asked:

We know all of the television signal YUV (YCrCb means synonymous). This is in a "channel" the brightness (luminance) and in two other "channels" each Farbdifferenzsignal (Chrominance to red, to blue chrominance) transferred.
When the digitization is now in the usual DV - standard with a Farbsampling of 4:2:0 worked, which means (as we know of several contributors) that the Y signal with full Resolutionabgetastet or in DV-Picture will be saved, while this is for the two Farbdifferenzsignale (Cr and Cb) with only a quarter of the Resolutiongeschieht.
http://www.dma.ufg.ac.at/app/link/Basics%3AVideo.Postproduction/module/3927
Wenn wir nun davon ausgehen, dass ein PAL-Picture eine Resolutionvon 720 mal 576 Bildpunkten ( Pixel ) besitzt, bedeutet dies, dass wir einen Y-Kanal with dieser Resolutionbesitzen and zwei weitere Farbdifferenzkanäle with einer Resolutionvon je 360 mal 288 Bildpunkten (im Produkt genau ein Viertel des anderen Produkts - 720*576=414720, 288*360=103680).

First question: Is my idea the extent or are somewhere in my reasoning inconsistencies arise?

Next, I wanted my idea with an example showing what happened to me but have not yet succeeded. I have a picture from a normal DV digital compressed video (*. avi) extracted (in Adobe Premiere 2.0 as * bmp without changing the aspect ratio of pixels and exported Halbbildverarbeitung).
Here's the picture:
zum Bild
Now I wanted to verify this theory that I am with any program, the three channels separately consider (as an 8-bit grayscale image). It actually should not matter that the picture in the meantime, as RGB is stored, the individual can Frabsysteme my knowledge, but easily converted into each other.
Looking at the individual channels, I would have expected the following:
The Y-channel is a normal 720 * 576 pixel Picture dar.
The other two channels are shown individually or jointly with another picture of Resolutionvon 720 * 576 pixels (because the channel size in a match must be *. bmp), but with the difference that the pixels, as it were twice as large, ie, four pixels the first image of (relative).

Unfortunately it was not possible to me, this attempt, since I with no program (either Photoshop or Premiere or After Effects with ColorFinesse) created, the individual channels YUV (YCrCb) activity.
The only other possibility I came across the LAB mode in Photoshop On: Again, there is indeed the Luminanzkanal difference of two signals separated. These difference signals are not with those obtained in YUV identical, but my thought would be after the secession of the LAB luminance remain the same as for the secession of the Y Knals (probably wrong in my opinion identical to the LAB-Luminance - Luminance is equal luminance ???).
Here my result pictures:
LAB luminance channel
zum Bild
(as expected 720 * 576 pixels with normal resolution):
LAB-a color:
zum Bild
LAB b_Farbkanal:
zum Bild
(both can still clearly recognize the half and are 720 * 576 with true dissolved, the expected Verpixelung remains but there are compression artefacts visible, but four neighboring pixels may have differing shades of gray)

Second question: anyone know an easy to use tool or sonsteinen way to make a DV movie Picture in his three YUV channels to decompose and individually as 8-bit grayscale channel?

Third question (s): Where is

Space


Antwort von LuFaHa:

Small addition:


Space


Antwort von HeikoS:

Can you find the above times magnified region selected post here? I can only Farbdifferenzbildern on the corresponding theory of large blocks detect.

Space


Antwort von Stefan:

Unfortunately it was not possible to me, this attempt, since I with no program (either Photoshop or Premiere or After Effects with ColorFinesse) created, the individual channels YUV (YCrCb) activity.

Not tested, only wanted ( "yuv viewer" @ google) and found:

* Https: / / yuvviewer.dev.java.net / (Mac binary)
* http://www.ee.columbia.edu/~ywang/Links/index.html (including source) and https: / / skye.ee.columbia.edu / ~ ywang / Research / YUVGenius.html
* http://www.garydickson.co.uk/software/yuvfileplayer/
* http://freshmeat.net/projects/yay/ (Debian Linux, Windows)
* http://www.elecard.com/products/products-pc/consumer/yuv-viewer/ (40 $) in StreamEye drin http://www.elecard.com/products/products-pc/consumer/streameye-tools/ than 21 days Trial

That was the 1st Page of 171,000 hits ...

View Empire perhaps synonymous

* http://www.geocities.com/siwalters_uk/excmyuv.html Extract / Combine YUV for Virtualdub

Good luck
The thick Stefan

Space


Antwort von LuFaHa:

@ HeikoS
Yes, I can:
zum Bild
The blocks are told how to recognize, but clearly much bigger than four blocks and also not so plain. These are probably the post of DV compression.
But still, no theory of the corresponding four blocks recognizable - veilleicht exist in the real footage not synonymous, because in the post-DV Compression block this composition is dissolved. But it is still surprising when you consider that there are still the fields are visible (ie, the compression is not so much upset again).

@ Stefan
Thank you for your tips! Will I immediately look at the tools.
As you can see once more helpful as a second thought donor. I searched for appropriate tools for several hours gegoogelt, but I just completely wrong suchbegriffe used (programs, tools, YUV, YCrCb, extraction, channel, channels, individually, etc.) - but all possible combinations of these keywords have no helpful findings gelifert.
So again THANK YOU! - Even if I do not think that the results change ...

Space


Antwort von HeikoS:

404ERR

Space


Antwort von TheBubble:

For the 4:2:0 color format:

It is true that in the 4:2:0 format Farbdifferenzsignale horizontally and vertically only sampled half as often, as the corresponding brightness signal. You're assuming that the samples for the reconstruction of the frames is simply doubled (this would correspond to a "nearest neighbor" interpolation), which led to the expected of you "steps" would lead s.schrägen edges. It can be used for an improved image quality for scanning and reconstruction synonymous but other filter functions, where no staircase artifacts. In this case you would not observe the effect can be.

Besides: There are different variants of the 4:2:0 sampling and different codecs (such as JPEG, MPEG-1, MPEG-2, DV-PAL) use different synonymous variants.

Quote: the individual channels YUV (YCrCb) indicate

Try it with YIQ, or ...

Quote: Second question: anyone know an easy to use tool or sonsteinen way to make a DV movie Picture in his three YUV channels to decompose and individually as 8-bit grayscale channel?

... You write an RGB -> YUV conversion itself would simply the best variant.

Quote: As the work with 4:2:0 sampling fields?

Either the half or the half irgnorieren separate process. Some codecs take the decision synonymous dynamically, depending on image content.

Space


Antwort von LuFaHa:

Thank you s.alle!
I think now I have the knowledge that I needed:
Farbsampling is only the measured values, such as those in the actual picture will be used depends of the Compression and other factors. Normally, one after the other two fields and sampled only during the compression (before or after) are nested.

Even a few ambiguities:
How does the DV codec at all (image, like MJPEG - I know)? - What I think: What information is stored.
Contains a (DV -). Avi file individual compressed images, plus a soundtrack. Where the DV fields stored separately or together in a picture - in which said strip effects (in motion, on the monitor) occur?
Sometimes yes in Premiere under "properties of video material (or similar) things such as issuing program, pixel aspect ratio, etc. Halbbildreihenfolge displayed. If such metadata easy to DV material embedded or interpreted as the program only because of other properties (720 * 576 = DV => 1.0667, lower field first etc?

Space


Antwort von Stefan:

How does the DV codec at all (image, like MJPEG - I know)? - What I think: What information is stored.
Rel exactly, you can do that in the patents of Sony After Reading.

Contains a (DV -). Avi file individual compressed images, plus a soundtrack.

Could be. Always the sound in the data of DV frames saved. For DV AVI type2 also as an audio stream in the AVI file.

Where the DV fields stored separately or together in a picture - in which said strip effects (in motion, on the monitor) occur?

Together.

Sometimes yes in Premiere under "properties of video material (or similar) things such as issuing program, pixel aspect ratio, etc. Halbbildreihenfolge displayed. If such metadata easy to DV material embedded or interpreted as the program only because of other properties (720 * 576 = DV => 1.0667, lower field first etc?

Both. There are in DV 16:9 frame will be a name and a PAL / NTSC interlaced an identifier and / Progressive ID.

Good luck
The thick Stefan

Space



Space


Antwort von JackJohnson:

"LuFaHa" wrote: like this then the actual picture used depends of the Compression and other factors.

The sampled values are (lossy) compression.

"LuFaHa" wrote:
Normally, one after the other two fields and sampled only during the compression (before or after) are nested.

DV compresses a frame consisting of two fields.

"LuFaHa" wrote:
How does the DV codec at all (image, like MJPEG - I know)? - What I think: What information is stored.


The exact details of the standard you have to see. Basically:
1. Sampling
2. Block Education
3. Switch
4. DCT
5. Quantization
6. "Zigzag" and actual rearrangements Compression

"LuFaHa" wrote:
Contains a (DV -). Avi file individual compressed images, plus a soundtrack.

Depending on the type, the soundtrack for a second time to enable programs to easily access them Lönn.

"LuFaHa" wrote:
Where the DV fields stored separately or together in a picture - in which said strip effects (in motion, on the monitor) occur?

In a frame together (see above).

The aspect ratio is with the compressed image data stored so that it can be properly displayed. Similarly, the system used (50/60), both variants differ not only in the picture Resolutionund frequency, but are synonymous s.anderen Make internal treated differently.

Space


Antwort von TheBubble:

The contribution is ( 'The sampled values are ...") was of me.

Why forget the forum to write my application?

Space


Antwort von BugsBunny:

Even small comment on "individual channels YUV (YCrCb) show":
with VirtualDub and my Gradation Curves filter can do this very well and it's all free open source software.
The filter can be found here


Space





slashCAM nutzt Cookies zur Optimierung des Angebots, auch Cookies Dritter. Die Speicherung von Cookies kann in den Browsereinstellungen unterbunden werden. Mehr Informationen erhalten Sie in unserer Datenschutzerklärung. Mehr Infos Verstanden!
RSS Suche YouTube Facebook Twitter slashCAM-Slash