Video compression - variation of size & quality - illogical outcomes in quality.

I just downloaded a number of various copies of different source material for a popular comedy central cartoon. sizes range from 87mb to 1GB and source material is HDTV, DVD and bluray. All use the same codec h264 and audio is very close to the same bitrate, not really an big issue in size comparison.

The thing is that the best quality is the DVD rip at about 120mb. I've found that a lot of video's I've found have looked better at 700-800MB vs some files that are 2-3.5GB in size (and these usually use the same codec).

I've done compression before and never had any results where the quality looks so bad at the larger file size that I see in some of these files (like the 2 - 3.5gb). Is there any reason that some of these files often look so bad? I've heard of some people hiding files within pictures and videos and was wondering if there is any way to determine if this has happened in any of these files.

Some people just don't know how to make the most efficient use of compression technology, and that's about as simple as it gets. I've seen very large encodes of movies (meaning file size) in excess of 10GB and then I've found smaller encodes (under 2GB) of the same movie and it looks pretty much exactly the same to my (now almost 50 year old) eyes, but I gave up caring about having the absolute best most awesome quality that's possible years ago anyway. As long as I can watch a movie or show, that's good enough for me and it's easier on the storage as well and lends itself to working well on my smartphone and tablet which tend to have limited storage capacity.

File size != the quality, at least not always. Yes it's true that a larger file size (and I mean the byte size, not resolution of the encode) traditionally means it's using "more bits" which tends to mean more data is there for display and not lost since it's lossy compression by default, but there's so much more going into making a great video encode that it's practically a black or dark art for most people that just don't realize such things.

I adopted a simple rule for myself a few years ago: if it looks good enough and I can watch it, so be it. I don't own an HDTV, probably never will to be honest, I just don't see a point in such things but that's just me personally. I tend to watch most material on my laptop, my smartphone, or my tablet and that's that.

I can say that with respect to "cartoons" or anime style material, the people encoding those from source material (whatever it is) love to placebo the hell out of it for whatever reason and by that I mean they will choose encoding settings that more often than not ends up with a finished encode that is oft-times larger than the source material was, sometimes 2-5x larger and yet they seem to think "Oh, my encode of <whatever> is even better than the Blu-ray!" or something equally ludicrous. If the source material is lossy by nature (as all DVD and Blu-ray content is), then it's technologically and factually impossible to improve on the quality by doing yet another round of lossy encoding no matter what settings are used. There's a reason x264 (the most popular h.264 encoder) has a preset setting known as "placebo" because it's just wasteful and stupid but people don't seem to realize this - even worse, they just don't care either.

If it looks good to you and you can watch it, that's what counts. Right?

h264 encoding has a lot of options beyond bitrate. There is a lot of tweaking you can do to squeeze out more image quality per bit.
http://www.adobe.com/devnet/adobe-media-server/articles/h264_encoding.html
This article just scratches the surface of it.

Part of the problem with "Gotstahas teh big filez!" encodes is you start getting limited by your storage medium's I/O.

So, you get a 1-2GB file that looks great and plays fine.

Then you get a 10-15GB file that looks like crap and/or has playback artifacts.

And all the while the larger file is choking simply because it's got a stupid, arbitrarily padded bitrate.

Thanks for the responses! I feel much the same as Tiberian. While reading the responses I started to think that it might be possible that the files may be not compressed as much so that they don't need as much CPU power to play (just more bus bandwidth...). I know some of my old laptops start to choke on some higher bitrate HD content so if the file was larger and needed less processing, the older computer could handle it better. Is that a possibility? I honestly think the actual reason isn't that well thought out and probably more along the "placebo" reasoning that was posted.

Sometimes yes and sometimes no. Using more compression may or may not mean you need more to decompress. The link that sharknice posted is a pretty good link to get into the details. Here is another that might be useful.

https://mattgadient.com/2013/06/12/a-best-settings-guide-for-handbrake-0-9-9/

Basically, what it comes down to is that the DVD one was probably ran through using a high profile which uses the most of the different techniques to keep the IQ up, and file size down. That means that they spent a significantly longer time encoding the file than some others that might be using a mainline or baseline file. The biggest files were probably used with a very fast setting that was only using baseline and thus much quicker to encode. (At expense of quality and file size)

For the high profile setting some of the options might increase cpu usage, where others might not require any additional cpu for playback. The difference is simply the algorithm used to create the file, and that might take longer to process on the encode.

Some also steal videos and recompress at a higher bit rate and claim it their own. You end up with crappier video and larger size.

basically count the colors in a frame the more different colors and the more they change to the next frame the larger the file. lazco can over come some of this with striping where the colors are added up across the frame and line of color parity is added at the right side and the bottom of the frame but if there are too many artifacts it has to just save more data. each color can be 0-255 or double word and use a hex code number. if a single value it is grey value and it has to store a inverse color across the whole image.
Encoding h264 without sound is smaller but a single frame or a million frame of a single color is about thirty bites, a movies with rapid motion can be gigs in size.