-
Converting from .MOV to ProRes 422?
Out of complete curiosity, what is the point of converting my files from .MOV to ProRes 422? How will that help my workflow/final quality/anything?
I’m using Premiere Pro CS6 and am shooting with a 5D Mark III
Thanks!
-
13 Replies
-
That’s the suggestion when working with FCP. PPRo…entirely different. Working with H.264 is fine…as long as you have a Graphics card that can enable the Mercury Engine, and a decent amount of RAM. Then editing native is a breeze.
Shane
Little Frog Post
Read my blog, Little Frog in High Def -
Converting the H264 files to Cineform would give better quality results with compositing or color grading with the greater color space and subsampling of Cineform. If Cineform is not an option then I would avoid converting to Pro Res unless you require it for editing.
Eric-ADK
Tech Manager -
[EricBowen] “If Cineform is not an option then I would avoid converting to Pro Res”
I’m not really certain I follow this logic…
As much of a CineForm fan as I am…and I am, I don’t understand how the difference between transcoding to one 4:2:2 10 bit format over another 4:2:2 10 bit format would be that distinct (in this case, most editors wouldn’t push either format to 4:4:4, etc…).
ProRes is a I-frame format and CineForm is an I-P format or an I-frame format (you can set the parameter as you wish) and both formats are very high quality and reduce the decoding load on the CPU over H264 (the files from your DSLR…though the newest Canon and Panasonic DSLRs also have an I-frame H264 codec option).
As far as compositing, etc… The footage arrives with 8 bits of precision…there is an advantage, on paper at least, to putting those 8 bits in a 10 bit pail (ProRes or CineForm), but Premiere Pro decodes and interpolates everything to 4:4:4 32 bit float anyway…so as Shane notes, if you have a system that can handle the DSLR footage natively, and you don’t have another software package in your workflow (Nuke, etc) that has a significant disadvantage to using 8 bit 4:2:0 footage, transcoding may very well not be necessary for you.
All that said, if you aren’t on Mac, ProRes is a non-issue.
TimK,
Director, Consultant
Kolb Productions,Adobe Certified Instructor
-
Cineform is a Wavelet based codec while Pro Res is Mpeg2 based which uses block based compression. That in itself changes the data results when adding images or changing color data.
You can also see features such as what Cineform brings to the table different than your standard 4 2 2 codecs below. I also included Adobe’s blog regarding color sub sampling. There is also Their First Light utility which is extremely nice for LUT changes.
https://cineform.com/compression-subsystem
https://blogs.adobe.com/VideoRoad/2010/06/color_subsampling_or_what_is_4.htmlThe Adobe I-frame codec effects Preview but not the original source data which is what is modified by the compositing or color grading. The greater color space and bit info available to modify increases the accuracy of the resulting encoding with out presenting artifacts. That is why the Space of the source media matters and not just the processing.
Eric-ADK
Tech Manager -
[Eric Ewing] “what is the point of converting my files from .MOV to ProRes 422?”
Hi Eric,
Good question. The standard answer would be that there is no point. Just ingest them natively. You should at least try a test.Transcoding comes into play with underpowered computers, so I don’t think it’s necessary unless your hardware is lacking. ProRes is not the best option for transcoding in CS6, but use it when necessary. I like the suggestion for Cineform. Premiere Pro CC works with ProRes much better now with smart rendering, better performance, etc.
Cheers,
KevinKevin Monahan
Social Support Lead
Adobe After Effects
Adobe Premiere Pro
Adobe Systems, Inc.
Follow Me on Twitter! -
Hi Eric,
[EricBowen] “Pro Res is Mpeg2 based”
Actually…Discrete Cosine Transform is used in many codecs, and does function with the macroblock structure, but MPEG 2 is a specific set of parameters that include using DCT. XDcam and DV variants as well as HDcamSR, DNxHD and ProRes are based on a Discrete Cosine Transform but XDcam variants are all MPEG2 while DV, DVCpro/HD, HDcamSR, DNxHD and ProRes are not MPEG2.
I am a fan of wavelet transforms (and the CineForm codec in general) as their degradation patterns are far less visible to our eye than DCT based codecs, all other things being equal…
I’ve worked with CineForm since Aspect HD was their only product…presented the CineForm workflow in Sony’s booth at DV Expo in New York, and later in Chicago at Sony’s offices along side a Sony Rep as we talked about HD workflow back in the early days. I was the CineForm editor for the 2003 Microsoft/Adobe “HD World Tour” where we met the guys who did ‘Dust to Glory’ who weren’t yet certain how they were going to post their feature. I traveled a bit with the Silicon Imaging camera when it was shooting CineForm raw, and was introduced to SpeedGrade when I was one of the demo guys for how OnSet operated in that ecosystem to create LUTs…and I wrote one of the early manuals for FirstLight…long story short, yes I’m familiar with CineForm.
CineForm is really, REALLY good stuff and while it has advantages over ProRes like being truly cross platform, having a raw variant, FirstLight metadata editor, etc… ProRes is a proven, high quality codec in wide use that has more frame size flexibility than DNxHD, and holds up very well under stress in post production, but to be most useful (encode and decode), is stranded on the Mac.
If you are converting DSLR footage to 10 bit 4:2:2 for editing purposes alone…both codecs are excellent…unless you’re working on Windows…then CineForm is excellent.
TimK,
Director, Consultant
Kolb Productions,Adobe Certified Instructor
-
Thank you for the correction and info. Checking it out now. I am interested in your DNxHD comparison though. Results clients and I have seen both with using differential in AE and Acquisition devices such as the Samurai show far better quality when editing with DNxHD over Pro Res. I know DNxHD is Open Jpeg based which is where the industry is moving at this point with 2K and greater. Why do you see Pro Res holding up better?
Eric-ADK
Tech Manager -
[EricBowen] “Results clients and I have seen both with using differential in AE and Acquisition devices such as the Samurai show far better quality when editing with DNxHD over Pro Res. I know DNxHD is Open Jpeg based”
Well…ProRes and DNxHD have a variety of settings…DNxHD actually has a dizzying variety of settings. I’m always interested in DNxHD-what vs ProRes-what? DNxHD has the largest selection of formats in an MXF wrapper, but many non-Avid devices have the QuickTime wrapped DNxHD codecs…which are a bit more limited…ProRes is definitely more flexible if you limit yourself to QuickTime wrapped files.
What I said in my previous post was that ProRes was more flexible than DNxHD as far as frame size. Media Composer…and therefore DNxHD is limited to HD resolution. At NAB, we were running 4K in PPro CC Windows in the AJA booth, but the vast majority of it was ProRes (we had a few DPX frame sequences…a little Sony 4K raw I think…some Red R3D of course).
Of course, CineForm has been 4K capable for some time…
TimK,
Director, Consultant
Kolb Productions,Adobe Certified Instructor
-
Also pages like this seem to state that the DCT processing creates the Mpeg Iframe for DVCPro HD. Is that accurate? Is that the same with other DCT processed codecs?
https://broadcastengineering.com/infrastructure/mpeg-2-and-h264avc
BTW the Pro Res to DNxHD flavor comparisons such as from the Samurai would be the 220. That had a noticeable quality difference editing to clients with DNxHD as the eventual choice. I agree regarding the spacial options of Pro Res but the CinemaDNG Quality is incredible with trancode to Cineform near completely marginal in differential.
Eric-ADK
Tech Manager -
[EricBowen] “Also pages like this seem to state that the DCT processing creates the Mpeg Iframe for DVCPro HD. Is that accurate? Is that the same with other DCT processed codecs?”
The article you link to is specifically comparing MPEG2 to H264 (MPEG4), and refers to long-GOP compression…
The transform isn’t the data compression really…it’s the quantizing/tossing-out process that decides how (and how much of) the data is going to enter the process of data compression in the first.
DCT is run on video frames for many of the video codecs we’ve been talking about with the exception of codecs that uses a wavelet transform like JPEG2000 (basis of Grass Valley Infinity and REDCODE) or CineForm.
It’s similar to the process of 4:1:1 or 4:2:0 or 4:2:2 color subsampling…that is done prior to the actual compression. Video compression is so effective at reducing the size of video data as much due to what data is simply ash-canned before the process even starts as it is because of the actual “compression” employed.
TimK,
Director, Consultant
Kolb Productions,Adobe Certified Instructor
Reply to this Discussion! Login or Sign Up