Home

Content

Shiv Kumar

Hobbyist Filmmaker / Editor

Adobe Premiere Pro CS3 and Canon 5D Mark II / Canon 7D footage	0votes
October 28, 2009 01:28 PM	Views: 7041 Favorited: 0 Comments: 12

	Filed Under: Post Production, Videography
Tags: Bitrate, Canon 5d Mark II, Canon 7D, Uncompress

The Canon 5D Mark II records HD video as compressed H.264 at a video bit rate of about 40Mbps. The audio is pcm 16bit at 1411kbps (that is CD quality audio).

However, when attempting to edit footage in any NLE, especially if you need to be able to scrub the time line you won't have a good experience. Mind you, if your computer hardware specs are really good (I mean really good), you may not have any issue, but for those whose hardware specs are (just) good enough for HD video editing, you'll need to uncompress the video your Canon 5D Mark II produces.

There are many ways in which you can uncompress your footage. This tutorial is about using Adobe Premiere Pro CS3/CS4 to do this.

Source File

I started out with a source file (straight from the camera) that had the following specs:

1. File Size: 166MB

2. Duration: 34 seconds

3. Video bit rate: 39612kbps (~40Mbps)

4. Audio bit rate: 1411kbps

Here are the steps (Uncompressed AVI):

Create a new project with using the HDV1080p 30 preset. You could use any appropriate preset really. For example you could use the XDCAM EX 1080p 30 preset as well.
Import the original .mov file into the project
Place it on the Sequence 01 timeline.
With your Sequence 01 selected, use the File menu and choose Export - Movie. Make sure you select Export - Movie and not Adobe Media Encoder or any other option.
In the dialog presented, click on the "Settings..." button. You will then be presented with another dialog window.

From this point make sure you choose the options shown in the images below. I've shown to basic options. Uncompressing to AVI and uncompressing to MOV.

Make sure you select, check, uncheck (as the case may be) all options as shown in the images below. A few things have changed in each screen (not just one setting).

Uncompressing to AVI

The process took 2:11 on my machine for a 34 second source clip.

The uncompressed avi file (steps below) has the following specs:

1. File Size: 3.93GB

2. Video bit rate: 994363kbps (~994Mbps)

3. Audio bit rate: 1411kbps

Figure 1

Figure 2

Figure 3

Figure 4

Uncompress to MOV

Follow the same steps as those for AVI (above). The only difference is in the setting you choose when presented with the dialog that show up after clicking on the "Settings..." button.

Instead of the settings shown in Figures 1 & 2 above, choose the settings shown in Figures 5 & 6 below. and then choose the same settings as shown in Figures 3 & 4 above.

The process took 2:20 on my machine for a 34 second source clip.

The uncompressed mov file (steps below) has the following specs:

1. File Size: 3.93GB

2. Video bit rate: 993376kbps (~993Mbps)

3. Audio bit rate: 1410kbps

Figure 5

Figure 6

You'll notice that in Figure 6 above we selected Uncompressed YUV 8bit 4:2:2, you can also choose the 10bit option. the specs of the file size thus are as follows:

1. File Size: 5.2GB

2. Video bit rate: 313921kbps (~314Mbps)

3. Audio bit rate: 1410kbps

So the file size is much larger than the 8bit version but the video bit rate is way less. I'm not sure why. either ways, I'd personally use the 8 bit version as I don't believe the source is 10bit so there is nothing to gain by using 10bit.

Other Formats

You can choose other formats that are uncompressed as well. The Figure 7 below shows you the other options available. Of those, the Targa and TIFF formats are uncompressed as well.

Figure 7

Comments

October 29, 2009 07:07 AM

Shiv, I'm not sure I fully agree with you on this one. Maybe it's a different terminology, but to my mind this is not 'uncompressing' the original compressed footage.

Is it not just converting the files into an intermediate file format - that happens to be uncompressed?

I also think that the file sizes you would generate from this procedure would be unnecessarily huge and I think my system would probably play the raw footage easier than an uncompressed file.

There are lots of alternatives out there for intermediate file formats which would make a lot more sense to use. A lossless codec would be a better way to go - or use one of the many converters like Cineform Neo Scene or Voltaic etc.

Shiv Kumar

October 29, 2009 08:20 AM

I’ll try and explain technically what is going on. Imagine you’re playing a video. This video is compressed (for argument sake, let’s say it’s compressed like the Canon 5D footage) at 40Mbps. What the video player is doing is decompressing the video at 30 frames per second and playing it back. That is the video needs to be decompressed before it can be played back and it is being decompressed at 30 frames per second for 30 fps footage (and that’s a lot of processing -I’ll get to this part later when I talk about the NLE).

So any video (HD or SD) is compressed. For example HDV footage is compressed at 25Mbps (Canon XH A1 footage for example), XDCAM HD in HQ mode is compressed at 35Mbps (Sony EX1/3 for example). This is what we consider "raw" footage. But it's not "raw" in technical terms, and in fact is compressed.

The process I describe is really converting the footage to "raw" in the sense that it decompresses, (just like the video player would have to decompress before playing back) and leaves it alone. That is it saves the decompressed video to file. If you were to encode then it would re-compress as per your settings. But in this case we want to leave it as uncompressed (or decompressed), which is why I call it uncompressed. I guess I could have called it decompress.

Now, in your NLE, when you scrub the time line, your NLE has to decompress every frame while you’re scrubbing the timeline. But in this case it has to decompress way more than 30 frames per second! That is just not something most computers can do today. Mind you, you’ll still get a semblance of scrubbing but if you were to compare scrubbing the footage that came out of your camera versus the uncompressed footage, you’ll see a marked difference. So much so, that you can actually edit the footage on a less capable computer. The NLE knows when to attempt decompressing and when not to, given certain footage and in this case it knows not to decompress and so the uncompressed footage scrubs a whole lot better and faster without skipping frames.

Yes, uncompressed footage will be much larger which is why cameras compress the footage in the first place so they can record the footage to some portable media (tape, CF cards, SxS cards etc.). Personally, I wouldn’t maintain the uncompressed footage beyond a certain period. I’ll keep the original compressed footage.

In PPro CS3, you have the option to “render” your timeline. When you do that using the footage straight out of the camera you can scrub the timeline normally. But as soon as you change something in your timeline, the render is rendered useless and you’d have to render again in order to get a decent scrubbing experience. With the uncompressed footage you don’t need to render. It just works and so the complete editing experience is much better.

I don’t have PPro CS4 so I can’t be certain, but PPro CS4 has the notion of proxy files. It’s possible that it does something different and that the footage straight out of the camera may work. Further, PPro CS4 utilizes the GPU (embarrassingly parallelized) to encode/decode so it’s possible that the experience is different, but I seriously doubt it, because the process of compression/decompression is extremely compute intensive as well as Hard disk I/O intensive and as a result typical HD video editing hardware requires really good disk I/O (50-60 Mbps) as well as multiple cores and tons of memory to support the cores.

Finally, yes there are other alternatives (software and/or hardware) but this post was about using the tools you already have. For those of us who don’t do video editing for a living it becomes difficult to justify additional expenses. I should mention also that "lossless" codecs won’t solve this problem because they are still codecs, meaning COmpress-DECompress, so the process is not eliminated.

October 29, 2009 10:25 AM

He he. That's a pretty detailed explanation, thank you. Let's not split hairs - but technically, CODEC is a contraction of COde-DECode and doesn't have to involve compression. It has evolved somewhat in terms of its meaning.

All that aside, in purely practical terms I would still argue against using uncompressed. There are an infinite number of intermediate codecs that will perform as well, probably better and take up a hell of a lot less space.

But each to his own!

RUBEN SENDEREY

October 30, 2009 11:52 AM

Hi,
we try your way,
but the file was still strobeing on the time line after we made the movie to avi. file on CS3,
any thoughts

Shiv Kumar

October 30, 2009 11:56 AM

Ruben,

What kinds of files don't strobe on your machine? I'm trying to get a sense of your computer's capabilities. Do you by any chance know the Hard disk configuration of your computer. I don't need the capacity but rather the I/O capabilities of your hard drive sub-system.

What Operating system? What is the resulting file size? What is the original file's size?

October 31, 2009 05:07 AM

It is most likely that your Hard Drive is not fast enough to handle the high rates you need to edit uncompressed material.

You can download a free utility to check the speed of your drive (the one you use for video editing) - Something like HD Tune has a free demo.

Enrico Perei

November 12, 2009 06:05 AM

hi,

I stumbled upon this Post looking for facts about 7D's HDMI output. Mainly if it is capable of outputting an uncompressed 1920x1080 format.
I believe so far it is not possible. However there seem to be some workarounds with downsized formats through this connection...I'll keep looking (;

Anyways, I found your Post to be a good tutorial explaining your workflow.

Whenever possible I am an advocate of uncompressed footage.
it holds the best quality and will be best suitable for image manipulations, colour grading, keying and other vfx work.
in professional equipment the pricetag has a lot to do with the fact that we need basically "fast everything" in order to work uncompressed.
budget,purpose and growing image sizes however request usually compromises.

Your workflow makes a lot of sense and is in fact used often but I'd like to point some things out and propose an alternative as well.
Also to make sure what my terminology is to describe things I have a short explanation at the end of my post.
Unfortunately due to the ever growing techniques and standards the meanings of many seem to shift, change and lead astray, which can cause misunderstandings during conversations even among veteran professionals.

for those who have struggling equipment I suggest to remember what we did before the machines became capable of SD,HD or even 2k/4k uncompressed RT playback.
do so called Offline editing by using proxy files.
you can convert your footage using higher compression or lower resolution or both.
this way you will be able to edit much easier, have smaller files to move around and so on.
after you are happy with your cut you do the so called online, where usually the EDL (your timeline) is rendered using the original footage instead of the low res proxies.

this is however extremely important to keep in mind !:
any effect and prerender you are doing should be made using the original footage and proxied to insert into the proxy timeline.
the final output needs to be rendered using the original footage.

this way you also will have only one additional generation of your footage minimizing quality loss due to recompression.
it is important to also remember that if you are doing pre renders for effects etc. to cut into your edit you will decrease the quality of that portion with every render (unless you render from and into a lossless format).
while you can and should do prerenders to increase responsiveness for smooth playback of your timeline, the final render should not use prerendered footage but should render those effects during the final rendering.

using this old technique you would be in theory able to edit your 4k blockbuster movie on your iphone. all you would do afterwards is to bring your edl into the online and let the machines crunch the numbers.....this is of course not the way it usually is done but it shows the potential.
in seriousness though, you will be able to edit on your low spec notebook any high resolution show.

also problems with editing for instance mpeg footage arise from the fact that the compression spans over several frames (GOPs)
you are only able to jump to a frame inside such a structure if the software decompresses the previous frame(s).
which means that instead of using an uncompressed format or a proxy with this kind of compression structure you should use a frame based compression.
this won't be as efficient but you will be able to jump to each frame independently and quicker.
another possible way is to export your footage as image sequences. provided your editing system can read those and maintain RT playback.
again preferrably lossless formats like .dpx,.cin,.tga etc.

most professional editing systems have their own proprietary file structure which accounts for most of the aforementioned issues.
they often have hardware designed to speed up compression processes as well.

uncompressed online editing:
pro:
if lossless workflow is maintained until final output,there is no further quality loss
less processing time is needed to playback, it is more an I/O issue
straight forward understandable workflow, hardly room for errors

contra
large to huge file sizes
fast hard drive or raid needed
longer times to render wipes/blendes etc. or RT capable hardware needed depending on used resolution
decompression(conversion) time required
(re)compressing into final format introduces second generation loss

online editing using source footage:
pro
no conversion time needed
generation loss minimized,(re)compressing into final format introduces first generation loss

contra
large to huge file sizes
fast harddrives/raid needed
fast processor and/or graphics card needed
possible compression assistant hardware speeds up responsiveness but at a budgetary cost
longer times to render wipes/blendes etc. or RT capable hardware needed depending on used resolution

offline editing
pro
smaller files, quicker software and system response
low hardware requirements
processing relies more on processor speed than I/O
generation loss minimized,(re)compressing into final format introduces first generation loss

contra
proxy generation(conversion) time needed
low quality to work at
errors during final output possible by using proxy instead of source material
certain effects need to be judged using full res footage

what is what:
uncompressed:
uncompressed means per se any file that has not been compressed (no pun intended hehe)

compression comes in two flavours. (mathematically) lossless and lossy.
lossless compression is for instance used in .zip and similar compressed files.
put simply the overall filesize is reduced by representing repeating byte patterns or patterns
identifieable by algorithms with less bytes than in the source file.
the integrity of all bytes of the source is guaranteed and the decompressed file is a 1:1 replica of the source.
.tga for instance is an image file format that can be used for lossless compression

lossy compression is found in many different image and movie formats (jpeg and mpeg being just two).
there are different approaches in compression, suffice it to say that usually algorithms are used to somewhat average certain informations in images.
this way lots of information can be replaced by data that takes up less bytes.
understandably this approach, while often with stunning good results in image quality, has the drawback of loosing original information.
a decompressed file is NOT a 1:1 copy of the source but rather an estimation.
the accuracy of said estimation depends on the quality settings and compression strength/algorithms.
it is to note that every time lossy files are decompressed and compressed again (even with the same settings) you loose more and more information.
(generation loss, very similar to that of creating a copy of a VHS tape that itself was a copy of a copy of the master....each of those copies is degraded in quality with respect to the previous)

raw:
raw footage is most commonly referred to in conjunction with digital cinematography and photography
a raw file can contain many things depending on the vendor.
ideally it should be the most basic least processed data coming from a chip in digital cameras.
raw files,again depending on vendor,can be debayered and further processed for sharpness etc. or be the absolute unprocessed information from the chip (rare)

EDL
edit decision list
this file contains the used footage of a cut, it's timecodes, effects applied etc.

RT
Real Time
used to describe the capabilities of a system to playback the final output at its desired framerate and resolution.

VHS
Video Home System
Before digital technology videocassettes using this format were the first choice for consumer oriented film recording and distribution.
if you know this system chances are you have been born about 15-20 years ago (;

Frame formats:
SD
standard definition (PAL 720x576, NTSC 640x480)

HD
1920x1080 (often referrred to as full HD)

2k
2048x1556

4k
4096×3072

debayering:

an image sensor has filter elements on each pixel which lets only red, green or blue light pass through and registered by that pixel.

the distribution of these filters generates image data comprised of a grid with the respective colours.

in order to generate a single image pixel with its colour information most commonly 2 green,1 red and 1 blue pixel are used.

I hope this helps and I didn'd confuse too much.
hopefully I didn't repeat too many things you guys already knew (;....
enjoy!
enrico perei

Shiv Kumar

November 12, 2009 11:07 AM

Thanks for your valuable input Enrico!

hannah huddy

November 30, 2009 08:35 AM

where is my HDV1080p 30 preset?

Hello I have premiere pro cs3. When I want to start a new project I have a folder for HDV but no options in it. I can't seem to find a plug in for it. What I am doing wrong or need to do to have HDV options for my time line?

Cheers

Hannah

Shon T

November 13, 2010 11:04 PM

I tried Shiv's technique and still had to render the file once it was imported into CS3. I don't think I missed a step...

Dominic Hunter

January 30, 2012 11:10 AM

way to go.

tnx shiv. So this is what is happening. personally, i compile all my videos(10-15video clip) into one video just so to aovid the "error" issues of CS3. My "compiling" settings are uncompressed microsoft AVI,8 bit. im quessing it's the same to your settings. and after i compile all the videos, is start editing. and yeah, i agree with you that i dont really have to render. God knows how annoying rendering is with a limited resources(shitty computer) and yeah the files are much larger.So Shiv, after editing all the "uncompressed" video, how would you or what setting do you use to compress again the resulted Video that can be played back? stupid me, i've exported the files with the same settings(uncompressed) and the result video cannot be played and is bothersome to store(100GB). Please help, i badly needed a mentor (being a young newbie) and i think it HAS to be you.

Shiv Kumar

February 10, 2012 10:27 AM

Dominic, Your export settings depend on what your target is. If it is for websites such as ExposureRoom use the settings we recommend here http://exposureroom.com/help/view.aspx/89/