Does HDV loses its quality when Color Correcting?
 

I am recording HDV with a Sony V1U, when I upload the footage and I see the plain video in the monitor it looks like HD, you can see the pixels and the quality of the video, however when I add color correction I notice that pristine quality of HD where you can see every detail is lost, the video starts to look like plain DV shot on widescreen... is this common for others? or is it my imagination?

By the way Im using Premiere with a Matrox card, the color correction is a Matrox feature, not Premiere feature.

Thanks

Premiere has a "feature" that adapts the preview image to the load on your cpu to try to keep a real-time playback. (You can turn this off). It is likely that the additional calculations for the color correction are causing Premiere to drop back to a lower res preview. Try rendering the clip first and playing back in high-quality mode and see if that's the problem.../ Battle Vaughan

even after rendering the quality of the video with the color correction doesnt look as pristine as the original HD file without color correction

"Does HDV loses its quality when Color Correcting?"

YES YES YES YES and also YES!!!!

You should transcode to another, higher quality codec. HDV does not store a lot of colour information in its codec, and to start, it is a highly compressed codec. It was designed for capture, not editing.

If all you do is straight cuts and nothing else, then you are ok and you can happily use HDV in its native format - most every app will not need to rerender on "cuts only". But as soon as you do ANYTHING else: fade, fx, or colour correct, the program will need to render and you will lose information(quality).

Take a look at the Cineform NeoScene codec(or similar product). For 99$ you can colour correct, apply fx, render and rereender to your hearts content and you won't notice a difference ...for a few generations as least. With HDV you WILL notice after one gen!!

HI MIKE, so u r telling that the best workfolw to get max quality colors is to trascode in intermediate right? For 3 years i have been editing and color correcting m2t files with Vegas without having this in mind:To trascode with CINEFORM!WHY?
because my hardware can handle hd very quickly and easy and never had particular problems(even with the bullet).
So if now i decide to start using cineform i'll notice better colors?

thx a lot and i appreciate your help

Rich, Try color correcting in Premiere and see if the same thing happens. To be honest, it sounds like a Matrox issue more than an inherent problem w/ CCing HDV.

I've been wondering about how cc actually gets applied to footage in this scenario and if the format of the source footage matters at all. It appears that cc in PP is non destructive. CC filters are applied to a clip as instructions to the program on how to manipulate the clip "data" when rendered for output. The changes then are manifest only in the output file and are subject to the limitations of the output codec. So, if one color corrects an HDV clip then renders it out to h.264 for BluRay shouldn't the limitations be only in the output format?

Now, if an editor's cc tools operate only within the color space of the source format, my argument is invalid, but wouldn't it make sense for a cc tool to operate in a 4:4:4 space or would that involve too much overhead for the conversions.

I raise this because I believe the answer to the OP's question hinges upon how cc really works inside of an editor.

This from Cineform Site: You can download a trial version and test yourself to see what you get for your $ 129. I have used it for 2 years now, I think.


"Chroma Interpolation: 4:2:0 → 4:2:2. Background: Most HDV and AVCHD camcorders record chroma (color) in a format known as 4:2:0. Without boring you with details, 4:2:0 chroma is half the color resolution of more professional 4:2:2 formats. When Neo Scene detects 4:2:0 chroma it properly interpolates the source chroma to 4:2:2 for more accurate color processing during editing and effects work. And if you ever "key" your material, CineForm’s chroma interpolation will substantially improve your resulting visual fidelity."


Here is site:

Cineform NeoScene : Features

Chris, I went to the cineform website and read the information and it actually looks interesting, if it does everything it says it does I think its a good investment, but how does it work specifically?? not even the website explains that.

Is it an individual program? you have to open it and load the footage and then save it in a different format? does it work INSIDE Premiere? do you have to call it once on premiere like an effect, how does it affect the performance on Premiere? I mean the website doesnt give you any information as to how exactly would be the functioning of the program, how is it?

Rich, the Cineform website actually explains how their products work in considerable detail. I suggest you look over the information in their Tech blog. Cineform Tech Blog CineForm Technotes

Essentially, your video is being transcoded to a new editable codec-- whether it comes from HDV or ACHVD, or other recording formats. For HDV, you capture via firewire, pretty much real time. For file based video, its a matter of locating the files, and telling the program to create a Cineform file.

By the way, my understanding is that Cineform was used with Slum Dog Millionaire, the big 2009 Oscar winner.

Chris, so basically it becomes a function inside Premiere? just like Matrox works "inside" Premiere, all the Cineform functions would be performed within the Premiere timeline like one more video effect??

Hey Rich.
The way you worded your question, is sort of open ended.

Basically, the Matrox card uses a proprietary I-Frame MPEG codec, that allows the board to do realtime CC..
As far as CC is concerned, 8 bid HDV is already a compromised structure (since it was designed as a delivery format). Any CC at this level, is rudimentary, and is sort of "mid level".

The advantage of Cineform, is it allows and does calculations with files that have way more colour information than HDV. You can tweak RGB parades to much finer levels than an 8 bit HDV file could ever get to. But unless you're working with some higher end monitoring, and delivering back to a 10bit signal, you probably won't gain much using Cineform vs Matrox.

But as far as your particular problem is concerned, there is something else wrong.
Since all my delivery is SD, I personally use a broadcast monitor, and have my Matrox card do a realtime down-convert for monitoring, via the RGB analogue cabling.

You haven't mentioned how you monitor, and to what format your renders are made..Perhaps a few more specifics, and we can solve your issues...

Rich:

Mostly, you use a utility to capture HDV or transcode files. It is called HDLink, a simple file for selecting the files or communicating with the HDV unit to capture and convert. From there, you open your project and import transcoded files for edit.

You will be able to render out to that codec, too, from within your NLE maintain files for further compiling of a final product. The point of using that codec in that way is that it is nearly lossless tool for success renders.

Chris,

Thank you, I think I understand how it works, let me ask you a question. I presently capture video through a firewire cable that comes out of the Matrox box. The way my system is set up, the only way to capture video thru this firewire is by using Premiere with Matrox.

If I install Cineform, would the program be able to use this same firewire connection to capture? do I have to program it to work that way? is it complicated?

Sorry I ask so many question, the thing is I have absolutely no knowledge of how to install, uninstall or set up programs, I can edit very well but I pay somebody to put the system together and do the entire programming and connections. I dont touch the programming as I can not afford to mess up with it and not being able to work so I am "extremely" careful what I install or how it works or what I could mess up with. If my system stops working I would never start installing or unistalling programs, I would take it to somebody who I'd pay to fix the problem, I would rather not do that.

Thanks

Sorry, I have never used Matrox stuff. About 5 years ago, I was running Premiere 6 with an edit card. Capture was through Premeire and the firewire on that card. When I later stopped using the card, I still used the firewire input for capture. Purely anecdotal, but my thought would be it would work the same way.

An easy way to test that is to download HDVSplit and see if you can control and capture from your through the Matrox firewire set up. Of course, if you are downloading the NeoScene trial, you could check it out that way too.

HDVSplit utility for HDV capturing with scene split - HDV capture utility

Has anyone tested this? I didn't know CF makes your keys better. I have a hard time pulling HDV keys especially in that I'm limited to keys within Sony Vegas. I don't have AE and haven't found any of the aftermarket keyers substantially better than the Vegas built in.

yeah it looks like its going to be too much downloading and testing in doing this, as I explained I dont even touch drive C on my editing system, I might be too neurotic about this but at least Im never complaining about my system crashing or not being able to work. Thank you anyways!

Brian, I have pulled at least one decent key with my 5D footage and NeoScene, in Vegas. I say decent because I am sure there are those who will notice a couple of messups and some green screen reflection due to my poor lighting technique. But within a week of getting camera we shot this little promo using a green painted canvas in my garage. It was easiest pull I ever did:

Nice job Chris. A significant benefit to Cineform is that it converts to 4:2:2 color space which provides more "depth" when color correcting / grading as well as green screen editing. Even if the footage wasn't shot as 4:2:2 video, it helps a great deal to convert it to 4:2:2 before editing because it holds up better when editing.

Rich, i think you are missing the point - all cineform is, is a file type (and it comes with small little program used for capturing, or changing files all ready on your computer)

When you capture from your HDV cam, you normally capture to .m2t files, but with cineform you capture just as you did before, but now the file type will be an .avi file.


Rich,
You can continue to have a rock solid computer with medicore renders - or you can take a small little chance and install cineform and have great quiality once again. BTW they have great support should any problems arise!

THANK YOU everybody for your answers and potential solutions to this problem.

Let me ask you all one more question.... by your answers I think one of the problems is that the codec that HDV use basically doesnt allow you to store too much information so that it loses resolution once you apply the color correction.

One of the solutions that you present is to be able to capture to a more pure HD format, and one of the solutions would be by transforming the file thru Cineform. On another forum they told me that another solution would be to capture thru the SDI HD connection on the Matrox box instead of using firewire as I am using now.

A friend showed me this product a while back but I didnt think of it until now, because basically it is a converter that would allow me to connect thru firewire on one side and export an SDI HD signal thru the other side:

Convergent Design, experts in HDMI, SD, HD, and HDV

Please check out the product and please let me know if you think this would help the problem by capturing a more pure HD signal with the codecs that dont lose so much resolution as the HDV does

Thank you!

I don't undertand how sdi will improve anything, it's still the same amount of data, it's just a better pipe. CF a lot though because it makes each frame a discreet frame and elimates the GOP.

Brian is correct...

Unless you're talking about bypassing the tape, and going straight to disk for uncompressed capture, i don't know what advantages you're getting...
Please be specific in your taping environment and final delivery.
Your opening comment talks about colour correction. Now we're being lead down the path of chroma keying.

If taped correctly, HDV will hold enough information to pull correct colour correction.

I can assume you have the AXIO box, since you mention SDI inputs, correct?

thank you Peter, I dont care much about chroma keying, still my main concern is how "I believe" color correction makes the HDV footage lose resolution when applied.

As mentioned I record on a Sony V1U on miniDV tapes, and load the footage onto the system thru firewire.

The reason why I brought the whole SDI is because one poster (on another forum) suggested that I should try uploading the footage thru the "component" option on the Matrox box, he said that I would be surprised how the component option "could" give me better quality of video than the firewire.

Since he suggested that, I thought... well if component "could" give me better quality than firewire then SDI should give me even better quality than that... that was just a conclusion I was arriving at.

Also I remember that when this person suggested that converter (mentioned on the link) a while back he did mention how uploading thru SDI could give me a better quality than firewire, that person was not a salesperson of the product, he was just a regular production guy who knew about the product.

I believe it's not so much that HDV will lose resolution as it will break down with excessive CC -- which isn't the same as a resolution hit.

If your camera is capable of RECORDING directly from component or SDI that allows for bypassing HDV altogether and recording uncompressed, but usually you need a fast RAID for that. The simple thing is use Pro Res if you're in Mac, Cineform in PC or the freebie Avid codec. Transferring your footage via firewire, sdi or whatever is all the same, only the speed will vary. That guy is wrong. M2t is more of a delivery format than one to edit in, although every generation of NLE seems to improve native m2t support.

Well I believe the advantage of capturing thru SDI instead of firewire is that when capturing on firewire the ONLY format I can capture in is MATROX MPEG-2 I-FRAME. When capturing thru SDI or component Matrox lets me capture the video on 4 different 4 formats incluing MATROX 8 BIT UNCOMPRESSED.

Isnt it better to capture the files on 8-bit uncompressed because Id have more information in the footage so that it would be less destructive when applying Color Correction?

Please let me know, I am just assuming all this, based on everything I have read on this thread.

Thank you.

wait, wait, wait

Very simply....

AFTER you record something - it doesn't matter how you transfer it to the computer - via firewire, SDI, hdmi... its still the same quality. (bit for bit transfer)

the SDI idea only works if you RECORD LIVE to through SDI (you bypass the cameras internal recorder/codec, say HDV, and you record/use a better codec on your computer , say cineform)

capisce?

Mike,

This is the advice I got from a person:

"IF YOU WISH TO USE A BETTER CODEC, YOU CAN TAKE THE COMPONENT HD OUTPUT OF YOUR CAMERA AND CAPTURE IT USING THE MATROX BREAKOUT BOX. THIS WILL CAPTURE TO THE MATROX MPEG-2 I-FRAME HD CODEC, WHICH HAS A DATA RATE UP TO 100MBPS AND ALSO HAS 4:2:2 COLOR SPACE. COMPARE TO HDV AT 25MBPS WITH 4:2:0 COLOR.

I'VE DONE THIS FOR GREEN SCREEN FOOTAGE I NEEDED TO KEY AND DID GET BETTER RESULTS. YOU MIGHT BE THINKING THAT USING ANALOG CAPTURE WOULD LOOK INFERIOR TO 1394, BUT TRY IT AND YOU WILL BE IMPRESSED. NOTE ANALOG CAPTURE DOES NOT WORK FOR PROGRESSIVE, MUST BE 1080I.

JEFF PULERA
SAFE HARBOR COMPUTERS"

Now I SINCERELY appreciate your input on this, but understand that you and this guy Jeff Pulera are 2 unknown people to me, I am not saying you are wrong and he is right or you are right and he is wrong. I read your input on this and I appreciate it very much but I need to hear opinions from MANY different people and MANY different points of view to then have a real idea of all my options. This has been the way that has worked for me for many years and just by looking at this thread I can see many perspectives on this.

And again thank you very much for taking an interest in my question and provide me with an answer.

Mike's right. Jeff, in this case, is wrong. Once it's been compressed to tape it doesn't matter how you capture it; the damage has been done.

Live capture is a whole 'nother story.

It doesn't matter how big your pipe is if there's only a trickle of water. You can pour a gallon of water onto a five gallon bucket, but it's still only a gallon of water.

The better codec gives you more headroom for color correction and other manipulation, but that has nothing to do with the capturing. You can accomplish the same thing with Cineform or any other intermediate codec.

Adam, in this case what you are saying is, no matter how I capture the video from the tape (firewire, SDI, Component) the codec will remain the same on the video, thus it doesnt matter how I capture it, the results when wanting to apply color correction will be the same, correct?

In terms of capture, yes. But you can always then transcode to a better codec which gives you more latitude... but it won't do anything to improve the original video.

Adam:

I think rich is talking about capture live using the component out like are available on the FX1. It was my understanding that that was a viable alternative to obtain the level of recording capability he was looking for. The idea was that feed is before any compression occurs. Anybody had experience that way ?

Adam,
so the only way (at least the only way mentioned on this thread) to transcode the video to a better codec, thus giving me more freedom when applying color correction or other effects without damaging the quality of the video, is by using a program like Cineform, correct?

Chris,

no I was never planning to capture live, my editing system is extremely heavy and in a very small room, I wouldnt be able to get much by doing that, I was always refering to the case of having the footage on tape

Chris--

Rick talked about recording HDV and uploading in his first post so I was pretty sure he meant recording to tape and then capturing.

Rich--

There are a variety of intermediate codecs, but Cineform seems to be among the best and most used for advanced post work.

If you can shoot to a better codec (using a better cam, like XDCAM EX, for example) or capture live uncompressed without ever recording to tape, you can get much better initial quality to start with.

Capturing LIVE via component or SDI basically grabs the picture information BEFORE the camera has a chance to process and convert the information to MPEG2 and record it to tape. However this requires speedy RAID as has been discussed. Once the picture info is recorded to tape, you can't do anything magical to convert it to 422 color -- that information has to be there to begin with. However, cineform, prores422 and others can store your image information in a 422 CONTAINER, which is comparable to carrying one gallon of paint in a 5 gallon container. You still only have a gallon of paint, but in that big 5 gallon container, you'll be less likely to spill it or for it to slop out the way it might in a 1 gallon container filled to the brim.

Ahh... my bad for not following thread closely. Well capturing from tape and into an intermediate codec is best you are going to get in that process, in my mind.

Capturing direct HDV mpeg files is really a matter of transferring 1 and 0s from the tape. There is no loss there. Converting at capture with an intermediate format like Cineform does involve a transcode, but its generally acceptable because controlled process provides a pristine and editable file. That has 4:2:2 interpolation allowing better access to color correction processes.

Rich, I think the point everyone here is trying to make is:
 

Rich, I think the point everyone here is trying to make is:

ONCE YOU RECORD ON CAMERA - THAT IS THE BEST QUALITY YOU’LL GET

And if you use HDV (or certain other codecs), it doesn’t really hold up well when adding FX, colour correction, etc during editing.

So what we do is, we copy HDV to a better quality codec that can withstand the adding of FX, colour correction, etc. and preserve the original HDV quality.

Let me remind you;

- copying HDV to a better quality codec (Cineform) will not increase the original HDV quality

- It will only preserve the original HDV quality.

ok Rich,
did a simple comparison for ya.

here is a HDV original screen shot,
and a cineform copy screen shot

i rendered both of them to an uncompressed format, and took a close up screen shot - ALREADY, even after ONE generation, the HDV copy has more artifacts! (look at blockiness to the left of the seagulls head in the HDV jpeg)

the first is HDV, the second Cineform (the crop between the two is not perfect, but negligible)


PS tech heads: I'm on my internet computer, not my workstation so I'm using Sony Vegas's pre-installed cineform V2.5 codec. Cineform's current NeoScene Codec is a MUCH MUCH improved codec over the Sony Vegas pre-installed cineform V2.5 codec. I added an simple S-curve and a simple Primary Color Corrector change to both clips.

the HDV JPEG is;
HDV > S-curve and a Primary Color Corrector > uncompressed render > jpeg capture > jpeg crop

the Cineform JPEG is;
HDV > Cineform V2.5 > S-curve and a Primary Color Corrector > uncompressed render > jpeg capture > jpeg crop