A while ago I rendered out a video at 1080p and then again at 720p.
As expected the 720p file size was considerably smaller than the 1080p one. However I thought I’d play a bit more with this as I was a little ‘uncertain’ with the results I was getting, after all not just image size (1080 / 720) had changed, but bit rate too. On play back both videos looked ‘good quality’ and even stretching the 720 up to fill a 1080 size window didn’t noticeably reduce its viewing quality.
So I took a short 20 second clip standing on a street corner which started with the camera held still (so a mix of non-moving buildings but with cars and people going past) and then at the 10 second point panned the camera around so everything was now moving relative to the camera.
Then it was a case of rendering this out at 1080p various times, the only thing being changed between each session being the bit rate, working up from around 800 Kbps to 32Mbps.
Down at the low bit rates stationary items looked all right, but moving items (people, cars, the buildings when the camera was panning around) were of very poor quality. At the high bit rate end of the scale everything looked crisp and nice, however both rendering time and file size had also grown.
What was also apparent was that the choice of playback program was significant. Different players, VLC, Windows Media Player, QuickTime, all will play back the same clip differently. Same applies to using different browsers as a play-back engine – they are not all equal. Also what you have recorded makes a difference. A clip with little movement and few colours will be very different from a clip with lots of movement and activity.
Then a matter of working up from the bottom to find the bit rate cut-off point between where a clip was unacceptable and one was watchable; and coming down from the high bit rates and at what stage starting to notice a degradation in quality. Then finding an overall mid-point balance between quality and file size .
So what this really means is that there is no one ideal bit rate setting that’s going to cover all situations. For my set-up, with my camera and my editing software and thinking of uploading to somewhere like Vimeo or YouTube, then I’ll be thinking of working at 8 Mbps for indoor videos where there’s not a lot going on, and 12Mbps for outdoor videos full of colour and movement.
But this just my experience. You will need to do your own experiments to find your best compromise settings that will suite your equipment and set-up.
A while ago I thought I’d take my camera out for an afternoon’s videoing and so headed off to Linlithgow Palace. Once I’d got some editing done I found I had a video of just under 10 minutes in length. Usually I render out at 720p for a YouTube or Vimeo upload, but this time I thought I’d render at 1080p just to see how things went. At the end of the processing I found myself with a file of just over 2 gigabyte. This I thought a little bit large for uploading from home so went up to Uni did my upload (to Vimeo) from there. That went nice and fast, in fact it uploaded faster than the length of time it took me to type out title, description and all that sort of stuff. Once back home thought I would re-render it but this time in my more usual 720p (other settings the same) just to do a compare and contrast. My just under 10 minute came down at just over 400 megabyte.
So we’re talking about one fifth the size of the 1080p video. Playing the two videos side by side it was hard to notice any difference in quality, it was there but you really had to pick around the edges to see it. Now and again I’d pause at a suitable moment where perhaps a a sign was in view or maybe a car number plate or something similar and try and read it, and it was more readable in the 1080p version. However as a flowing video it was quite difficult to tell the difference between the two, yet the 720 giving me so much more free hard drive space compared to 1080.
I also rendered it out in standard definition quality at 360p size (looking towards uploads for mobile devices or older smart phones). Here we had a file size of just under 60 megabyte, or a little under 3% of the 1080p file size. Comparing the video here to the others there was a very definite reduction in quality. It still wasn’t too bad to watch, however when you had this 360 next to either the 720 or 1080 the difference really jumped out at you. On the other hand it was still more than good enough for a portable device if you were on the train.
Something to think about. Take the 1080 HD video size, which is a 1920 by 1080 rectangle. Multiply that out and it comes to just over 2 million pixels. Or to put it another way, if you’re in a shop talking about video and cameras and the sales staff are busy trying to sell you this camera rather than that camera because this camera has more mega pixels on the sensor… well, if you’re going to actually film at the native HD resolution then all you really need is this just over 2 mega-pixel size sensor. Bigger does not always mean better.