HDV has a 16:9 aspect ratio. But many cameras have 1080 by 1440 senors, e.g. Canon's XH-A1. I've read that they essentially have rectangular pixels instead of square ones, so the horizontal pixels cover the 1920 field but only with 1440 descrete measurements. I'm okay with that.

But when I capture the video in my NLE (in my case an Avid Xpress Pro 1080i 59.94 HDV project), does the data stay as rectangular pixels or does the NLE convert it into its square equivalent? And does this behavior vary by NLE or is it consistent across the board?

And finally, do cameras with native 1920 sensors convert their HDV signals to 1440 as well or do their HDV recordings have 1920 by 1080 square pixels?

I apologize if these questions have been answered elsewhere. I'm just feeling a little confused right now, even by Wikipedia, LOL!

Thanks as always for the help!

P.S. This post was motivated by reading about Avid's DNxHD 145-TR codec designed for some HD/HDV applications. "TR" stands for "Thin Raster," i.e. 1440 instead of 1920. I had always thought ALL varieties of HD and HDV were 16:9 out of the camera. But Avid has me actually doubting this. (I'm sure I'm the person mistaken somewhere along the line.) Thanks again.

P.P.S. I believe this is more of an Avid ? than an HDV one, so I've posted on Avid's forum.

P.P.P.S Acutally the Avid community has me believing there is more to the HDV codec than I currently understand in terms of horizontal pixel counts and aspect ratio. Sorry for all the addendums, it's getting late around here.

The pixel aspect ratio for HDV and DVCPro HD is 1.33333333:1, i.e it's a rectangular pixel in the codec.

Your computer displays square pixels, so it loads up the original image, then stretches it when you play back in the correct pixel aspect ratio (as long as your settings are correctly set up.)

It's just like Anamorphic SD images, which are squished into a 4:3 container, but then stretched out by wide screen tv's (or squished down vertically if going to be letterboxed on a 4:3 DVD player for instance).

Times 1440 by 1.3333333 and you get 1920.

So, in native format your pixels would play back square on your computer, but the should be 'stretched' on playback/preview by your NLE on the fly so you would only see this if you loaded them up in a player that doesn't check this, or if you export a single still image (in which case the still image will be squished and need to be stretched out).

When you output your footage, you have to make sure your export settings are correct for the final output. Any titles etc you create in Photoshop should preferably be created with the correct pixel aspect ratio settings for your project as well to match the original footage (the last few versions of Photoshop can all adjust the pixel aspect ratio to account for HD pixel ratios).

Different or incorrect interpretations of pixel aspect ratios can cause some major headaches (for example, Color for Final Cut Studio had(/has?) a known issue where it didn't interpret the pixel aspect ratio conversion of DVCProHD correctly in the xml send through, and you got some fairly funky things happen in terms of the image you ended up going back into Final Cut and had to manually change clip/timeline settings to new settings to get the aspect ratio correct.

DvcproHD has a slightly different PAR due to only having 1280 pixels horizontally for 1080p output, but the way that's handled in the computer works the same way.