How do you avoid recompression with DSLR footage?

[David Cherniack]

Unprocessed footage cannot go straight through unless it’s uncompressed.

It depends what your end use is but if it’s for later processing your best bet is to render to a high quality codec. If you have FCP on your system then Prores, if not then DnxHD or 10 bit uncompressed if you have the drive space.

David
AllinOneFilms.com

[David Cherniack]

BTW if you stop to think about it for long GOP material such as DSLR it’s not possible to pass cuts through without recompression. If the cuts occur anywhere in the middle of the GOP the surrounding frames have to be rendered.

David
AllinOneFilms.com

[Phil Radelat]

I don’t think you understand what I’m talking about. When I edit with MiniDV footage, any footage not run through any process goes through completely un-re-rendered when outputting to a file. Only anything that was applied some sort of process will be re-rendered. Not only does this maintain the original MiniDV image quality, it makes for fast write time on the render because only areas processed get re-rendered. The unprocessed data goes straight through to the final project output file untouched.

I want to do the same with the h2.64 footage coming from my camera. I want a project/output setting that will NOT re-render anything that wasn’t processed. Is this possible to do in PP CS5?

[David Cherniack]

Phil, that may be true for DV which is an intra frame codec ie each frame is a complete whole. But as I tried to explain above it’s not true for your camera AVCHD codecs which are inter frame (Long GOP) and thus there’s no way any NLE can directly pass your cut clips through on export. At the very least the partial GOPs surrounding your edits would need to be rendered. I think MainConcept’s Smart Rendering plugins do that for mpeg2 but you’d have to check with them to see if AVCHD is supported on output.

David
AllinOneFilms.com

[Tim Kolb]

David’s got this one spot-on. Any footage that is temporally compressed can’t be just simply “cut” even in theory (unless you only cut on the I-frame of the GOP if you knew where they were…) as those GOPs need to re-order.

If you cut out 5 frames that you don’t like…that 20 frame GOP is now only 15…or maybe one GOP is 2 frames shorter on the tail, and the next one is 3 frames shorter at the head and since you’ve cut out it’s I-frame so it has to generate a new one…etc, etc.

Intra-frame means every frame is independent. DV, ProRes, DVCProHD, etc are all Intra-frame and you can cut them up to your heart’s content as each frame exists in whole.

TimK,
Director, Consultant
Kolb Productions,

Adobe Certified Instructor

[Phil Radelat]

So ultimately the answer is no, you can’t pass-through unprocessed data without re-processing. At least, not in PP.

This doesn’t really make any sense to me, I can’t see why you can’t pick up from a full frame and continue with the original data stream until a deliberate process has been induced somewhere on the timeline. It’s not like the software can’t tell where the next I-frame is, it’s clearly identified in the data stream.

Oh well. So then what’s the next best option in PP? An intermediate format like ProRes is a complete waste of time and an additional unnecessary compression hit for the type of work I want to do with this footage. So it comes down to what’s the least damaging export setting, I suppose.

[Tim Kolb]

[Phil Radelat] “So ultimately the answer is no, you can’t pass-through unprocessed data without re-processing. At least, not in PP.”

There may be some dedicated MPEG utility or something that does this, but unless you plan on counting frames and making your cuts every “X” (assuming you know how long the GOP is) frames, you’ll be sliding the GOP structure. This is not new as HDV and XDcam are the precedent being MPEG2…to AVCHD and DSLR MPEG4 (h264)

Anytime you go to a less aggressive compression scheme, you’re typically less degrading to the image. DSLR footage has some different data rates depending on your brand…so it isn’t as if it’s “one” format per se.

The best forensic preservation of the image will typically have a direct relationship to the data rate…the best visual quality depends on who’s looking and what they’re viewing on more often than not.

The advantage with PPro over say FCP is that you don’t have to convert all your raw footage, you can cut it natively and master the edit to something more robust.

TimK,
Director, Consultant
Kolb Productions,

Adobe Certified Instructor

[Phil Radelat]

Well, I guess Adobe has some work to do in this area. THIS is what it needs:
https://www.ulead.com/spotlight/hd/runme.htm

If this program can do with HDV footage with every editor has done with MiniDV for ages, so can PP. There’s simply no excuse for it. This is exactly what I was referring to.

I will be downloading a trail version of this editor to see if it lives up to the claim. I used to work with this particular program years ago in the MiniDV age, so I’m somewhat familiar with it.

While PP certainly has a lot of features and functions that I like, unnecessarily re-compressing HDV footage is a real turn-off.

[Jon Barrie]

Hi Phil,

I’m not sure I understand where you are coming from here.

Adobe Premiere Pro has supported native HDV editing/capture (using the MPEG-2 Stream) since it started supporting HDV back around version 1.5.1

It doesn’t convert it to anything else. Sure there are Cineform options but that’s a 3rd party product.

I don’t see anything that the Ulead software does that’s any different to what you could do with Premiere Pro.

– JB

Jon Barrie
aJBprods
Jon’s YouTube Tutorial Page
follow Jon with twitter

[Tim Kolb]

I believe he’s referring to the promise it will only recompress altered frames…though that is a touch vague. I have heard of this being done with MPEG2…but not MPEG4.

I suppose someone will come up with a way to do sooner or later (use a partial GOP where it’s been edited so it doesn’t alter the surrounding GOP cadence, triggering complete GOP re-ordering downstream), but since MPEG4 is using not only pixel redundancy, but also vector edge tracking and motion tracking, etc, it may take a while before we see it.

HDV is probably the main format where a large percentage of users would desire to ingest, edit, and go back to the camera format to master…I think an awful lot of folks these days are acquiring on whatever they acquire on and mastering or exporting to whatever is needed, so this request to go back to DSLR format may not have come up before…

…sounds like a feature request.

TimK,
Director, Consultant
Kolb Productions,

Adobe Certified Instructor