Alright,

Let me ask you guys something that I am kind of struggling with; last night just playing in my mac I decided to take one of my captured clips from the A1 - 24f and manipulate it, I ran it thru MPEGSTREAMCLIP and converted to Uncompressed 10bit 4.2.2, and heavily color correct it and apply some "bleach bypass" to it. I did the same with the regular HDV file

Amaizinly enough I ended up with a better file output from the Uncompressed file (IT TOOK THE ABUSE) while the HDV file looked like it couldn't handle it as well.

Help me understand this, 25mbps is 25mbps as the source, even converting the file to Uncompressed which ended up being HUGE, the color space is what it is, how come it looks better and took so much abuse better than the other I am crazy here?

Sounds interesting - could you post a clip of each for viewing? If you need server space, let me know. Cheers!

I am in the office at the moment, but when I get home I'll see if I can.

Here is my uneducated and speculative thought about what is going on:

Once you output to a file that fixes what each frame is, you can manipulate from there without the HDV codec interpolation coming into play. When you play or edit an hdv file, it is relying on other frames to arrive at the the "right" pricture based on the codec "rules". This is what Cineform and other intermediate editing systems are about, providing frame accurate editing capabilities.

Here is my uneducated and speculative thought about what is going on:

Once you output to a file that fixes what each frame is, you can manipulate from there without the HDV codec interpolation coming into play. When you play or edit an hdv file, it is relying on other frames to arrive at the the "right" pricture based on the codec "rules". This is what Cineform and other intermediate editing systems are about, providing frame accurate editing capabilities.


Interesting.......that is actually what I was working on as my theory. So I am not alone on this concept.

PS: Man, if we are right this is actually nice, but uncompressed takes A LOOOOOOTTTT of space, need to try with DVCPRO HD

DVCproHD is another compressed codec that also down-rez'es the footage. Try SheerVideo or other lossless codec. Have you tried the Apple intermediate? How does it hold up?

I would guess that taking any compressed format to an uncompressed format would make colour correction easier.

I know for a DV shot project we did, we did an online in Avid through 10bit uncompressed, spat this out as Digibeta, took it to a grading facility to grade on Baselight to get our digibeta master. What we ended up with on the Digibeta looks amazing. When we dub it back down to MiniDV it still looks amazing, but not as amazing as the digibeta.

I'm pretty sure if we had just been mucking around with DV the colour correction wouldn't looked nearly as nice, because while you can't pull information back that isn't there, you can add information in that you couldn't add into your compressed format.

I really don't have any technical background at all, so this is all just guess work, but from a practical point of view I would say getting your footage to an uncompressed state before doing any colour grading is always going to yield better results.

Whether those results show up when you go back down to your distribution format is the more interesting question - but I reckon the idea that you are never going to improve on your original shooting format no matter what you post on has crept in when it really isn't true.

You may not be able to get more resolution, but surely when colour correcting you are able to add information that isn't there when you are shooting because your compression system discarded it? Just because you can't find what was meant to be there, doesn't mean that your grading system (if beefy enough) won't fill those gaps with something else if you tell it to, right?

I'm guessing the improvement you're seeing is a result of working in a 10bit format rather than 8 - take a look at Stu Maschwitz's 'DV Rebels Guide' (www.rebelsguide.com). He discusses the advantages of working in a higher bit-depth format than your source material when color correcting/grading.

While an uncompressed source certainly can speed things up by taking the decompression load off the processor, I don't think having it in an uncompressed source format makes much difference quality-wise. Regardless of the source, any modern video app is working internally with a 4:4:4 uncompressed image. Some plugins/effects only use 8bit internal processing though, while others go as high as 32bit - and some applications do render back to the source format which can degrade your final output. Try taking the HDV source into AE for your color correction, but switch to 16bpc mode, and render to an uncompressed file - I'll bet you'll see similar results to your 10bit source test, even if you go to an 8bit master.

I notice this as well when I work with 24F footage from my H1. I transcode the .M2T to Apple's 10-bit Uncompressed Codec and use the footage in my online process. There is a noticeable increase in quality of the footage for color grading. I also use the Nattress Chroma Smooth and Sharpen filters before adding my cc, which make's the footage look even better and it grades easier ...

I think the combination of MPEG streamclip along with the Nattress chroma plugins and Colorista truly make HDV footage look a lot higher end. I'm quite pleased ...