Copy link to clipboard
Copy link to clipboard
Look at it this way:
Every time a video file is re-encoded to a different codec and/or datarate there will be quality loss. With higher datarate formats like Prores, Dnx, Cineform, AVC-Intra etc the loss is negligible. And with these formats, we are talking much, much bigger datarates than your original 5Mbps files or even the YouTube Preset at 16Mbps.
As you go down in datarates quality loss is more noticable.
The simple answer (in your case) is: the YouTube preset is your better option. You could probably get away with 10-12Mbps but is the file size really that much of a concern?
If it is then 10Mbps would be my next step down. 5Mbps is probably too low, but it's entirely up to you. If you can't see a difference, or it does not worry you, maybe it doesn't matter!
As you get towards the lower datarates you'll start to see more compression artifacts most noticably in text and graphics, so this may affect your intro/outro and lower thirds. Have a close look and see for yourself.
A nice trick to see the difference is to re-import your exported file into Premiere Pro, drop it onto your master sequence (on a higher video track) then in effect controls under 'opacity' 'blend modes' choose 'difference'. The image will go mostly black - but wherever you see anything, this indicates a difference between the original and the encoded file. A high quality exported file will show up as almost black (or totally black).
In comparison to Prores at around 250Mbps - 16Mbps is nothing 🙂 🙂
Another thing to remember - file sharing platforms like YouTube will re-encode your video again (introducing some more quality loss). So it's always better to start with a high quality source.
Wow, thank you for that difference effects trick! I tried it out and the screen was pitch black through the whole recording, even the intro and outro. Only the lower thirds showed some difference, mostly in the text. That's with 8mbps. Does that mean I'm good?
I understand now that a higher export bitrate causes less compression, and going even higher than the source footage's bitrate yields a better result. But there is one thing I don't understand: If I render my video with 16mbps, the file size is ~2.5x as large as the source footage (which has ~5mbps). How is it possible that I get a bigger file even though I'm losing information (video quality)? If I would repeat that process a couple times, I would end up with a huge file that has very low video quality. I would expect the file size to shrink with video quality. Is this just a weird quirk of how video rendering works?
Your original video may be using H.265 to encode. H.265 provides higher quality video at the same or lower datarates than h.264.
To get the same visual quality with h.264 the datarate needs to be higher than with h.265.
You can export using h.265 if you want - test it out and see what you get at 5Mbps.
Note h.265 is not as 'common' as H.264 but is used a lot for recording - Drones, gopro etc often use this format to keep datasizes to a minimum.
If 8Mbps looks OK, go with it - IF smaller files sizes are something you need. Otherwise I'd choose between 10 and 16Mbps for max quality.
Why do you still lose quality with higher datarates?
going to h.264 might in your case be part of the reason but everytime a video is re-encoded to a different format or datarate you get what's called 'concantenation'. At lower datarates and using highly compressed format (like h.265 and even h.265) a lot of image data is 'thrown away'. Each time it's re-encoded there are slight differences in the encode and the encoding loses will get worse each time. It's not a great analogy but think of photocopying a photocopy.
You've noticed the differences in the text of your lower thirds. Text is often the first thing to 'lose' quality in highly compressed files. Other things that will show compression artifacts easily are very highly detailed scenes - trees, water etc.
Thank you again! I understand that rendering a video multiple times lowers its quality. That makes sense. But I would expect a smaller file size in return, not a 2.5x increase. In every other kind of file/data, less information usually results in a smaller file.
The 2.5x increase may because your chosen export format is h.264 but you original recording codec was h.265.
Find out - open the original recording in VLC (free app). Find the 'media information' window (found under the 'windows' menu in VLC). Click on 'codec details' and see what codec is listed.
No, both files have the h.264 codec. The file increase comes from the larger bitrate.