Still trying to figure out about the NT 525 vs 486 vs 480 video lines thing

[Videogamer555]

I know 525 lines is all the lines in the picture, including the v-sync pulses, but then supposedly the "active video" has 486 lines, but most video capture hardware only has 480 lines that get saved to a video file on your computer. This leaves me wondiering what the extra six lines are for. And if in fact they don't contain any image data (but are simply left black), isn't it more reasonable to consider them to be actually part of the vertical blanking interval (and not the "active video" interval of the signal)? If the extra 6 lines are part of "active video" then the vertical blanking interval consists of 39 lines. And does closed caption fall within these (normally not displayed) extra 6 lines of active video? Or does closed caption data fall within the true vertical blanking interval?

And it gets more complicated than that. There's actually 2 vertical blanking intervals in a frame (one for each field). So how are the 39 VBI lines divided between the 2 fields? Does one field have 20 VBI lines, and the other field have 19 VBI lines? And which of those 2 sets of VBI lines contains the closed caption data?

And regarding the extra 6 lines of "active video", where exactly do those come in? At the begining of the first field, the ending of the first field, the beginning of the second field, or the of the second field? Or are they split between the 2 fields (3 in each field)?

And lastly, how are the lines numbered? Is the first line of video considered to be "line 0" or "line 1"? This will also determine the correct naming of the fields. If the first line in an image is "line 0" then the even field comes first. If the first line of video is "line 1" then the odd field comes first. I'd really like some clarification here as well.

[Videogamer555]

And I just noticed that that according to http://www.sxlist.com/techref/io/video/ntsc.htm each field has 253 lines of normal image (and the first field, which that site designates as field-0, has an extra half-line for a total of 506.5 lines per frame of active video (as opposed to the standard 486 lines of active video per frame that I've read about on other websites). Is it true that there are actually 506.5 lines of active video in each frame (and not the 486 lines that I've read about elsewhere)?

[lingyi]

This thread https://forum.videohelp.com/threads/377493-Questions-about-structure-of-NTSC-TV-video-signalhas a lot of info (most of which goes completely over head) about your inquiry.

[Videogamer555]

I have another question. When certain fields cut off the last line half way through, does that mean half the image line (such that the h-sync gets sent + half an image line)? Or does it mean half the total line gets (such that the h-sync gets sent + LESS THAN half of an image line)?

[jagabo]

Horizontal sync pulses are always at regular intervals. The first line of the top field only has image data in the second half. The last line of the bottom field only has image data in the first half. The other half of those lines are black.

[pandy]

Half line in PAL are not blanked and this is clearly visible in first half of line 23 - first half of this line carry WSS signaling which belongs normally to VBI signal (normally not visible) however in video codec this part of video (out of active video) is encoded as regular video data - it is common to see WSS stripes in many PAL captures.

[Videogamer555]

I've created a program that I believe creates a single frame (both fields) according to what I believe is correctly implemented NTSC standard. Here's what it creates:

There are 525 lines total (including both picture and V-sync regions) .
It is divided into 2 fields.

In the first field you have:
3 lines containing 2 equalization pulses each (for a total of 6 equalization pulses)
3 lines containing 2 V-sync pulses each (for a total of 6 V-sync pulses)
3 lines containing 2 equalization pulses each (for a total of 6 equalization pulses)
2 lines containing black image data (even though technically part of the image, it is not displayed, and is often used to hold digital data relating to the image)
240 lines containing the first set of interlaced image lines for the image that is to be displayed
11 lines containing black image data (even though technically part of the image, it is not displayed, and is often used to hold digital data relating to the image)
0.5 lines containing black image data (this is just padding to make sure that the fields line up)

In the second field you have:
3 lines containing 2 equalization pulses each (for a total of 6 equalization pulses)
3 lines containing 2 V-sync pulses each (for a total of 6 V-sync pulses)
3.5 lines containing 2 equalization pulses each in each full line, and 1 in the half-line (for a total of 7 equalization pulses)
2 lines containing black image data (even though technically part of the image, it is not displayed, and is often used to hold digital data relating to the image)
240 lines containing the second set of interlaced image lines for the image that is to be displayed
11 lines containing black image data (even though technically part of the image, it is not displayed, and is often used to hold digital data relating to the image)



When the 2 fields are combined so that a proper image is displayed, we now have:
0.5 lines containing black image data (this is just padding to make sure that the fields line up)
6 lines containing 2 equalization pulses each (for a total of 12 equalization pulses)
6 lines containing 2 V-sync pulses each (for a total of 12 V-sync pulses)
6.5 lines containing 2 equalization pulses in each full line, and 1 in the half-line (for a total of 13 equalization pulses)
4 lines containing black image data (even though technically part of the image, it is not displayed, and is often used to hold digital data relating to the image)
480 lines containing the image that is to be displayed
22 lines containing black image data (even though technically part of the image, it is not displayed, and is often used to hold digital data relating to the image)



So in total there are:
480 lines for the displayed image
484 lines of image data (including 4 lines of non-displayed image above the displayed image)
486 lines of image data (including 4 lines of non-displayed image above the displayed image, and 2 lines of non-displayed image below the displayed image)
506.5 lines of image data (including all non-displayed image lines, and the half-line of image padding)


I have no idea why it is often stated that there are 486 lines of image in tho NTSC standard. The really important numbers appear to be, 480, 484, and 506.5 lines. 480 is the the actual number of lines in the picture. 484 is counts the first 4 lines of non-displayed image above the displayed image, and 506.5 counts all of what are technically "image lines". 486 is not really a significant number at all. It only counts 2 additional non-displayed image lines below bottom of the displayed image. Are those 2 additional lines used for something important?



And one last question in this post. Are all of my facts here actually compatible with the NTSC standard?
If not, please let me know, because this info is part of what I'm using in my program that converts normal picture files into (what I hope are) NTSC compatible frames. I will need to know anywhere where my numbers might not be up to official NTSC specifications, because those will be corrections that I need to make in my program.

[JVRaines]

This is the final NTSC studio standard of the Society of Motion Picture and Television Engineers.

http://atarionline.pl/forum/?PostBackAction=Download&AttachmentID=6601

[Videogamer555]

This is the final NTSC studio standard of the Society of Motion Picture and Television Engineers.

http://atarionline.pl/forum/?PostBackAction=Download&AttachmentID=6601

Thanks for the document. That was very informative.

It also revealed something strange. The chroma standard for NTSC is based on YCbCr, not YIQ. That is, the QAM modulator is not fed with I and Q signals, but rather is fed with Cb and Cr signals. If I and Q signals are used, a phase offset of 33 degrees must be added to the QAM modulator. I always thought that NTSC used I and Q chroma signals, and that Europe's PAL standard used Cb and Cr signals. When did NTSC switch to using Cb and Cr chroma signals?

I thought that "I" meant "in phase", which means "this signal is in phase with the cosine signal", and that "Q" meant "quadrature", which means "this signal is 90deg to the cosine signal". In that case, there should be no need to offset the phase by an additional 33 degrees, yet it appears that when using IQ based chroma, you do indeed end up offsetting the chroma subcarrier's phase by 33 degrees. What's up with that?

[pandy]

https://www.itu.int/rec/R-REC-BT.1700-0-200502-I/en

[JVRaines]

It also revealed something strange. The chroma standard for NTSC is based on YCbCr, not YIQ. That is, the QAM modulator is not fed with I and Q signals, but rather is fed with Cb and Cr signals. If I and Q signals are used, a phase offset of 33 degrees must be added to the QAM modulator. I always thought that NTSC used I and Q chroma signals, and that Europe's PAL standard used Cb and Cr signals. When did NTSC switch to using Cb and Cr chroma signals?

I thought that "I" meant "in phase", which means "this signal is in phase with the cosine signal", and that "Q" meant "quadrature", which means "this signal is 90deg to the cosine signal". In that case, there should be no need to offset the phase by an additional 33 degrees, yet it appears that when using IQ based chroma, you do indeed end up offsetting the chroma subcarrier's phase by 33 degrees. What's up with that?

This is a studio standard which provides for using either Y/B-Y/R-Y or YIQ color encoding. The broadcast standard mandates YIQ.

[Videogamer555]

It also revealed something strange. The chroma standard for NTSC is based on YCbCr, not YIQ. That is, the QAM modulator is not fed with I and Q signals, but rather is fed with Cb and Cr signals. If I and Q signals are used, a phase offset of 33 degrees must be added to the QAM modulator. I always thought that NTSC used I and Q chroma signals, and that Europe's PAL standard used Cb and Cr signals. When did NTSC switch to using Cb and Cr chroma signals?

I thought that "I" meant "in phase", which means "this signal is in phase with the cosine signal", and that "Q" meant "quadrature", which means "this signal is 90deg to the cosine signal". In that case, there should be no need to offset the phase by an additional 33 degrees, yet it appears that when using IQ based chroma, you do indeed end up offsetting the chroma subcarrier's phase by 33 degrees. What's up with that?

This is a studio standard which provides for using either Y/B-Y/R-Y or YIQ color encoding. The broadcast standard mandates YIQ.

Where can I find the full broadcast (not studio) specification for NTSC? Is there a PDF file of this specification?

And are the 2 mathematically equivalent? I mean, is using YIQ with a constant 33deg phase offset, equivalent to using YUV with no phase offset?

And also I saw it mentioned in section A.4 about "equiband encocding, that "I" and "Q" signals can either have the same band width (known as "equiband"), or have "I" using a different bandwidth from "Q" (known as "nonequiband").
In section A.5 it says that when using equiband encoding, either YUV or YIQ can be used, but that for nonequiband encoding YIQ must be used. You have stated that YIQ must be used in the broadcast specification, but what about equiband encoding? Does the broadcast standard permit equiband encoding or does it require nonequiband encoding?

[JVRaines]

Google "NTSC equiband." I'm finding some references that say equiband was adopted for broadcasting back in 1990.

[Videogamer555]

Still I have 2 unanswered questions

Where can I find the full broadcast (not studio) specification for NTSC?

Is using YIQ encoding with a constant 33deg phase offset, equivalent to using YUV with no phase offset? That is, do they both produce the same signal (prior to any filtering for equiband or nonequiband final signal)?