What's the "best" video format to use in Hitfilm Pro

I am using the Canon EOS 5D Mark III.  It outputs MOV Quicktime.  Is that a "good" format to use in Hitfilm Pro?  I read that Quicktime is a 32 bit format.   My computer is 64 bit windows.

So overall, what is the "best" format to use in Hitfilm Pro?  For fastest editing and rendering.  In editing, 3D and VFX  mode.

Thanks

Rowby

Comments

  • Triem23Triem23 Moderator
    edited April 2015

    @Rowby I am going to link another thread here. Scroll down to my very long post, which also links to an outside article on editing vs delivery codecs. 

    Short answer, DNxHD. In the other thread I discuss this, and do a quick run down on conversion of h.264 MOV to DNxHD AVI.

    hitfilm.com/forum/discussion/6132/stupid-technology-questions#latest

  • Thanks!  That link is great.  ...Rowby

  • hi rowby,

    actually, the 32bit image format is not related to the 64bit"ness" of your windows computer... also, what people mean by XXbit color is not consistent either. :P

    hang on to your hat:

    - computers operate on bits of data. a single bit is like an on/off switch with only those two states, on or off... we refer to the on and off as 0 or 1. at the most base level, computing takes place purely with 0s and 1s and we call that binary code.

    - a single bit can only represent two "numbers"... 0 and 1. in order to represent numbers higher than that, you need more bits.  so  a 1 bit computer can represent 2 distinct numbers and/or count to 2. a 2 bit computer can count to 4 by representing those numbers like this: 00, 01, 10, 11... to represent a new number requires a new combination of 0s and 1s (which we established previously are the only things that a computer really knows about). a 3 bit computer can count to 8: 000, 001, 010, 011, 100, 101, 110, 111...

    - basically, the rule is when someone says, XXbit, they mean 2^XX (the ^ means "to the power of" or exponent).  so as we've seen, 2^1=2, 2^2=4, 2^3=8, etc.

    - when someone says that a computer is XXbit, they're referring to the largest numbers that the computer can work with at a time and with perhaps its greatest impact in the largest amount of memory the computer can use at once... every memory location has to have an "address" that the computer can look up or pass on... if you can only count to 4, you can't possibly tell anybody about addresses that extend past 4.

    - the prime advantage in our modern operating systems being 64bit is that we can address and use more RAM. previously we were limited to 8gb of ram... now we can have much much more (if we can afford it... that stuff is still relatively pricey).

    =======================================

    now all of that was for computers that are 8bit or 16bit or 32bit or 64bit.

    this has NO RELATION to image and video formats that are referred to as being 8bit, 16bit, etc. except for the fact that we are still referring to the amount of numbers we can use to describe something (in video's case, COLOR).

    =========================================

    when we say an image or video is 32bit, we're talking about the amount of colors that can be stored in the format.

    so to start from the ground floor again, a 1 bit image format would be able to call a pixel black or white.  that's it. no gradations of any kind.

    a 2bit image could encode 4 shades of gray.  a 3 bit image could encode 8 shades of gray. etc. as you can see, the more bits you have, the more smooth your gradations between colors can be.

    an 8bit image can encode 256 colors.  BUT... usually, when people say that an image is 8bit, they are saying that the image is 8bits PER CHANNEL.

    for color images, we usually (but not always) have 3 channels: Red, Green, Blue. and 8bits per channel means that we have 256 different intensities for EACH CHANNEL. a more accurate way to call that kind of image is 24bit (8+8+8) which is capable (by mixing the r,g,b with 256 intensities each) of showing 2^24=16,777,216 colors.

    for the longest time, this was basically "true color" for video displays. they had lesser technologies like EGA and CGA which had a far more limited color palette once we hit 8bits per channel, we stagnated there for a while.

    you might be fooled into thinking that 32bit color was the next step up in color representation but 32bit color simply refers to the addition of an additional 8bit channel that encodes for how transparent a pixel is called an "alpha channel". so instead of just RGB, it's RGBA.  same colors but now you can just make them have 256 levels of transparency/opaqueness.

    the REAL next step up is 16bit color. this used to mean RGBA at 4bits each but nobody uses that kind of restricted color anymore, least of all for movies.

    nowadays, we mean 16bits PER CHANNEL again... so R=16bit, G=16bit, B=16bit. another way to call this scheme is 48bit color. (told you it was confusing).

    now we're in the realm of "high dynamic range". 8bits per channel is good but we still see more colors with our eyes in real life. 16bits per channel gets us closer.

    if only our monitors were not almost all universally 8bits per channel!

    but within programs like hitfilm, having the extra unviewable colors helps in the blending of elements and effects and is less likely to incur artifacts like "BANDING" (where you can see distinct stair stepping of colors in what's supposed to be smooth gradient).

    =========================

    tldr:

    1. xxbit color has no relation to xxbit computing

    2. when people tell you that a video is xxbit, ask them if they mean xxbit or xxbit PER CHANNEL.

  • Triem23Triem23 Moderator
    edited April 2015

    @jinchounggmailcom the other thread I linked Rowby to, above, talked briefly about 32 vs 64 bit memory addressing (which is what he meant when discussing "32-bit QuickTime," but you just gave a great walkthrough of color depth there. Very well written and put together. :-) 

    Just to quickly add to what you wrote, of course some HDRI programs calculate in 32-bpcc color, but nothing can display that, it's a bit of overkill, and for Hitfilm/video purposes we can pretty much ignore it for now. :-) 

  • thanks much! good info on your link as well.

  • Aladdin4dAladdin4d Moderator

    @jinchoung Great breakdown! This thread is now totally bookmarked

  • I think DNxHD is the correct video format for Hitfilm. That's the reason why I need to encode Canon MOV into DNxHD via a professional tool.

  • why cant I follow the link! =(

  • edited April 2016

    @Dezzy

    At the very top of the post, next to the post title, right under the word menu, there is a star. Click on the star it will turn gold and you will get a notification every time someone adds to this link.

     Cheers

  • Triem23 posted where he talks about AVxHD when I click on the link, it just redirects me back to the comments section of this thread and I currently need to convert MTS to AVxHD because I have the express version and cannot edit dolby audio, and the camera I used records in dolby audio, so I need to convert everything I filmed into a different format, or at least advice on how to conver my audio format

  • Triem23Triem23 Moderator

    Hmmmm. That link should take you to a different thread (it does for me) where the relevant information is hidden in the posts.

    Fortunately we have cut and paste. ;-)

    This link take you to another forum thread with full instructions on how to convert footage using MPEG Streamclip: hitfilm.com/forum/discussion/6091/step-by-step-on-how-to-do-render-to-an-intermediate-codec-then-work-with-it-in-hf#latest

    Also:

    Download DnxHD here: http://www.avid.com/US/industries/workflow/DNxHD-Codec (it's free!)

    Download MPEG Streamclip here: http://www.squared5.com/ (it's free!)

    In MPEG Streamclip, you'll load in your files, tehn you'll need to go to export settings. Select Export to AVI, then, in the "Compresson" drop-down menu, look for "AVID DNxHD" as shown here: https://flic.kr/p/pMd3nF

    Then, click the "Options" button to open up the DnxHD settings and use the resolution drop-down to select the proper resolution and desired data rate as shown here: https://flic.kr/p/qrqCSC

    Or this tutorial by Simon Jones covers transcoding footage in Handbrake: https://www.youtube.com/watch?v=kmBmkIrekBk

    While this tutorial has supplemental information on optimized h.264 settings that work faster in hitfilm: https://www.youtube.com/watch?v=4wWI_9Sscjc

  • Hey im trying to convert my M2TS files to AVI files using Streamclip but when i start converting it says it cant find the first frame and the video wont play. Any advice?

  • Try using Handbrake instead. I don't think MPEG Streamclip supports MTS/M2TS footage.

  • I'd installed the QuickTime Alternative 1.81 and launch MPEG Streamclip 1.2 on my Windows 10 PC. After that I installed the AVID codec (that Triem23 linked) I relaunch the application, but the avid coded wasn't listed. What's wrong actually?

  • @iFang for DNxHD you select Quicktime MOV as the export, then look for either "Codec" or "Compression" (on phone now, so can't check) and select the DNxHD settings there. 

  • @Triem23 Do you mean the one inside the option like this?

     

  • @iFang right menu, wrong codec. DNxHD is two above the one you have highlighted. 

  • edited February 2017

    @Triem23 Thanks. Anyway, ueould like to ask is there someway to transcode the video files into ProRes?

  • @iFang In the transcode thread I provided some ffmpeg window scripts to do various "intermediate" codec transcodes including Prores. Others like DNxHD and DNxHR. Of course there is also the fast decode AVC listed in that thread as well.

     https://hitfilm.com/forum/discussion/42349/transcoding-for-better-performance-and-easier-editing

Sign in to comment