-
DSLR transcoding/decompressing. Necessary?
I have been reading for a while about the necessity of decompressing DSLR footage, so I have been doing so using the Squared 5 software. I have always been told that this is required if you plan to do a lot of post-work.
The Adobe suite (I use CS6) is advertising it’s ability to work with DSLR footage natively. Does this mean decompression no longer has any benefits?
Lastly, what is the purpose of this step in the first place? Is it simply to make the AVC / H.264 compression (I shoot with a 7D) more edit friendly? Or does it have anything to do with color space? I mean, transcoding to ProRes422 for example can’t add any image information than what was written to the card in the first place can it?
If my system (computer/software) can handle it, should I just skip this step, or will I be losing any image information?
Thanks,
Bryan -
9 Replies
-
Bryan,
H.264 is an intra-frame codec. So a frame of video will contain information that references another frame somewhere else in the file. This is why it is not ‘edit friendly’ to some NLEs and some computers. But if your computer can handle it (CS6 can defintely handle it) than by all means edit natively.
Sometimes when doing motion effects in After Effects, such as rotoscoping or many h264 FX layers, that intra-frame codec will cause sluggishness or blank frames. That is what people mean when they say “a lot of post-work”.If you don’t experience any of these problems, there is no need to transcode before you edit, just start cutting natively. Transcoding does not give you any image quality advantages either, because like you said, you can’t get any more image data out of that H.264 file once it has been compressed to H.264.
JohnJohn Young
Surrounding Media -
CS6 rocks with Native footage. I just cut a video that was over a hour long and was shot with three different codecs and it worked great. I was extremely happy with the results. Premiere has limited ability to get its RAW footage into Color Grading programs and I hope that is addressed in the next version.
Stephen Smith
Utah Video ProductionsCheck out my Motion Training DVD
Check out my Vimeo page
-
can you say, open API Dynamic Link!
-
Bryan, you don’t need to transcode anything. Shoot, drag your files into Premiere and edit – that’s the best workflow.
In the 10 years that I’ve been in this business, I have yet to see one single example of where an intermediary codec (Cineform, Prores or DNxHD) has any impact on the quality of the footage.
Computers today are more than capable of handling interframe codecs like AVCHD and H.264 in real time, and more. You don’t need the high data rates of intraframe codecs.
Hope this helps.
https://www.wolfcrow.com – Workflow information and support for filmmakers, photographers, audiographers and videographers.
-
Sareesh and the others are correct. It appears that since CS6, the only place transcoding is necessary in the workflow is where the computer is not up to spec, and in the delivery phase of the project, where, of course, proper codecs continue to be critical to the quality vs. bandwidth issues.
Joe Bourke
Owner/Creative Director
Bourke Media
http://www.bourkemedia.com -
Bryan,
I’ve asked a similar question of a couple of friends, both of whom work ongoing with RED on the job and HD enabled DSLR’s for hand held.
The subject came up regarding HDMI direct capture, to an external device from the D800 Nikon, and Sony A99.
The short answer, according to them is if one is using heavy AE motion graphics and effects, then there is a case to be made for ingesting I frame specific codec.
Generally as noted, CS6 works fine without that step and yes, if one has an older computer or previous version of Premiere then trans coding also as noted (can) make the process smoother reducing the stutter.
RM
-
there’s more to it than that…
Nikon D4 and D800, GH2, GH3, and soon 5D Mark III are all 4:2:2 hdmi uncompressed with Blackmagic Shuttle, Atomos Ninja, or Aja Ki Pro Mini.
why?
ProRes is a lossy intermediate codec that gives you a much higher bit rate and less compression artifacts than your DSLRs’ H.264 or AVCHD file.With the Ninja, you not only have the ability to capture clean HDMI in 1080p at 24fps from your camera’s sensor, you’re also working with harddrive storage. This means that you no longer have recording limitations of 10 to 15 minutes per clip, and you get focus peaking, zebra etc.
Ninja Tested with popular DSLR cameras with videos
https://atomos.activehosted.com/kb/article/ninja/ninja-tests-with-popular-dslr-cameras -
[John Young] “H.264 is an intra-frame codec. So a frame of video will contain information that references another frame somewhere else in the file.”
Sorry to correct you John, but is right the opposite.
Intraframe has all the info on just one frame. No need of others for encoding/decoding.
Those who needs other frames to be encoded/decoded are GOP or “interframe”.
H264 can be both, Intraframe (as on the Canon), or GOP, as the H264 flavors that we use for BluRay, YouTube, etc.
rafael -
Hello there,
I hope that one day I understand codecs as well as you do.
Are there any books that deal with all the technical/objective aspect of video?
Reply to this Discussion! Login or Sign Up