Playback Lossy vs lossless

[Toshi.des]

I've heard and read that lossless video is easier on resources to playback than lossy due to that the computer has to decompress it first.

The only thing is, it's always seemed the opposite to me.

I've two computers I use frequently:

First:
i7 2.0 2nd Generation
16 GB DDR3 Memory
nVidia 460M 2GB VRAM
160GB SSD OCZ Vertex

Second:
Core 2 Quad Q6600
4GB DDR2 Memory
nVidia 260 896MB
150GB WD Raptor 10,000RPM

If I re-encode Full HD Video with UTvideo it always plays back a little jerky, where as any form of compressed video that I've come across has never had a probably playing back.

So just curious if that's really true.

[Kevmaster]

I've heard and read that lossless video is easier on resources to playback than lossy due to that the computer has to decompress it first.

I never heard of that. Lossless had always been harder to playback for me. Couldn't playback on my old PC at all and even on my new one, depending on how much stuff I have opened it'll stop and needs to buffer.

EDIT: However, maybe this is what you meant you heard: lossless works better for editing purposes. Editing softwares just seem to not handle compressed video as well.

[Cannonaire]

Lossless codecs such as UTVideo and Lagarith are not intended for normal playback, and as such you probably will not get reasonable results trying to open them in something like MPC-HC. However, for editing purposes I have never had trouble with them as far as playback speed goes, even on my old Pentium D e5200.

In regards to comparing lossy codecs to lossless for editing, one of the most important benefits comes from intra-frame compression: lossless codecs like the ones mentioned above store complete frame information for every frame, whereas lossy codecs most often use keyframes at set intervals, and between them only the changes in picture from the previous frame are stored. Practically, this means that if you try to seek backwards with lossy codecs it is likely that you will get mixed, often inaccurate results.

For a fun test, try opening some footage encoded with UTVideo in VirtualDub. If everything is configured properly, it should have no trouble playing the footage back in realtime. You should also be able to randomly jump to any part of the footage without delay or image quality problems.

[Cannonaire]

Sorry for the double-post, but I forgot to add something.

For lossless full HD footage, it is likely that your hard drive throughput is causing problems with playback speed. The actual decoding of the file should be more or less easier on your processor, but the file size will be much bigger. AFAIK professionals edit with lossy footage, but they generally have higher quality footage to begin with and they also use only intra-frame compression with their intermediate (editing) codecs. This gives them the advantage of maintaining quality and striking a good balance between decoding speed and hard drive reads. Because AMV editors generally edit with compressed footage from DVDs (or worse...), using lossless compression minimizes generation loss.

[Mister Hatt]

UTVideo is designed for high-speed ENCODING, not decoding. Lagarith is designed for compression and use in editing, not playback. Lossless AVC is designed for fast playback. They all rely on being read off the disk fast enough though, especially UTV and AVC due to their higher bitrates. Lagarith will usually lag more noticably due to it being a significantly more complicated codec.

Either way, playback speed depends on the speed of the disk it's on, and the complexity of the codec. The fact that it's lossy or lossless has little to nothing to do with it. Other than that, what Cannonaire said.

[Kionon]

I always insisted on editing with lossless source even when I was doing pro work. Not everyone agreed with me, but since we often recorded in lossy, I saw absolutely no reason to covert lossy to lossy to edit with lossy and export to lossy like some of my colleagues did. I always tried to minimise generation loss no matter how "high quality" my lossy footage started off with. Unfortunately, given the limited storage space on our cameras, of course the video had to be stored originally in a lossy codec, but I didn't want to keep it that way.

I have typically used Apple Intermediate Codec or ProRes422 if absolutely demanded of me, but I prefer not to. Instead I prefer to use Animation. When dealing with AMVs, I still use HUFFYUV, but I'll be making the change to UTVideo for my newest projects (some of my projects are years old, and source is already stored in HUFFYUV, so I see no reason to change them over).