Jeremiah Black
Forum Replies Created
-
Jeremiah…please let up on the name calling here. Calling Walter a liar is going a bit overboard.
Did I call him a liar? IF I did, I apologize. I remeber saying he’s spreading techincal and mathematical misinformation, which is true.
“All he is saying, and I am saying, is that capturing upcompressed HD from DVCPRO HD does not improve the quality of the original image. They will look identical. He isn’t wrong there, he isn’t lying. I am not an engineer but a creative editor, and I know that both will look identical.”
You are both correct here. 100%. But Wather is saying that even with rendering there is NO differnce. That is incorrect. The differnce may be such that he and his clients dont care. That’s cool. Actually, that’s more than cool. That’s less money to spend! But they are NOT identical. Nope.
“Where the path diverges is when rendering comes into play. And all that is happening here is a misunderstanding. Walter isn’t saying that whether you render in DVCPRO HD or uncompressed HD the quality is the same. No. He is saying that he gets perfectly fine quality working in the DVCPRO HD codec. As do I. So he isn’t wrong. Our workflow, as you admitted, is viable. The confusion lay in the rendering process, which you explained nicely.”
I know, and I’m glad his workflow works for him- cheaper and faster!. But he refuses to admit, that there’s a slightly better way. And there is.
“But I don’t see anywhere in this listing, nor in previous posts, where Walter out and out lied. The symantics of this discussion hinge on the quality loss when you render color correction…not initial quality. I know that misunderstanding one another can be frustrating, but please…name calling.”
Walter claims he “did the math”, “did the scopes” and found NO difference. Either (1) he did both incorrectly or (2) he didn’t do them and is falsey claiming that he did. I have proof of neither, so I’ll just let the reader decide which it is.
-
“How’s a 65′ screen in a movie theater? We produced our first independent film on the Varicam and cut it using the DVCPro HD codec, color corrected it in FCP, mastered it back to my 1200A and played it directly off the 1200A to a Panasonic projector to an audience of 200, mostly local filmmakers and producers. It was flawless and looked just like 35mm film. 24p production all the way and it was outstanding on the screen with all positive comments. So for all these professionals in the business it looked really really good.”
If you think it looked as good as 35mm, then you are insane and/or blind. Super 16 (which is far better then HD) that’s had a D5 DI done doesn’t look as good as 35mm. You obviously have no eye for image quality, which is a shame since you work in post. The fact that eveyone liked it is probably a testament to your really good creative skills as an editor and/or animator. Or a great DP which can make all the differnce. Hell, I once saw an HDCAM indie movie that everyone thought was DV, because the sound was so poor. And I sound desinged a minidv movie years ago that everyone in the theater thought was shot on HD because the sound was really good and they were super engaged. There’s a lot of varibles that go into audience perception. BUt this conversation is about compression math, and you seem to only want to talk about your feelings.
“Have you even worked with DVCPro HD yet? I don’t think so because you obviously don’t understand how good it looks. Our clients are broadcast networks and they are telling us the work is outstanding. ”
Been working with it for years, and I’m working on it right now, starting at DVCPROHD footae on an HD monitor as we speak. I think a question for you is “have you worked with anything better than DVCPROHD, because you obviously don’t understand how much better other formats (like 35mm) are” And the answer is probably no. Try to get your hands on a 10 bit log file from a 2k 35mm negative scan. POST THAT, Walter, and then come back to the board and we’ll talk about image quality and what it could be.
I’m a colorsit. I work with minidv, dvcpro50, DVCPROHD, HDCAM, 10 bit log files, whatever. I am intimately familar with DVCPROHD, in both compressed and uncompressed forms. I just never capture it because that work has usually been done before it gets to me, which is why I initially posted a question about my capture settings.
“Ok, you’re calling me a liar and putting false words in my mouth? Dude, I’m completely done with this thread and really don’t care how you want to edit your footage.”
At last count, you claimed there was no mathmatical difference. If you’d like to recent you can do so. If not, you remain incorect.
-
“HDCAM is 3:1:1? Well shoot, I forgot about that. OK, next time I master to D5.”
yeah, wierd, but it’s 3:1:1. Don’t know why, but, yeah, you get a little color shift going from 4:2:2 codec to a 3:1:1 one. Also, HDCAM is more compressed than D5. So, I it’s a little more expensive, but D5 keeps a purer video migration.
But, like you said, the video will NEVER stay pure for long. It’ll get compressed to mpeg2 for digital cable or whatever. NOTHING stays pure because the files are too big and can’t be transmitted. Some say, “oh well, it’s going to be impure sooner or later, so who cares”. Some say, “since it’s going to be messed with and go through the wash, I want to start with the cleanest master possible.” Whichever. Mostly just comes down to whos paying for it, and how much they care.
-
“But I am with Walter for the quality…it looks fine. I have upconverted 720p24 DVCPRO HD to 1080p 23.98 HDCAM a couple times and both times it looked great. Even when we went in for fixes to a lower third at another post facility, the online editor, looking at it on the HUGE Sony HD CRT commented that it looked good. QC came back clean from the networks. So that is a perfectly viable workflow. ”
Yeah, it just depends on how anal you want to be, and what type of color correction you’re doing. It’s a good codec, and so it probably will end up looking good. But if you’re blowing up to film or showing someone on a big screen, it’s no good. BUt for most TV, it’s probably prefectly great. You do it, and your clinets are happy, so go for it.
“As is what you are doing. Capturing at 10-bit or 8-bit uncompressed and coloring that way will get you better results due to the compression you are dealing with. And side by side on similar monitors you WILL see a difference. BUT that is not how the viewing audience will see it. When it goes to air the footage is compressed into an MPEG-2 stream (high end, mind you) and sent out over the airwaves. Then it lands on their consumer sets and can look any NUMBER of ways. At that point you might not notice the difference in quality.”
Oh yeah, totally right. Hey, DVDs look great, too, and they’re crappy Mpeg2 compression. But when your making a D5 master for someone who’s spent a looot of money on something, and they want the best, then they should get the best.
But that is not to say abandon your high standards. Making it the best it can be should be our goal. Given the budgets that many of us have to deal with, coloring the DVCPRO HD footage in its native codec with the tools at hand (or even with Final Touch) and outputting to DVCPRO HD or upconverting to D5 via a capture card is all that our budget allows. We don’t have time to recapture the footage, and we might not have the drive space. And since the workflow we use looks great and passes QC with flying colors we go with it. Given the time and budget I might do the uncompressed thing. But I can’t, so I don’t. But to call what me and Walter are doing as substandard and spreading of misinformation is going a bit far. It is a viable workflow that works well.
I NEVER said “substandard”. It is totally viable. And I never said what you’re spreading is misinformation. What Walter is spreading is misinformation because he says “there is no quality differnce. Recompressing to a highly lossy codec doesn’t lose information”. That is a lie and an untruth. The fact is there is an actual quality differnce, that may or may not be negligible depending on what you want and how deeply you want to care. No one ever said your worklfow is no good. Just that it’s not the best and the purist. And whether someone wants to do it or not isn’t the issue for me, the issue is in saying “there’s no mathematical difference” which is untrue and unscientific. Also, it’s important to learn HOW everything works, and what exacty is happening to your footage at every stage to all become better at post. So, I get upset, not becuase some choose to render back into the DVCPROHD codec, but because some claim that “nothing’s happening”. Which is wrong.
Another great discussion is the merits of different ways of up/cross converting 720p footage to 1080i. Some use Telanex, some go to HDCAM, or maybe use a Kona 3, or Blackmagic HDPRO. I personally, would never go to HDCAM, since HDcam is a 3:1:1 color space and it will shift my color timing that I did in a 4:2:2 space. Some will for a variety of reasons. And that’s cool. But, regardless, I think it’s good to talk and make sure thescientific facts of whats happening are out in the open.
-
“But if I do not render, then there is no render file to refer to. Real time effects are effects that create no renders…dissolves, audio dissolves, 3-way color correction. I don’t render, so no render file is created. If I start with a new project and add real time effects, and look in the RENDER folder…it is empty.
If it renders them, where is it putting them?
Not trying to sound snide or incredulous…just curious as to why you say that it renders when I don’t see the render file.”
That’s a great question. I think they go into the RAM. But if you were to export your timeline (and hence be writing files to disk) I think you’d find it takes a lot longer to write than if you didn’t have the 3 way correcter and the grreen render bar.
Realtime means realtime RENDER, not no render. The data and calculations have to made. It just makes them fast on the fly because the filters are simple and it can do so, so the writing of a separate file to play isn’t necessary. BUt it’s still rendering and recompresing.
-
Nope…no rendering…no matter what render option I choose. Dark Green no render bar.
The green render bar is still a render bar. It just means your computer is doing it really fast and in full quality (realtime). But, you can’t just change footage and not have a render. A complete video file must be created somwhere (either in RAM or an actual file) that your computer then “plays”.
-
I believe realtime coloring still requires rendering. You just don’t have to WAIT for it (realtime) because FCP and the G5 do it so fast. It plays it in realtime. But it’s still rendering. It still has to make a NEW file, because something has changed.
-
“Who’s recompressing? If you capture via firewire, it is at native compression…no recompression occuring. Or are you talking about taking the footage from FCP into After Effects then back to FCP? If so, then yes, you would be better off working in an uncompressed environment. But if you are staying in FCP, and using Final Touch to color correct, then editing DVCPRO HD native is fine.”
Or, are you saying that when you render ANY filter that you are re-compressing the image? Hmmm…didn’t think of that.
Yeah, that’s what I’m saying. Sorry if it’s not well put. BUt, essentially when you render ANY filter to the image it has to write data and make a new file snd compress. If you’re in a DVCPROHD timeline it’s compressing (recompressing) your file into the DVCPROHD codec. If you are in an uncompressed space, it isn’t recompressing your footage. That’s why if you do a cuts only edit, uncompressed doesn’t matter, because the footage is untouched, and won’t be rendered/recompresed
-
“Actually we go by what the networks tell us which is we’re delivering some of the best HD material they’re receiving “by far.” So I know the DVCPro HD workflow is incredibly clean.”
Hey that’s great. Now just realize that it would be even better if you stopped recompressing it and throwing away information.
“Done the tests, did the math, did the tests again, ran some more math, did split screen with Avid DS/HD, it’s all the same to us. We’ve owned our 1200A for almost 2 years now so we’ve tested this to death.”
Um, if you don’t notice a difference than that’s cool. But you obvioulsy didn’t “do the math”, because the math says not to recompress your footage into a lossy codec. There’s no differnt “math” you could’ve done, ace. And I doubt you “did the scopes? because I have, and there’s always a difference. How could you not expect an image to look ANY different after compressing it? Geez, you can see the difference even on a software scope in FCP.
“If uncompressed makes you feel better, then by all means work in uncompressed. That last thing I want to do is convince you to work in a format you’re not comfortable with.”
Don’t patronize me, Walter. You are simply wrong, wrong, wonrg. and everyone on these boards knows it. I’ve read you argue time and time again with people that have 10 times your intelligence and knowledge (graeme nattress comes to mind), and you are wrong everytime. No one agrees with you. Let it go, dude. If you want to work that way, that’s cool. And if your material looks great, even better! But quit spreading factualy innacurate info on the boards, just because you have some desire to think that the way that you do it is “the best way”. It’s not. You’re wrong. Don’t pollute the boards with bad facts. People come here to get answers. The truth is that the DVCPROHD codec is awesome, and looks great. You can even recompresses footage into it twice (as you do) and it’ll still look good! But the way to preserve as many bits of your image as possible is to either capture it uncompressed for rendering back into uncompressed, or just redering the DVCPROHD footage into uncompressed out of a program like after effects. This is the best way to preserve all your original information. If you are doing a “cuts only” edit with no rendering, then don’t bother with uncompressed. But if you’re rendering alterations, then don’t recompress. These are the facts. Please stop subsituting your “reality”.
-
“I can understand the need to go uncompressed for coloring. Better compression than DVCPRO HD, although no gain in quality. But I know that coloring in an uncompressed timeline can give better results. So I get ya.”
Yeah, that’s what I’m always saying, “sure, there’s no GAIN in quality, just less loss in the process.”
“But I am not sure about the drive speed. Uncompressed 10 bit, even at 23.98, requires a lot, so I don’t think 100MB/s is gonna cut it. You can try, but I don’t think that will be enough.”
Yeah, well 8 bit uncompressed 4:2:2 footage with a frame rate of 59.94 requires about 110 MB/sec (132 for 10 bit). But if FCP is removing flagged frames and only capturing the 23.98 frames that are wanted, then the Hard drive speed suddenly drops to a requirement of around 66 MB/sec. you can do a test youself. Take a clip and render it out of AE or whatever program @ 10 bit uncompressed 1280×720 @ 23.98. Open it in quickitme and you’ll see the data rate is only around 66 mb/sec. The published uncompressed frame rates for HD are really high because the specs are 60i or 59.94. but if you’re only using 23.98, then the specs are much lower.