5D Mk II - Raw Video Conversion Options? at DVinfo.net

Billy Griffin · February 22nd, 2010, 02:36 PM

Okay, novice needs help from you guys/gals.....

So I shoot the raw video on my 5D Mk II and then upload it to my computer. I am using Sony Vegas 9.0 as my NLE software. Obviously the RAW video is not user-friendly when dumped into Vegas 9.0. So thus comes the need to convert the RAW file to an .AVI file, etc. etc. Basically, if I understand it correctly, all NeoScene does is convert the RAW .mov file from the Canon codec into a user-friendly .AVI file for better playback and stability in your editing software, (???)

My question is: what advantage does using NeoScene to convert the raw footage to .AVI have, versus simply dragging the entire raw file into Vegas, then selecting what type of file I want to reder it as (AVI, WMV, MOV, etc.) and also selecting the frame rate I want to 'convert' the video to, and then once that file is created, drag it (the newly rendered file) back into Vegas for proper viewing / playback / editing ?

I hope this makes sense, but I'm trying to learn the justification for purchasing NeoScene to convert my raw Canon 5D files to a .AVI file, versus just saving the $ and doing the above-type conversion using Sony Vegas.

Is it a time, quality, functionality issue? Please explain. Newbie hoping to learn!

Matthew Roddy · February 22nd, 2010, 05:55 PM

I use NeoScene, and here's why:

1) Conversion is easy

2) Editing is infinately easier using Cineform files

3) The codec (Cineform) is more robust, allowing for a great deal more processing (see web site for more details about how deep you can go); ie: more CC, compositing, etc.

If you go to their web site, you'll get many more reasons and facts and figures to back it all up, but the above is from an end-user's POV.

Hope that helps a bit.

Jon Fairhurst · February 22nd, 2010, 06:02 PM

Hi Billy,

When you drag a RAW file to the Vegas timeline, it uses the Quicktime decoder to work with it. Unfortunately, QT applies some gamma conversion to the signal that screws up the levels. At least the version that I last tried did that.

The poor man's solution is to open each clip manually in QT and re-wrap it as an M4V file. You can drag that directly into Vegas, and the results are good.

Unfortunately, there are still some problems that can be fixed in Vegas:
* Manually set the height of each clip to 1080. It erroneously says 1088, creating a black line at the bottom and skipping some lines in the image.
* Add a filter that converts Computer (0-255) to Studio (16-235) RGB.
* Slow the video to 0.999 the rate to convert from 30p to 29.97p.
* Slow the audio to match.

And, unfortunately, the files still don't playback efficiently for editing due to long GOP. You can render the clips in lower res and use these as proxies, if you'd like. Vegas is kind of lame with managing proxies though...

Or you could just buy Cineform. Convert each file once in a batch process. It's roughly a real-time operation with a quad core machine, so it's no worse than capturing DV tapes, time wise. Cineform applies all of the fixes above and gives you a 10-bit file to boot.

It's $100 that you will not regret spending.

Andree Markefors · February 28th, 2010, 01:27 PM

Billy,

Normally RAW is reserved for just that- RAW sensor data. That is something that we currently don't have for our DSLRs just yet.

A better term might be "native 5D" files... or whatever...

Anyway, as you know these are h.264 .MOV files- and that is basically a compressed delivery format.

Nonetheless, compression isn't all that bad at roughly 40 MBit/s.

I'm just mentioning this as a FYI. Might come in handy down the road. Especially if we actually come up with some way to get the RAW data. Maybe live sensor capture via HDMI?

Good luck with your project!

Jon Fairhurst · February 28th, 2010, 04:07 PM

You're right. In my post above, I should have written "native" rather than "RAW".

And, yes, Canon's compression is quite good. The only place where it falls down is on very flat, smooth areas. That's where the block contours can be the most visible. If you have time in post, you can mask those areas and apply a blur filter to smooth things back out.

The encoder is much stronger than Nikon's. When the D90 sees too much detail, like on a gravel driveway or cobblestone street, it often delivers some blocks with detail and others that are smooth. Canon's encoder generally encodes detail very consistently. As long as you're not aliasing, detail encoding is quite good..

February 22nd, 2010, 02:36 PM	#1
Billy Griffin Regular Crew Join Date: Dec 2006 Location: Enumclaw, WA Posts: 58	5D Mk II - Raw Video Conversion Options? Okay, novice needs help from you guys/gals..... So I shoot the raw video on my 5D Mk II and then upload it to my computer. I am using Sony Vegas 9.0 as my NLE software. Obviously the RAW video is not user-friendly when dumped into Vegas 9.0. So thus comes the need to convert the RAW file to an .AVI file, etc. etc. Basically, if I understand it correctly, all NeoScene does is convert the RAW .mov file from the Canon codec into a user-friendly .AVI file for better playback and stability in your editing software, (???) My question is: what advantage does using NeoScene to convert the raw footage to .AVI have, versus simply dragging the entire raw file into Vegas, then selecting what type of file I want to reder it as (AVI, WMV, MOV, etc.) and also selecting the frame rate I want to 'convert' the video to, and then once that file is created, drag it (the newly rendered file) back into Vegas for proper viewing / playback / editing ? I hope this makes sense, but I'm trying to learn the justification for purchasing NeoScene to convert my raw Canon 5D files to a .AVI file, versus just saving the $ and doing the above-type conversion using Sony Vegas. Is it a time, quality, functionality issue? Please explain. Newbie hoping to learn!

February 22nd, 2010, 06:02 PM	#3
Jon Fairhurst Inner Circle Join Date: May 2006 Location: Camas, WA, USA Posts: 5,513	Hi Billy, When you drag a RAW file to the Vegas timeline, it uses the Quicktime decoder to work with it. Unfortunately, QT applies some gamma conversion to the signal that screws up the levels. At least the version that I last tried did that. The poor man's solution is to open each clip manually in QT and re-wrap it as an M4V file. You can drag that directly into Vegas, and the results are good. Unfortunately, there are still some problems that can be fixed in Vegas: * Manually set the height of each clip to 1080. It erroneously says 1088, creating a black line at the bottom and skipping some lines in the image. * Add a filter that converts Computer (0-255) to Studio (16-235) RGB. * Slow the video to 0.999 the rate to convert from 30p to 29.97p. * Slow the audio to match. And, unfortunately, the files still don't playback efficiently for editing due to long GOP. You can render the clips in lower res and use these as proxies, if you'd like. Vegas is kind of lame with managing proxies though... Or you could just buy Cineform. Convert each file once in a batch process. It's roughly a real-time operation with a quad core machine, so it's no worse than capturing DV tapes, time wise. Cineform applies all of the fixes above and gives you a 10-bit file to boot. It's $100 that you will not regret spending. __________________ Jon Fairhurst

February 28th, 2010, 04:07 PM	#5
Jon Fairhurst Inner Circle Join Date: May 2006 Location: Camas, WA, USA Posts: 5,513	You're right. In my post above, I should have written "native" rather than "RAW". And, yes, Canon's compression is quite good. The only place where it falls down is on very flat, smooth areas. That's where the block contours can be the most visible. If you have time in post, you can mask those areas and apply a blur filter to smooth things back out. The encoder is much stronger than Nikon's. When the D90 sees too much detail, like on a gravel driveway or cobblestone street, it often delivers some blocks with detail and others that are smooth. Canon's encoder generally encodes detail very consistently. As long as you're not aliasing, detail encoding is quite good.. __________________ Jon Fairhurst

February 22nd, 2010, 05:55 PM	#2
Matthew Roddy Major Player Join Date: Jan 2006 Location: Riverside, Ca Posts: 307	I use NeoScene, and here's why: 1) Conversion is easy 2) Editing is infinately easier using Cineform files 3) The codec (Cineform) is more robust, allowing for a great deal more processing (see web site for more details about how deep you can go); ie: more CC, compositing, etc. If you go to their web site, you'll get many more reasons and facts and figures to back it all up, but the above is from an end-user's POV. Hope that helps a bit.

February 28th, 2010, 01:27 PM	#4
Andree Markefors Regular Crew Join Date: Sep 2009 Location: Zürich Switzerland Posts: 81	Billy, Normally RAW is reserved for just that- RAW sensor data. That is something that we currently don't have for our DSLRs just yet. A better term might be "native 5D" files... or whatever... Anyway, as you know these are h.264 .MOV files- and that is basically a compressed delivery format. Nonetheless, compression isn't all that bad at roughly 40 MBit/s. I'm just mentioning this as a FYI. Might come in handy down the road. Especially if we actually come up with some way to get the RAW data. Maybe live sensor capture via HDMI? Good luck with your project!