- Last Updated on Sunday, 24 November 2013 03:17
Distribution Codecs and Codecs for Editing. This distinction is pretty important, because distribution codecs are not suited for editing, they often cause problems, will always result in disappointing quality when re-delivered and frequently require conversion to another editable format before you can start editing.
Let's start with the basics. A video Codec is a way to Compress and Decompress video.
The higher the resolution of the video the more complex it becomes. SD with only 720 x 480 resolution is pretty simple, HDV or HD with either 1440 x 1080 or 1920 x 1080 is getting a bit more complex and RED 4K or even EPIC 5K makes it hugely more complex, because of the sheer number of pixels to compress. The balance is always to keep file size and compression ratio manageable.
Let's make a small detour to show the difference between Distribution codecs and Editing codecs. Imagine two forms of vacation, backpacking, and a 5-star luxury resort to really show the differences and it all relates to the luggage you haul along.
Backpacking: All the luggage is stuffed in a single backpack, it does not matter if it comes out all wrinkled, all that matters it is lightweight, does not contain stuff you don't need, is highly compressed to keep the size of the backpack small and gives you the clothing you need when the weather turns cold, wet and windy, CWW. The same with distribution codecs, they are also CWW, Compressed, but watch-able on the WEB. A backpacker is ill-equipped to join a luxury diner-dansant where 'black-tie' is requested, since he did not pack it in the first place, all he has is heavily wrinkled casual clothing, and his dirty trekking boots. He is ill-equipped to join a golf match, since he did not bring his clubs, etc. The same with distribution codecs, they are ill-equipped to do anything but play-back. Too much has been lost to compression, getting rid of stuff not needed and weight considerations.
5-Star Luxury Resort: All the luggage is loosely packed in a number of Louis Vuitton cases to avoid wrinkles, the number of cases is irrelevant when there is a Bellboy to take care of them, the tuxedo, golf clubs, and dancing shoes and whatever else comes to mind for such a vacation are also packed. Nothing is forgotten and who cares if you packed stuff you did not need? This is like an editable codec, lightly compressed to avoid wrinkles, a number of cases, so you never end-up empty handed and are prepared for every eventuality.
It is obvious that backpacking is far less expensive than a 5-Star Luxury Resort, and the same applies to the computer requirements for distribution or editable codecs. For only playing a video, you need a simple computer, to edit you need much more muscle from the computer, like a Bellboy, to carry all the extra luggage. The following chart shows that:
Material downloaded from the Internet tends to use distribution codecs and - generally - is ill-suited for editing. The material has been compressed so heavily, that it is extremely difficult if not impossible, to edit at all.
Conversion of a distribution codec to an editable format, if it is successful, will only lead to disappointing results. During the initial compression, so much video information has been irretrievably lost and can never be recovered, that the new 'source' material is at best only mediocre to start with. Upon final delivery, it is compressed again with further quality losses. The general advise is to get the original material in editable format from the copyright owner, avoid legal issues and improve the editing experience and end result. Do not think that because you try to edit an .AVI file, that it is an editable format. AVI is just a container that can hold very different codecs inside. Whenever you have trouble editing an .AVI file, it is a very safe bet that it only contains a distribution codec inside, not an editable codec.
If you have access to a video camera, most of the codecs used are very editable. In contrast to some other editors, PR allows editing in native format, so all you need to do is either 'capture' over an IEEE 1394 firewire connection for tape based cameras, 'import' media from card based cameras by copying the entire contents of the memory card to the hard disk, or 'ingest' over a HD-SDI (or the 'poor-man's choice. over HDMI) connection from the camera to PC. There is no need to convert this material to an intermediate format in order to edit comfortably, unless the computer capabilities are really marginal. Of course, a faster PC will give a 'snappier' feeling, stutters less during playback and is more responsive, but you don't need to convert to another format.
The drawbacks of converting from a native format to some intermediate format are:
- Extra time to create the conversions,
- Extra disk space for the converted material,
- The lower CPU load (because it is easier on the CPU), may be offset by the increased demands on memory and disk I/O,
- The intermediate codec may not be (visually) loss-less, so quality deterioration can occur.
The general advise is to edit natively, unless it really is cumbersome and limits productivity. The alternative to upgrade hardware, may be the better solution.
For PC users there is a further complication and that relates to source material that uses the QuickTime 32 Server add-on, for .MOV or ProRes or even DNxHD in a QT container. QuickTime on a PC is a 32-bit application. It not only shows the dreaded 'gamma-shift' problems, but the major problem is the 32 bit nature of QuickTime. That simply means that the processing, when using the QT 32 Server add-on, occurs in a 4 GB memory space, even when 64 GB of memory is installed. The consequence is much more page-file swaps than would be normal on a system with lots of memory, causing severe performance degradation, but the other drawback is the very limited threading capability of this add-on, causing a far from efficient use of all the logical cores in the system. That is the simple reason ProRes 444, ProRes 422 and DSLR QT are ranked in the most CPU demanding column in the graph above.
If you use any version prior to PR CC 184.108.40.206, avoid using anything QT related, like QT, ProRes or DNxHD .MOV. You will be doing yourself a huge favor.
The camera you own, rent or lend to get your original source material from, is not something you can do much about. Your choices for the codec in use are limited or non-existent. Take it as you go. But realize that some codecs are very demanding on the system, and others are 'easy' on the system. That has consequences for the hardware needed to edit comfortably without stuttering, without jerkiness. If you have the choice between 422 and 420 encoding in the camera, always use 422, since it holds up better in post, if you have the choice between different compression rates, choose the least compressed one, that is the one with the highest xx Mb/s. It is much easier on the CPU and gives better quality. It takes more disk space, but disk space is cheap.
Typical codec problems one can encounter.
- My .AVI or .MOV file plays fine in a software player like WMP, VLC or QT, but gives an error message in PR.
- This is typically caused by a distribution codec in an .AVI or .MOV wrapper, that can contain numerous different codecs. The solution is to convert these files to an editable codec externally and take the quality hit or get the source material from the copyright owner.
- Codec missing or unavailable message.
- Same as above.
- Playing the timeline gives stutters, jerky behavior and does not play fluidly.
- This indicates that the hardware is not fast enough to handle the data-rate of the source material with the codec in question. Tuning the system, upgrading the bottlenecks or simplifying the timeline can help to solve the issue. Often it is a matter of adding disks or memory to the system, but if that is not feasible, using an intermediate codec like Cineform or Lagarith can also help to avoid the stuttering and jerky behavior.
- Audio and video turn out-of-sync.
- Often this is caused by not using Media Browser to import clips into a project or not heeding the advice to always copy the entire content of the memory card from your camera to the hard disk, including the complete directory structure. Alternatively, it could be caused by importing material that was first ingested on another platform, Mac for instance, where FCP created an intermediate format. Again the solution is to get the original material and ingest again.
- After import the material either does not fill the complete screen or is zoomed in.
- That is typical of OE, Operator Error, where the sequence settings do not match the source material. Create a new sequence that matches the source material, using 'New Sequence from Clip'.