# miniDV - Same resolution but lost vertical space in 16:9 widescreen?

I have an old SD PAL camcorder, which records on miniDV tapes at a resolution of 720x576 pixels.

This is the output from ffmpeg of the 4:3 recording:

Stream #0:0: Video: dvvideo, yuv420p, 720x576 [SAR 16:15 DAR 4:3], 28800 kb/s, 25 fps, 25 tbr, 25 tbn, 25 tbc


This is the output instead for the 16:9 widescreen footage:

Stream #0:0: Video: dvvideo, yuv420p, 720x576 [SAR 64:45 DAR 16:9], 28800 kb/s, 25 fps, 25 tbr, 25 tbn, 25 tbc


I noticed that the resolution is the same in both the cases, and of course the SAR is different, in one to fit the 4:3 resolution, in the second to fit 16:9. This leads me to think that the footage should be stretched to fit the aspect ratio but keep the same resolution in both cases.

However, the actual video footage is quite different. It doesn't look like they use the same resolution, but more that the 16:9 footage is cropped vertically. This would make sense logically, since the sensor is 4:3. However, shouldn't that mean that the resolution is lower? How is it possible that the footage has the same resolution while recording but a different one while reproducing?

EDIT: To clarify what I mean above, I took a screenshot of both video. As expected, both videos fit the two formats: 768x576 for 4:3 (16:15 stretch), and 1024x576 for 16:9 (64:45 stretch). Which again, is what is expected.

However, the image is evidently cropped vertically in the second case. For example, if I put a ball on the top and the bottom of the screen, it would be visible in 4:3, but would be cut out in 16:9. Video quality is also much lower in this case, as expected from an upscale. It looks like the image is first cropped at something like 720x405, and then upscaled to 1024x576.

I wonder why this happens, instead of having just the image at 720x576 stretched out to 16:9.

Okay, after some reasoning and research I think I understand what happens.

Basically, the sensor is fixed in size, but the lens is too. Which means the only way to get a wider sample to fit the image to 16:9 is to crop down the image on the sensor to about 720x405 pixels, each pixel catching a sample of ratio 64:45 (SAR).

The image is then stretched vertically to fit the whole 576 lines of the PAL standard. This means the original 5:4 footage captures has the same horizontal FOV as the footage captured in 4:3 mode, but a lower one vertically, stretched up to fit the whole 576 lines. This allows for the recorded media to then be stretched back into 16:9 by the media player.

This last step (the vertical stretching done by the camcorder) is still something that sounds really puzzling to me, but after analyzing some more footage, it is rather obvious that it is a digital upscaling and not an analog capture of the samples, since a lot of the digital noise is stretched vertically too.

It is an interesting way to get around the hardware limitations of the camcorder, but in the end not really worth it, considering the loss in quality.

Believe it or not, this is correct. The sensor in your camera is something like 720x576 regardless of whether you're recording in 4:3 or 16:9. What that means is that the lens needs to squish the 16:9 image into a 5:4 (720:576) area and your display device needs to stretch it back out. This is why the SAR (sample aspect ratio, AKA pixel aspect ratio) is not the same in both cases. (And also, DV even at 4:3 does not have square pixels which is why a 720x576 image is displayed 4:3!)