Posted: Tue 31 Oct 2006, 3:09 Post subject: Source file aspect ratios in TMPGEnc XPress 3.X
Hi,
I'm relatively new to video encoding and DVD authoring. While the information on this web site is very helpful, I have a question. TSUNAMI MPEG Video Encoder XPress (which is the same as TMPGEnc XPress 3) provides the following options for aspect ratio for source files:
In which cases each of these ratios should be used?
I want to encode PAL DivX .avi file, 640 x 360, into PAL MPEG-2. The Encoder XPress defaults aspect ratio to Pixel 1:1 (square pixel) for this input file. What would be the right option?
Joined: 04 Feb 2003 Posts: 587 Location: Lisboa, Portugal
Posted: Tue 31 Oct 2006, 3:22 Post subject:
First of all, 640x360 is certainly not PAL. The file might play at 25 frames per second (if indeed it does), but PAL would be 720x576 (or, at worst, 704x576).
Your AVI file may have been made from a PAL source, but currently it's (probably) a non-interlaced, square-pixel AVI, with a lower resolution. The 640x360 resolution suggests a 16:9 image with square pixels (16:9 PAL is also 720x576, with wide pixels).
In other words, the source pixel aspect ratio would be 1:1.
Now, converting a DivX (MPEG-4) file to MPEG-2 would be a bad idea even if the file was the correct resolution, since it would introduce new compression artifacts. Taking the resizing into account, and the fact that your AVI file is probably non-interlaced, you're not only going to get compression artifacts, but also a bit of blurring, caused by the resampling, and possible interlacing problems.
In other words: you're better off getting a DVD player that can read DivX files (and will usually do some tricks to try to display them correctly on the TV screen). Or just get a graphics card with TV-out and connect your PC to your TV. But if you want decent quality, get a real PAL original, not a resized, deinterlaced DivX file.
Thanks for your explanation. The file is 25 fps and Encoder XPress does defaults to progressive that proves you right. Is there a way to find out if AVI file interlaced or not and which pixels, square or wide, its image has? Does it also mean that all DivX AVI files are non-interlaved and have square pixels?
Joined: 04 Feb 2003 Posts: 587 Location: Lisboa, Portugal
Posted: Tue 31 Oct 2006, 21:55 Post subject:
Interlaced AVIs will show two distinct fields when you pause them on a PC screen (i.e., the odd and even lines will show different moments, clearly noticeable if you pause during fast subject or camera motion). See this image for a (zoomed-in) example.
AVI files (compressed in DivX or any other codec) can have pretty much any resolution. Full-resolution files will be 720x576 (PAL) or 720x480 (NTSC). Anything below that indicates the file was resized and (probably) deinterlaced (assuming it was interlaced to begin with, of course).
Thanks again for very good explanation. Returning to aspect ratio options in TMPGEnc Xpress 3.x that I listed in previous question, I have an idea now what "Pixel 1:1 (square pixel)" aspect ratio means. From your posts on this forum I know you did not have much experience with Xpress but if you have to make an educated guess, could you explain what source files each of the rest of those options could be used for? Here is the list again:
Joined: 04 Feb 2003 Posts: 587 Location: Lisboa, Portugal
Posted: Wed 1 Nov 2006, 17:48 Post subject:
Well, they mean exactly what they mean.
Pixel aspect ratio is the ratio between the width and height of the pixels. Image aspect ratio is the ratio between the width and height of the image. Where the width and height are measured in physical units (i.e., millimetres or inches or whatever, not number of pixels or bits or anything like that).
Taking one of the aspect ratio values, plus the image resolution (in pixels), you can calculate the other. Also, taking both aspect ratio values and one of the image dimensions, you can calculate the other.
If the vertical resolution is 576 pixels (PAL standard), then horizontal resolution will have to be 576 x 1.7777 / 1.4545, which is 704. 704 pixels per line is the area actually considered as "visible" by the CCIR standard (the full width is 720, but the edges are not counted, due to some issues related to analog signals). I wrote a bit about that here (fourth question from the top).
DVD supports both 704 and 720 pixels per line, but it's really up to the player to decide which of those two formats it considers as 4:3. In some players 704xN will be exactly 4:3, and 720xN will be 4.1:3, and in others 720xN will be exactly 4:3 and 704xN will be 3.9:3 (the actual TV set can also skew this slightly). The difference is very small, so it's not really a problem, and it's always best to keep the original resolution and avoid resizing. Not that "N" stands for 576 in PAL and 480 in NTSC.
Generally, if your input and output formats are the same, you should simply pick the same aspect ratio option for input and output (ex., set both input and output to "Image 4:3"). If you're converting from a square-pixel file to a PAL file, for example, set the input to "pixel 1:1" and the output to the image aspect ratio (ex., 16:9).
RMN
~~~
Last edited by RMN on Thu 2 Nov 2006, 16:15; edited 1 time in total
Understood for the first four source aspect ratios in Xpress as pixel aspect ratio is mentioned explicitly. Then what is pixel aspect ratio is supposed to be for, say, "Image 4:3" or "Image 16:9"? If it is 1:1 (square), then how does it differ from "Pixel 1:1 (square pixel)"?
Joined: 04 Feb 2003 Posts: 587 Location: Lisboa, Portugal
Posted: Thu 2 Nov 2006, 16:09 Post subject:
As I said, to calculate the pixel aspect ratio you need to have the image aspect ratio and the image's dimensions in pixels (or vice-versa).
So the pixel aspect ratio for "Image: 4:3" will depend on the image dimension (in pixels). If the image's resolution is 800x600 pixels, for example, then the pixel aspect ratio is square (1.0, or 1:1), because 800:600 = 4:3. If the resolution is 400x600, for example (and the image aspect ratio is still 4:3), then the pixel aspect ratio is 2.0 or 2:1 (width of the pixels is twice their height), and so on.
There are four variables:
Image width (in pixels)
Image height (in pixels)
Image aspect ratio ("physical")
Pixel aspect ratio ("physical")
Where Iar = Iw / Ih x Par.
If you know any three of those, you can calculate the other (just invert the formula).
In TMPGEnc you always set the image width and height explicitly, so it will only let you pick one of the aspect ratios (image or pixel), and will calculate the other one from those values. Some programs let you set one image dimension (ex., height) and both aspect ratios, and will then calculate the other dimension (ex., width).
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum