Block Scan Order: Zig Zag or Alternate??? Please advise.

[ChachiFace]

Okay, here's yet another encoding question. I'm using CCE-Basic and there is an option for Block scan order to use zig zag or alternate. The default is alternate. I read that zig zag is really for progressive scan. Would it make a difference if I used zig zag even if the footage was not HD? Anybody familiar with these settings?

Please advise.

chachiface

[SaSi]

Don't confuse HD with progressive. Even normal camcorder or TV captures can be progressive, if they are de-interlaced or whatever.

The Zigzag scan patern is most suitable for progressive frames while the alternative method is optimized for interlaced frames.

You should always want to use zigzag for progressive and alternative for interlaced as each produces the best effect with each type of frame

[ChachiFace]

SaSi,

How do I know which one I'm supposed to use based on my source? The source is AVI caputured with a DAC. Is there a progarm I can load the AVI file into that would tell me if it's interlaced or progressive?

[FulciLives]

SaSi,

How do I know which one I'm supposed to use based on my source? The source is AVI caputured with a DAC. Is there a progarm I can load the AVI file into that would tell me if it's interlaced or progressive?

Anything you capture with your DAC will be interlaced be it from NTSC or PAL video sources.

You should leave it that way too.

So use ALTERNATE for the BLOCK SCAN ORDER

- John "FulciLives" Coleman

[ChachiFace]

Thanks FulciLives!!!

[FulciLives]

Thanks FulciLives!!!

Well something just occurred to me ...

It is possible that some of the NTSC video material you capture might be telecined which means you can restore it to progressive by doing an IVTC but that process can be a bit tricky.

LOL I just heard my girlfriend calling me a dork in my mind's eye.

How many people other than us video nuts can understand that sentance I just typed!

Anyways there are a few guides on doom9 about deinterlacing and IVTC

- John "FulciLives" Coleman

[ChachiFace]

Okay, I just ran a test (after all of these weeks) and encoded some AVI footage to .m2v with CCE using the ZIG ZAG scan order then encoded the same AVI footage using the ALTERNATE scan order. I burned both onto a DVD and watched on my TV (regular not HD) using a standard topset DVD player. After visually comparing both, I'd have to say that the ZIG ZAG was slightly cleaner looking (less artifacts) than the ALTERNATE encoded material.

Is it bad to use the ZIG ZAG scan order if your "supposed" to use ALTERNATE? From what I'm told in the previous responses to this post, I should be using ALTERNATE but I wanted to test ZIG ZAG since that's what my old DAZZLE encoder automatically did for regualar Analog to Digital VIDEO encodes.

[SaSi]

What kind of bitrate did you encode at?

Keep in mind that the encoding is about storing as much quality at a lower bitrate. Therefore, "poor" choices will only become visible at low bitrates.

Try to encode full DVD resolution at 2kbps or less VBR and you will see the difference.

If you encode at 8Mbps CBR, none of the numerous parameters and options makes much difference - I think...

[FulciLives]

Is it bad to use the ZIG ZAG scan order if your "supposed" to use ALTERNATE? From what I'm told in the previous responses to this post, I should be using ALTERNATE but I wanted to test ZIG ZAG since that's what my old DAZZLE encoder automatically did for regualar Analog to Digital VIDEO encodes.

Well respectfully ... we didn't make this shit up!

Those that have programmed MPEG encoders such as TMPGEnc and CINEMA CRAFT ENCODER have said "do this" or "do that" and it is all based on the MPEG-2 specs.

So why you want to "fight" this is beyond me.

- John "FulciLives" Coleman

[andie41]

I have wondered about this, as well. If I use TMPGenc, Bitrate viewer says it is Alternate; if I use Procoder, it is Zig Zag.

[ChachiFace]

Well respectfully ... we didn't make this shit up!

Those that have programmed MPEG encoders such as TMPGEnc and CINEMA CRAFT ENCODER have said "do this" or "do that" and it is all based on the MPEG-2 specs.

So why you want to "fight" this is beyond me.

- John "FulciLives" Coleman

I'm not fighting it Fulci, I just don't understand it. A "do this" or "do that" answer just doesn't cut it sometimes...this is when an explanation is needed. Can you explain to me how I can know for sure that the video I capture with a DAC-100 is indeed INTERLACED and not PROGRESSIVE?

[SaSi]

Don't take for granted that the MPEG2 encoding s/w are precice implementations without any flaws. In the contrary, they are programs that are full of flows (trust me, I know, I am a programmer at times and an expert in making flaws 8) )

Studying the scan order and reasoning behind it, makes perfect sense to have different scan orders for interlaced and progressive frames.

Using the wrong scan order will waste bitrate or quality (depending on encoding method).

If an encoder (s/w) produces the same quality regardless of the scan order, it is my humble opinion that this s/w is not exploiting the full potential of the selected bitrate in any case.

Don't get me wrong, I'm not saying the encoders are crap.

However, encoding at 1500kbps vs encoding at 6000kbps is a whole different story altogether. Optimization decisions would be totaly different and priorities would change. It's a miracle that these small and nice programs behave as they do.

We should just not take their behaviour as "the truth". They are close to it, each one from a different angle.

ChachiFace, I believe you would find it rewarding to study the MPEG-2 video specs. It will make using anyone of the tools much more obvious and productive.

[ChachiFace]

Sasi, thanks for your reply. Where can I find the MPEG-2 video specs? (somewhere on this site?) Studying them can only help. Thanks again.

[SaSi]

I would do a google search with "13818-2" as a keyword.

Of course, the official source is www.iso.org

[FOO]

The specs don't explain anything. They just set the rules.
I have seen good explanaitions of the scan order , but can't
remember where. I'll see if I can find it.

[ChachiFace]

This helped me understand a lot!

http://www.digitalfaq.com/capture/interlace.htm

[FOO]

No help on block scan order in there however

[SaSi]

Well, I don't know how relevant this is, but I just got a fine book

H.264 and MPEG-4 video compression.
by Iain E.G. Richardson, John Wiley publishing.
ISBN 0-470-84837-5

This book discusses the MPEG encoding including MPEG-2 and tries to explain the various elements of the MPEG-2 structure. It is excellent reading (300 pages) for anyone wanting to understand the encoding methodology.

I am quoting two paragraphs from pages 57 and 58:
Start Quote:
Scan
After quantisation, the DCT coefficients for a block are reordered to group together nonzero coefficients, enabling efficient representation of the remaining zero-valued quantised coefficients. The optimum reordering path (scan order) depends on the distribution of nonzero DCT coefficients. For a typical frame block with a distribution similar to Figure 3.38, a suitable scan order is a zigzag starting from the DC (top-left) coefficient. Starting with the DC coefficient, each quantised coefficient is copied into a one-dimensional array in the order shown in Figure 3.41. Nonzero coefficients tend to be grouped together at the start of the reordered
array, followed by long sequences of zeros.

The zig-zag scan may not be ideal for a field block because of the skewed coefficient distribution (Figure 3.40) and a modified scan order such as Figure 3.42 may be more effective, in which coefficients on the left-hand side of the block are scanned before those on the righthand side.
End Quote