Why anamorphic?

[Russia]

Hello everyone. I couldn't find the answer for the very simple question... What is the reason for doing anamorphic encodings? The filesize is bigger, the dimensions of the movie is bigger...quality is better also :) (taking in to the consideration higher bitrate)...So why to do anamorhic encodes? As far as i know only NTSC movies can be anamorphic, wright? And how would i know whether movie needs to be encoded in anamorphic? Thanx in advance.:thanks:

[Sharc]

Very simply speaking:
- The picture appears sharper the more details are stored (Resolution, width x hight in pixels).
- The human perception is more sensitive to details in the vertical than in the horizontal dimension. Hence it makes sense to reduce the resolution in the horizontal (sqeeze the picture) and to increase the resolution in the vertical dimension = anamorphic storage ("eggheads").

For more technical information about conversion of analog TV signals to digital storage media and movie aspect ratios versus DVD standard resolutions you may want to google the web and search wikipedia for "anamorphic".

Anamorphic encoding is used for PAL and for NTSC.
The decision to encode anamorphic or 1:1 is up to you. However one has to make sure that the player will stretch an anamorphic encoded picture correctly. Non-anamorphic encodes are perhaps less critical in this respect.

[LoRd_MuldeR]

"Anamorphic" means that the native size of the video (number of pixels) differs from the display aspect ratio.
It is not used only for NTSC, but also for PAL. For example the PAL resolution is 720x576, which obviously is neither 4:3 nor 16:9.
So in addition to the resolution you need another information, the DAR (Display Aspect Ratio) or PAR (Pixel Aspect Ratio) to display it correctly.
DAR and PAR do the same job. They are not identical, but you can calculate the DAR from a given PAR and vice versa.
The player will read the PAR (or DAR) and stretch the video accordingly, so it appears with the correct aspect ratio on your screen.

I think if your Source Video is anamorphic (PAR unequal 1:1), it should be encoded as anamorphic again.
If you have an anamorphic source and you want to encode it non-anamorphic (PAR = 1:1 aka "square pixels"), you'll necessarily have to resize the video.
For example a 16:9 PAL video comes in as 720x576. To encode that at a PAR of 1:1, you'd have to upscale it to 1024x576 or downscale it to 720x400.
Unfortunately resize unavoidably kills some detail. So keeping it anamorphic usually is the best way, as it avoids resizing.
But of course you'll need to set PAR/DAR correctly again!

If your source is NOT anamorphic, there usually is NO reason to encode it anamorphic, as this again would require resizing.
Of course if you want to put a none-anamorphic source on a Video-DVD, you'll have to encode it either at PAL or NTSC resolution.
In that case you can't avoid making your "square pixel" source anamorphic again...

[Russia]

"Anamorphic" means that the native size of the video (number of pixels) differs from the display aspect ratio.
It is not used only for NTSC, but also for PAL. For example the PAL resolution is 720x576, which obviously is neither 4:3 nor 16:9.
So in addition to the resolution you need another information, the DAR (Display Aspect Ratio) or PAR (Pixel Aspect Ratio) to display it correctly.
DAR and PAR do the same job. They are not identical, but you can calculate the DAR from a given PAR and vice versa.
The player will read the PAR (or DAR) and stretch the video accordingly, so it appears with the correct aspect ratio on your screen.

I think if your Source Video is anamorphic (PAR unequal 1:1), it should be encoded as anamorphic again.
If you have an anamorphic source and you want to encode it non-anamorphic (PAR = 1:1 aka "square pixels"), you'll necessarily have to resize the video.
For example a 16:9 PAL video comes in as 720x576. To encode that at a PAR of 1:1, you'd have to upscale it to 1024x576 or downscale it to 720x400.
Unfortunately resize unavoidably kills some detail. So keeping it anamorphic usually is the best way, as it avoids resizing.
But of course you'll need to set PAR/DAR correctly again!

If your source is NOT anamorphic, there usually is NO reason to encode it anamorphic, as this again would require resizing.
Of course if you want to put a none-anamorphic source on a Video-DVD, you'll have to encode it either at PAL or NTSC resolution.
In that case you can't avoid making your "square pixel" source anamorphic again...

Thanx for your reply. I think i started getting in to it... Correct me if i'm wrong. I'm doing x264 CRF18 encodes, so i'm trying not to resize and achive best quality possible. So what you saying is that whenever my resolution in Megui changes from the original i need to do anamorphic, wright? Lets say i have ITU 16:9 NTSC (1.822784) 720x480 movie. After cropping and without anamorphic option selected in Megui my encode will be 720x288 resolutionwise, with anamorphic option selected it'll be 720x480. In this case it's propper to go anamorhic, wright?:thanks:

[stax76]

You have to remember CRF 18 produces extremely large files, should be 50% more then what is common, CRF 22. You might want to consider not to recompress at all.

[Russia]

You have to remember CRF 18 produces extremely large files, should be 50% more then what is common, CRF 22. You might want to consider not to recompress at all.

I'm not going to recompress anything, dont get me wrong. I just trying to clearify the question i have regarding anamorphic encodes for myself. Still didnt get the whole idea...:stupid:

[Blue_MiSfit]

The idea is that you don't touch your video with a resizer. Preserving as much vertical resolution is important, so anamorphic encoding lets you preserve every line, so you can expand horizontally (852,480 for true 16x9 anamorphic). I'd rather interpolate on playback than throw away information when encoding.

~MiSfit

[LoRd_MuldeR]

I think you can keep the full vertical resolution of an anamorphic source by only upscaling, e.g. resize an 16:9 PAL video from 720x576 to 1024x576.
But then you have to encode more pixels per frame, which potentially leads to a bigger file (using CRF) or to a worse quality (using 2-Pass with fixed target bitrate).

Keeping it anamorphic will preserve the full resolution without upscaling before encoding, so it's the most efficient way to encode such source...

[Russia]

:thanks:everyone who answered. I think i got it now.

[yaous]

1) 720x480 anamophic dvd --> anamorphic transcoding --> displaying 852x480 in media player

2) 720x480 anamophic dvd --> resize to 852x480 when transcode --> displaying 852x480 in media player

In case 1),How media player display anamophic encoded video? Isn't it resize?

If so, is there no difference between 1) and 2) when we see them?
(given some more bitrate to resized video in proportion to resolution)

[Sharktooth]

the difference is you need more bitrate to retain the same quality for 2)

[Octo-puss]

I am still not sure if I understand this thing correctly. It's been bugging me for months and I still couldn't get it, lol.
Firstly though, how do I tell whether the source is anamorphic or not? 1:1 PAR, okay, but how do I tell? Maybe I am just confused by the terminology being used...

Mulder wrote:
It is not used only for NTSC, but also for PAL. For example the PAL resolution is 720x576, which obviously is neither 4:3 nor 16:9.
a) I got this one video that's PAL 720x576 4:3. I'm lost. Encoded it just as it is (no resizing, CAE, or specifying the AR at all) and it plays (looks) fine. EDIT: No way! it's slightly stretched vertically (just smaller image, hm) when CAE was used. Passing --sar 16:15 to x264 produced image identical to the source. I am even more confused.
b) The above statement would indicate that everything that has 720x576 resolution while being 16:9/4:3 (that's the DAR, right? the AR at which it should play at) is anamorphic. Yes? That would make sense to me if there wasn't a) - it didn't need anything, it looks ok already. But seeing what Mulder wrote.. eh.

Also. How is adding --sar x:y different from using CAE in the script creator?
I am not fond of ANY resizing/cropping at all under any circumstances.

[stax76]

Aspect ratios are discussed to death here and there can be found numerous guides on the topic, look for guides and posts by Brother John or SeeMoreDigital for instance.

[Octo-puss]

Saw crapload of guides but still confused. I think I could describe it as "I understand the words but don't get the story fully".

[Octo-puss]

Ok, nevermind everyone, I think I got it. CAE is apparently only good when you resize the image somehow. When wanting to keep the original resolution, --sar is your friend.

[stax76]

You really don't understand aspect ratios and CAE so you have to continue to read guides and threads. This thread is about aspect ratios in general however and not about how to use MeGUI so you have to read/create threads containing MeGUI in it's title.

[Octo-puss]

Thanks so much for valuable tips. I'll wait for someone more helpful.
Meanwhile, I will use what WORKS.

[Adub]

The filesize is bigger

Just in case this hasn't been said already, this is false. When anamorphically encoding something, you are actually saving tons of space, instead of encoding a properly resized version.

[smok3]

in film:
at some point they figured out that wider is 'better', but since the film stock was not so wide, one would have to either:
a. use a smaller portion of film for picture and waste the other parts of the film with black.
or
b. squeeze the picture when shooting and unsqueeze on playback time
in video:
very similar, only that it happen later.