Thread: Why do people rar videos?

Oh, it's MPEG4 alright. DivX to be exact. It's also MPEG4 and an AVI container.
Let's see how it fares against MPEG4 AVC, shall we?

UPDATE:
And the results are in. RAR: 150 MB vs 150 MB.
7zip: 152 MB vs 150 MB.
I can instruct how to recompress and verify these results, if you want.
But it goes to show that compressing a video using good video compressing yields in little or no saved space.

Or one could turn the tables on you and say you are hand picking your videos to be not compressed much, so that it would show you to be correct. Fact of the matter is, most times people share media across the net, especially audio and video, it is compressed to the point that using a standard file compression program is frivolous. Take anime for example, I have about 150gigs of anime, and about 150 gigs of that wont compress worth a damn if I ran it through 7zip, RAR, or what ever compression software you want to use.

Originally Posted by abachler

except again, you are choosing specific examples, whereas I have also shown examples that show excellent compression adn are also Mpeg4. Your claim is that they compress little or not at all, where I have shown that not to be the case. I can show you just as many examples where Mpeg4 does compress as you can show that it does not. You are taking a few specific examples where the Mpeg4 was highly compressed to begin with and makign a broad sweepign statement that those examples are representative of MPEG4 as a whole. I have shown that they are not.

I've chosen files that represent about 90% of the material on internet. Every video you come across, except really old ones, uses either Xvid, DivX or H264, usually x264.
MPEG4 ASP compresses about 2-3% and H264 pretty much none.

You have to have some kind of dearchiver to begin with, so that complaint is mute. It is common practive to compress files for distribution. Once you have spent the time to compress the file there is no good reason not to distribute the smaller of compressed/uncompressed. The situation where you do not have enough disk space to decompress the file is rare. I would hazard that 99.9% of users have plenty of extra space to decrompress. If you are that low on disk space then perhaps the real issue is why you havent upgraded to a larger HDD.

Decompressing a RAR is just like opening a folder. WinRAR doesnt fully decompress the entire image, just the directory, so you can see whats there before decompressing it. I realyl don't see how that is any more of a pain in the butt than using an uncompressed file. Particularly if you are so low on disk space as you claim to be, you shoudl be keeping all yoru files in compressed archives and only decompressing the ones you need to use.

As for the limitation in the samples, its simple. She claims that A cannot B because C. I show an example D where A can B. Therefor her argument that A cannot B is false by counterexample regardless of how unrepresentative C and D are.

It is common practive to compress files for distribution.

I think Wraithan already addressed this point: "In most cases, people have already compressed the video before it is RAR'd/7zip'd/ZIP'd so the effects of a second compression is negligable, the point of the OP was that no type of compression is much better than typical video compression."

An anology for (open source) software would be compressing a compressed tarball.

Originally Posted by abachler

As for the limitation in the samples, its simple. She claims that A cannot B because C. I show an example D where A can B. Therefor her argument that A cannot B is false by counterexample regardless of how unrepresentative C and D are.

So what are implying is that my test case is pretty irrelevant simply because there are other compressions out there can be far better compressed?

Originally Posted by Elysia

So what are implying is that my test case is pretty irrelevant simply because there are other compressions out there can be far better compressed?

Not irrelevant, just not deterministic.

The reason that it is common practice to compress files before distro is that they usually DO compress and once the processing time has been spent to compress them, even a small savings in size makes it better to distribute the compressed version. So essentially the cost to use the compressed versus uncompressed is zero and any savings in bandwidth is free. I'm not saying that MPEG4 cannot achieve good compression, I am saying that for a lot of reasons it is usually not the case that mpeg4 files are compressed at their highest settings and that RAR can usually achieve some improvement. This improvement can range from 2% in elysia's example to 33% or more as in my example. I figure that '10% on average' is a good rule of thumb then.

Yes, of course, but a few kilobytes saving is nothing. It all depends on the compression used, because video compression is far better at saving size than compression utilities. And as shown, with H264 and very high compression settings, compressing it using winrar saves nothing. It's just a waste of time.

Also - if archive is password protected - it will NOT be scanned by the Antivirus software on server and on Firewall, saving the CPU/HDD access times on server

Originally Posted by abachler

As I stated, video is not always compressed using the highest settigns for various reasons of quality/speed. Therefor it is prudent to at least TRY to compress the video first. Once the RAR ( or zip etc) compression has been performed, there is no reason not to use the smaller of the compressed or uncompressed versions. 2% may not seem like much, but its enough to save on the upload time which is realyl what most distributers are worried about.

For quality reasons? The only argument against heavy video compression is decoding speed and encoding time.
Still, if you encode using MPEG4 ASP/AVC, it's hardly worth the time compressing it. For other sources, perhaps... but 90% of the material is MPEG4 ASP/AVC.