Codecs, standing for "compression/decompression" are mathematical algorithms to shrink overall filesize of assets. There are codecs for both video & audio. Some image form compression algorithms translate to video, others do not, since video is just a bunch of images stitched together.
RAW video, usually, is a series of images & audio. Depending on format, they can be independent entities (like Quicktime's "tracks"), or just called tracks, but really a mesh (like MPEG for example). Usually, even RAW video is broken down into mJPEG (motion jpeg) to be more managable since although mJPEG is a lossy codec, at such a high setting (like 90), the video still looks good, but drops significantly in filesize. Codecs, like image formats, come in 2 flavors: lossy and lossless. Lossy compression means you lose image quality when using the codec. JPEG does this by removing colors the human eye cannot see (nor can a lot of computer video monitors render). Most go farther depending on how low you set the quality level. Bottom line, the more you compress something with a lossy codec, the more image degradation will occur and overall, it becomes to look worse and worse (pixelated, blurry, smudged, weird colors). Lossless codecs, like PNG, do not negitively affect the image quality in anyway, they mere use common mathetical algorithms to shrink the filesize of each frame. Lossless, however, have a set value, are usually not too configurable, and you have little to no control over "how compressed" something gets. Lossy codecs use 2 basic methods to compress video. Frame compression and time-lapse compression. Most codecs use something called a keyframe. There are other types of frames too, but the point is, you take a point in the video, a frame that is pretty high quality, and then save the information. So, for a talking head video for instance, they'll compress the backround more because if it's all 1 color, say bluescreened or white, it's really easy to not only use losslesss compression (like GIF where it uses lzw and turns a ton of white pixels it 1 white pixel). Timelapse, they'll use that 1 keyframe as a guide to how to compress the rest of the frames, say the next 30. If it's a person talking, usually only parts of the face change, while the rest doesn't, so there is little point in redrawing each of the other frames if the face doesn't change. Additionally, you don't even need to store pixel information about those frmaes if you are just redrawing really small parts. There are other types of frames used in different codecs, but those are the basic 2. Spark & On2 are both lossy codecs. As such, you can control the level of compression (usually by the datarate, or how much kilobytes is used per second). So, if you have 30 frames per second, you'll effectively have 700k to distribute amongst 30 small images, assuming you go with On2's highest, default datarate. You can quickly see how lowering framerate of video drastically increases quality since going from 30 to 15 frames per second doubles the amount of kilobtyes each frame gets to use. ...obvoiusly, audio usually uses mp3. Now, re-compression of compressed video usually screws this up on a number of levels, resulting in worse looking video, and higher filesizes. 2 main reasons for this. First off, the video already looks bad. No video codec in the world makes something look BETTER after you'ev used it, even lossless. You are losing infromation somewhere, and with lossy, you can be sure image quality will degrade. If you do it again, you are degrading something that is already degraded, thus degrading it more. What does that mean? You took something that looked bad and made it worse. Secondly, codecs are designed to find common light and color patterns in video and compress based on those. The pixels that are left over via redraw, as well as the blurring of color, and added noise to the compressed video not only confuses the codec, but gives it less information to work with. Go take a JPEG that';s comprssed to 50 percent, and then compress it again, and you'll see how both visually it looks like crap, and filesize doesn';t improve. It is common in the video world, however, to not save the source. Since uncomprssed video, even using mJPEG to compress it once still takes up gigs and gigs of space (usually a hard drive or two), you can't just "have it around" unless you work in the video industry and have the space for such things (like DV tapes, DVD-ROM storage devices, huge RAIDs, etc.). ----- Original Message ----- From: "Merrill, Jason" <[EMAIL PROTECTED]> To: "Flashcoders mailing list" <[email protected]> Sent: Monday, November 21, 2005 3:15 PM Subject: RE: [Flashcoders] Flash 8 grossly inflates .flv file? Yeah, I think I will just tell them to bite me. Not really, but something like that. Is this true of any video editing software? - that compressed WMV files get inflated when you try and convert them to another format because the codec freaks out - or is it just an anomaly/bug with Flash 8 and/or the available codecs? Jason Merrill | E-Learning Solutions | icfconsulting.com >>-----Original Message----- >>From: [EMAIL PROTECTED] [mailto:flashcoders- >>[EMAIL PROTECTED] On Behalf Of JesterXL >>Sent: Monday, November 21, 2005 3:11 PM >>To: Flashcoders mailing list >>Subject: Re: [Flashcoders] Flash 8 grossly inflates .flv file? >> >>No one ever does it seems. You're best bet is to use Spark then, although, >>the size will still be unnacceptable. NOTICE: This message is for the designated recipient only and may contain privileged or confidential information. If you have received it in error, please notify the sender immediately and delete the original. Any other use of this e-mail by you is prohibited. _______________________________________________ Flashcoders mailing list [email protected] http://chattyfig.figleaf.com/mailman/listinfo/flashcoders _______________________________________________ Flashcoders mailing list [email protected] http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
