Should I convert MOV files to ProRes 422?

[jnorman34]

I am still rather new to video after spending many years doing still photography of architecture.  My new project is documentary video footage of architecture, construction sites, public spaces, etc.  I am currently using a D5300, and have a GH4 on pre-order.

 

After processing some footage from the D5300 in Premiere, I do notice a slight degradation in the rendered footage compared to the original footage.  It was suggested to me that I convert the original MOV files to ProRes 422 using mpegstreamclip.  With many hours of video to wade through, converting every clip to a new format for editing would add a lot of time to my post processing work, and I need to be certain that it is worth it.

 

From some of the threads I have read, it sounds like only people who use FCP really need to do this, but it is not clear to me.

 

So, how necessary is it that I convert my MOV files to ProRes 422 for editing?  Does it actually result in cleaner rendered files than if I do not convert prior to editing and rendering?

 

And, when I begin using the new GH4 at 100mbps bit rate for 4k, will converting to ProRes still be a necessary addition to my workflow?

 

Thanks.

[Pascal Garnier]

I always convert all my footage to ProRes422 (not HQ).  I use 5DtorRGB Batch and convert everything right after offloading.  MPEG Streamclip is faster (and free), but there is a quality difference between the two.

 

[Sean Cunningham]

5DtoRGB does a better job of up-sampling MP4 than Adobe products.  It's often subtle but it's definitely not a FCP-only issue.

[Guest Ebrahim Saadawi]

Only convert to ProRes if your machine can't edit native h.264 files. Converting to ProRes doesn't improve the quality as far as I can tell, at least, not enough to justify the additional time required.

[Bioskop.Inc]

Simple answer is, if using:

FCP7 use 5DtoRGB

FCPX don't bother

[Sean Cunningham]

Only convert to ProRes if your machine can't edit native h.264 files. Converting to ProRes doesn't improve the quality as far as I can tell, at least, not enough to justify the additional time required.

 

Converting to ProresHQ via 5DtoRGB (there may be others that do an equal or better job) does mild improvement via chroma filtering.  It's not the kind of thing that jumps out and bites you but if you know what kind of footage shows off the limitations of 4:2:0 or 4:1:1 footage it's easy enough to see.

 

Adobe for one does not filter chroma when you import low fidelity file types like h.264 though YMMV with FCP, I dunno there.  The effort and additional storage is only worth how much you care.  I figure anyone who even has the presence of mind to ask is a valid candidate for someone who might care enough to go through the effort.  You should do it after locking your edit, however, so that you can build just a list of the clips to "bump up" and substitute.  Unless you can't edit with H.264 native and need to just do it all anyway.

[jcs]

Going from 420 to 422 is only going to (possibly) improve the horizontal chroma. If the tool is improving quality better than an NLE, ideally you'd go to 444 (both horizontal and vertical). One way to test to see what's going on is to bring in both clips (transcoded and original) to the NLE and place on the timeline/sequence, one above the other. Then set the top track blend mode to Subtract. Add a gamma filter/effect to the top clip and crank it up until you can start to see differences.

 

Another test is to bring the clips in and just A/B (track toggle) the two clips. If you can't see any difference visually, not worth it to transcode (even if you see a difference with the Subtract+gamma test).

 

Pascal's two clips differ in a major gamma/contrast shift.

[Sean Cunningham]

I confirmed just in the last week my suspicion that 5DtoRGB improved even All-Intra footage and this was visible with a simple A-B between the two clips lined up on top of each other, in both Premiere and After Effects, without any special tricks to highlight the differences (being slightly zoomed in did help, however).  Boundaries where the edges of very colorful objects meet darker parts of the image were smoother in the transcoded Prores.  In the instance of two brightly colored objects that overlap or otherwise share an edge this smoothness was even more apparent.  I also noticed areas with chroma noise being overall a little smoother (though that was only really apparent under magnification)

 

Not all footage will show off the improvement of course.  You need to look at the kinds footage or subject scenarios that hits color under-sampling where it lives.  Then decide how much "better" is worth the investment to actually do work with the better methodology.  

 

Premiere and After Effects show a slight difference in their handling of the AVCHD footage.  My guess is because you engage Premiere's precision as a render setting, meaning any interactive color correction is only semi WYSIWYG, unlike After Effects which has a color managed viewport.  Premiere does no chroma filtering of AVCHD to the viewer.  It might when you crank up the rendering precision and be somewhat closer to After Effects quality.  

 

After Effects does seem to do some form of chroma filtering to AVCHD footage when dropped into a 32bit project file.  It just doesn't do as good a job as 5DtoRGB.  Unlike Premiere, however, now I figure I can manually filter the chroma on the AVCD in After Effects and be as good as the 5DtoRGB filtering and skip the transcode if I'm finishing in After Effects, which is generally the case.