About size of videos

[shinigamie]

When someone uploads something and lets say the size of the file is 1,690,875,858, then someone else downloads that file but then uploads it to a different host or the same host just under their account but the size increases by one tiny byte from 1,690,875,858 to 1,690,875,859. Should i be worried about that? Or is that just something that happens?

 

[SamKook]

It should not happen assuming you're checking the size locally, on the same computer and in the same way since different system or software can have small differences like that depending on how they round things up.
That means part of it got corrupt at some point during a transfer.
That's why I always use rar archives instead of binary split since it does an hash check when extracting to make sure it's still intact and also include a md5 hash file so people can check the integrity of all my uploads(although I'm sure almost nobody does).

With that said, you probably won't notice a 1 byte change which would probably show up as a very small glitch in either audio or video.

 

[shinigamie]

Well, guess my ocd is going to be an even bigger problem for this.

 

[SamKook]

Your best bet is to stick with the original uploader as much as possible and you should be fine. It's not something that should happen very often unless a specific uploader is using a crappy software/connection for download/upload in which case, you should avoid his post if possible.

 

[shinigamie]

So the file with that one byte difference is obsolete compared to the other? This isnt just a misread from the hard drive. Basically its just utter garbage compared to the one with that 1 less byte??? ;____;

 

[SamKook]

If you want to make sure it's not a misread from the hard drive or something, just hash both and compare them. If the hash is identical, that means both file are identical. If not then one of them is obsolete compared to the other. I wouldn't say complete garbage since you most likely won't see a difference or maybe the reuploader just added his name in the file metadata or something which wouldn't affect anything at all other than the size.

 

[shinigamie]

So that one single tiny byte is a big difference?:cry5:

Im trying to comprehend to the fullest extent how data works. I mean, will it make a difference at all if i download one that has that on byte difference? I may be asking you things that u probably arent incredibly knowledgeable in, so my bad if it starts to seem that way, ocd.

I try finding the files without that one byte difference but sometimes links go dead and u have only one choice.

Would you give a shit I guess is the question. Im sure u have downloaded at least some files here and there with that 1 byte difference. Did you care?

Plus a added watermark would be way more than just 1 byte.

 

[SamKook]

It's certainly not a big difference, depending on the codec from the video and which bit is different, it can range from 1 pixel out of hundred of thousands in 1 frame out of hundred of thousands in the video that is the wrong color or at worst, on older codecs, a series of pixels that are flipped for a second or two.
With audio, you either won't hear it at all or it will make a little out of place sound for a fraction of a second.

If you don't already have both files in you possession, you can't know if the 1 byte difference is real or not since different OS/software round up the values differently so unless you're 100% sure the testing condition are exactly the same for both files, you can't know if there's really a 1 byte difference between them.

And even if there is, you shouldn't really care since the chance you'll notice a difference are slim to none. If you can get it from the original uploader, than do it since it reduce the risk, if not, then it's no big deal and get what's available.
Back when I did fansubbing, the only copy we had of a couple of our oldest files had the wrong hash(meaning at least one byte in the files somehow changed since their creation), but we looked at them and didn't find anything wrong with them so we just used the new hash and re-released them like that. Unless you actually see a problem with it, there's no reason to care if it's a little different.

I didn't mean an added watermark when I talked about metadata, I meant that video file containers(mkv, avi, mp4, ...) have a section that contains data unrelated to the actual video and audio file. In this section, one could add his own name as the creator of the video file for example and this would in no way affect the playing of the video. All it would do is that with a player that can read those information, you could display them if you wanted to. You could even add junk in the container and as long as it's in the right place, it won't affect the movie at all other than the size.
For (an extreme) example, I could add a 2GB file containing only 0s in a .mkv and the video would play perfectly as intended and be exactly the same as if it didn't have that extra 2GB file inside.

 

[shinigamie]

AH i see what u mean, like not a watermark but like that extra info added to a file when u play it. That would explain the 1 byte increase.

Would filling up the file name with a bunch of Japanese writing increase it by 1 byte?


MAYBE yes no ?

 

[SamKook]

I'm not 100% certain, but I don't think the name of a file is stored inside the file itself, I think it's stored inside the partition filesystem(NTFS, EXT4, etc...) and associated to a file, so it wouldn't affect its size.

 

[shinigamie]

I'm not 100% certain, but I don't think the name of a file is stored inside the file itself, I think it's stored inside the partition filesystem(NTFS, EXT4, etc...) and associated to a file, so it wouldn't affect its size.

Wait a second, i just thought about something. Yipman uploads his stuff in rar files. And you said its better to upload stuff in packed in rar's because it hash checks when being unpacked.

Okay so if this is the case then your size will be the exact size as the original uploaders right? There wont be that one small byte difference right?

So why does it still happen then?

Ive seen people put yipmans uploads in torrents. But in order to make torrents out of them the person would first have to download yipmans uploads and then unpack the file and make a torrent.

So why is the 1 byte over still appearing if yipmans uses rar to pack his files?:huh:

 

[SamKook]

Because once it's unpacked, nothing is stopping it from changing. The guy who downloaded it can do whatever he wants with it or a problem could happen during the transfer.
The only way to avoid it is if he redistributed the same rar files as Yipman instead of the unpacked movie, then you'd be certain it was the same as the original.

Also, I've experience unreliable behavior with torrents.
I downloaded a file which contained an hash in it's filename, when I checked it, it was different(meaning the file I downloaded was different from the original).
I told the torrent client to rehash it, to make sure it wasn't a mistake the client didn't catch during download by skipping a check or something and it told me it was fine.
I then downloaded the file from another non-torrent source and that one had the right hash once checked(meaning itwas different in some way from the one I got from the torrent).
I deleted the one I got from the torrent, replaced it with the one I got from an other source and asked the torrent client to rehash it to check it it thought it was ok.
To my big surprised, it told me everything was fine with it which means it saw 2 different files as being the exact same one. I've never trusted torrents ever since.

It might have been a bug in the software I used at the time(utorrent v2. something) and be fixed today though, if it's not a problem with the protocol itself.