Innovative solutions for post production professionals

Anamorphic Widescreen – Understanding Pixel Aspect Ratios

The old 4:3 video standard is quickly becoming a thing of the past; all new LCD displays are 16:9 widescreen models, all HD camcorders record a 16:9 widescreen image, and even most older DV camcorders offer the option to record in 16:9 mode. After editing widescreen content with your favorite NLE software, you may want to export the video to a format that viewers can watch on a website or portable device, and that can lead to unforeseen issues due to variations in how the image is recorded.

For instance, when shooting NTSC DV video, the dimensions of the video are 720×480, and this is true for BOTH the 4:3 and 16:9 recording modes. How can both formats record at the same dimensions? Easy – the Pixel Aspect Ratio, or PAR, is changed to create the illusion of widescreen video, recording with a PAR of 1.2 rather than the 0.9 PAR of 4:3 DV. This results in rectangular (non-square) pixels, creating the widescreen look upon playback, and is referred to as anamorphic widescreen since some stretching of the image is taking place to achieve the widescreen look.

When a computer displays a still image, it is assumed that the pixels in the image are square, with a PAR of 1.0, and the same holds true for video clips in most software players like Apple QuickTime Player and Windows Media Player. To display anamorphic video correctly, the playback software or device needs to know the PAR of the video and must be able to adjust the video output to compensate for it, but many can only display square pixels. When viewing an anamorphic video clip with these players, the image may look horizontally squished, appearing more like 4:3 video than 16:9.

In non-linear editing programs, the PAR is determined for you when you choose the editing preset, based on the video format you will be working with. When you import a 16:9 DV clip into a 16:9 DV editing project, the video will look correct in the preview display, but if you import that same 16:9 clip into a 4:3 project, the image may appear horizontally squished, making people look taller than they are. In Adobe Premiere for example, you can use the Interpret Footage command to manually change the PAR of the clip to 1.2 widescreen so it displays correctly in the project, should you need to mix footage with different aspect ratios.

Interpreting footage in Premiere Pro CS5

Interpreting footage in Premiere Pro CS5

Let’s look at a few different video formats now to determine if they are native widescreen, or anamorphic. A true widescreen format will deliver a 16:9 aspect ratio using square pixels. We’ve just discussed DV, so let’s consider 720p HD. Its dimensions are 1280×720, and you’ll find that dividing each number by 80 results in 16×9, so 720p is natively 16:9 video. The full 1080i and 1080p video formats at 1920×1080 are also proven to be native 16:9 video when divided by 120.

The HDV 1080i/p formats record at 1440×1080 resolution with a 1.333 PAR, so these would be considered anamorphic HD formats. Note that 1440 times 1.333 equals 1920, so when this video is played back from an appropriate device like an HDV camera or a Blu-ray player, the device will respect the 1.333 PAR and display the video at 1920×1080, while Windows Media Player would display the image at the native 1440×1080 size using square pixels, losing the widescreen appearance. Another anamorphic HD format is DVCPRO HD 1080i at 1280×1080, which uses a 1.5 PAR to achieve a 1920×1080 playback display.

Some HD camcorders that record full 1920×1080 video may also offer the economy option to record 1080i video at 1440×1080, allowing longer record times on the memory card. The Sony EX1 camera shooting XDCAM EX (SP mode) and the Panasonic AG-HMC40 shooting AVCCAM (HE mode) are just two examples I’m aware of, so pay attention to the recording mode you shoot with so that you choose the correct project presets and workflow in your editor.

In case you’re wondering why HD camera manufacturers would want to market anamorphic HD formats with reduced resolution in the first place, it’s because at 1920×1080, there was simply too much data to record to tape, even after compression. By reducing the resolution of the files, it allowed existing tape formats to be modified to record HD video, as with the HDV format using standard DV cassettes and tape transports in the cameras, keeping the cost down and also insuring backwards compatibility with DV video. The latest generation of HD camcorders use highly efficient H.264 encoding and record direct to memory cards or hard drives, so recording the full 1920×1080 video is no longer the issue that it was with tape, thus even consumer-grade cameras now commonly offer full HD recording.

I mentioned still images earlier, and any still from a scanner or digital camera should always use square pixels. Where you could run into trouble is if you or a contractor are creating a still graphic in Photoshop or another paint program for use in a video project that uses anamorphic video, you have to keep the PAR in mind while designing the graphic screen.


Default 1.0 PAR in Photoshop CS5 vs. Proper HDV 1.33 PAR in Photoshop CS5

Optionally, you can design the graphic at the native size of the video clip, but you then must tell the graphics program the PAR that you wish to work in! In Photoshop, there is a menu where you can select the PAR from a list of common video sizes, and PS will then stretch the image to look correct while you are working on it, otherwise what you draw as a circle might end up looking like an oval on the video.Generally, I would create the graphic using square pixels, which would be the default in the paint program. So if creating a graphic for use in a 1440×1080 HDV video project, you could create the graphic at the display size of 1920×1080 with 1.0 PAR, and when you import the graphic into the NLE, it should interpret it correctly.

Setting the PAR in Photoshop CS5

Setting the PAR in Photoshop CS5

Now that you have a better understanding of anamorphic video and PAR, we can move on to the point of this article, which would be getting the proper aspect ratio when exporting your widescreen videos for public consumption.

For widescreen export to DVD, from DV widescreen or HD sources, just be sure to set the DVD project as 16:9. With some DVD players, you may have to go into the menu and tell it whether you have a 4:3 or 16:9 display connected, but most handle this automatically and will add letterboxing on 4:3 displays or display full-screen on 16:9 displays when playing back your widescreen DVD.

For Blu-ray 1080i/p, video may be encoded as 1920×1080 (1.0) or 1440×1080 (1.333) and the player will output 1920×1080 to the display in either case, and 720p should of course be encoded at 1280×720.

The real issue comes up when encoding clips for computer use, where the DVD or Blu-ray player isn’t handling aspect ratio conversions for us. Computer and web video players use square pixels and that is how your video should be encoded, so if your source is anamorphic, you’ll need to adjust the video dimensions to get the correct aspect ratio using square pixels.

For 16:9 DV video at 720×480, set the output size to 864×480, 1.0 PAR, to get a proper output display. Note that 720 times 1.2 equals 864, and dividing 864×480 by 54 gives us 16×9, so we know the display will look correct to the viewer. Of course, you will likely want a smaller size for web streaming, iPod, or other uses, so just knock the size down while maintaining the 16:9 dimensions, for instance 640×360 or 480×270.

For HD web clips, 1920×1080 requires too much bandwidth over the internet, so the standard practice for sites like YouTube or Vimeo is to use 720p HD video encoded at 1280×720, whether starting with a 1920×1080 or 1440×1080 source.

If starting with NTSC or 1080i interlaced video, be sure to enable deinterlacing in the encoder software, since all computer displays are progressive. If you don’t deinterlace your footage, it will create nasty motion artifacts on computer screens due to the fields being used in the source video.

I should also mention that shooting in HD provides an opportunity for videographers to produce actual photo prints from the video due to the high quality of the source, but again, you may need to adjust for the PAR. If you grab a still image from 1080i HDV in your NLE, then open that image in Photoshop, it will not look widescreen, but more like 4:3. It would print out the same way, whether on your inkjet at home or the photo kiosk at the local drug store, so you must adjust the still image to square pixels.

For interlaced source video (1080i), the first step is to apply the Video > Deinterlace filter in Photoshop prior to doing any resizing to eliminate interlace artifacting in the image caused by motion in the video. Next, resize the image from 1440×1080 to 1920×1080, making sure to uncheck the Constrain Proportions box first. The resized image will then look correct on screen, and in print.

Whether your frame grab needed resizing or not, you’ll probably want to use the Crop tool in PS to size the image to fit a photo print properly, for instance 6×4 or 7×5. Save the results to .jpg format and either upload to a print service like Walgreen’s or Wal-Mart, or take the files into the store on a thumb drive or CD. Clean HD video grabs will look very nice as a glossy photo print, almost undistinguishable from a digital photograph.

To recap, if you are shooting 1920×1080 or 1280×720 HD formats, then you already have square pixels and are good to go, but for DV widescreen, HDV, DVCPRO HD and other anamorphic video sources, mind those pixel aspect ratios throughout the workflow and the whole process will be a better experience.

Share with your networks.
Anamorphic Widescreen – Understanding Pixel Aspect Ratios