bronxjragon
Well-known member
mattsand i was hoping that you'd reply to my posting : ) only one who seems to know how to make a good d90 image
Picture Perfect: a 1080p24 D90 - the noiseless, the stepless and the bandless
- Thread starter beckspace
- Start date
bronxjragon
Well-known member
matt,
here's some stuff i shot today with color correction and using your d90 rescaler
http://www.vimeo.com/2347733
without buying neat video i wish i could scale up to 1080p because from teh first posting and images, it looks like it also cleans up the artifacts as well.
*p.s. don't bother full screening that video, artifacts galore
here's some stuff i shot today with color correction and using your d90 rescaler
http://www.vimeo.com/2347733
without buying neat video i wish i could scale up to 1080p because from teh first posting and images, it looks like it also cleans up the artifacts as well.
*p.s. don't bother full screening that video, artifacts galore
mrmoe
Well-known member
Emanuel
Cross-Examiner
Maybe using an intermediate codec in a 2nd step?
i think you misunderstand the problem. it's not that the codec doesn't work with neat video, it's that neat video tries to recognize and learn the noise pattern of the image and with h.264 encoding it's not constant.
/matt
Emanuel
Cross-Examiner
I've got it. I'm just wondering if the h.264 encoding is not the proper environment for Neat Video, why not using an intermediate codec at post-production which can be suitable for the task? I mean, for a similar result as it occurs with the D90's intraframe MJPEG.
how would an intermediate codec remove the blocks created by the origination codec? it might add some new that you could remove more easily, but that doesn't solve the original problem, does it?
/matt
Robert Altman
Active member
I would doubt that any further recompression will help. Once the original codec has done its damage, you are not going to achieve anything by a new codec (and its inevitable damage).
Emanuel
Cross-Examiner
...suitable for the task?
...suitable for the task?
Not sure. When we deal with technology, a lot of what does seem impossible at times, it becomes possible not only sometimes but often. Didn't you notice yet? I did.
Maybe an intermediate codec may bring more garbage, easily removable. If not, why might we use it? A 10-bit CineForm intermediate codec, for example, shall offer what a 8-bit codec (with its blocks) cannot. Introducing less disturbance, for sure.
Even though far away from the techie side (not by background but for immersion in other waters), I don't need to place any credits on the table for wondering if, anyways, there is not a solution for a plugin that deals in a more effective manner with a intraframe MJPEG than another one which doesn't offer a pattern.
Nevertheless, as the original poster takes notice, «probably Neat Video will blur the 5D more than a reconstruction of the lost information», the point is to know how a superior intermediate codec might handle the problem over a codec such as that (5D Mark II). Therefore, the natural question: 'suitable for the task?'
...suitable for the task?
Not sure. When we deal with technology, a lot of what does seem impossible at times, it becomes possible not only sometimes but often. Didn't you notice yet? I did.
Maybe an intermediate codec may bring more garbage, easily removable. If not, why might we use it? A 10-bit CineForm intermediate codec, for example, shall offer what a 8-bit codec (with its blocks) cannot. Introducing less disturbance, for sure.
Even though far away from the techie side (not by background but for immersion in other waters), I don't need to place any credits on the table for wondering if, anyways, there is not a solution for a plugin that deals in a more effective manner with a intraframe MJPEG than another one which doesn't offer a pattern.
Nevertheless, as the original poster takes notice, «probably Neat Video will blur the 5D more than a reconstruction of the lost information», the point is to know how a superior intermediate codec might handle the problem over a codec such as that (5D Mark II). Therefore, the natural question: 'suitable for the task?'
rawfa
Veteran
#1 The Nikon guys are ***holes for taking this idea so close and yet so far. CREATE a freaking firmware to fix this already!!!
#2 I refuse to believe there isn't tech head around here capable of creating an AF script for this. DO IT and then sell it to Nikon or AF...or both (but use us as guinea pigs for free before)
#3 In the meanwhile could someone write a "D90 Stair-Stepping Removal for Dummies" type of tutorial? I mean a VERY BASIC in your face step by step tutorial for (technically speaking) mentally delayed individuals such as my self. Or maybe do like one of those Creative Cow tutorials where they film the computer screen while narrating what they're doing (and then post it on vimeo).
#2 I refuse to believe there isn't tech head around here capable of creating an AF script for this. DO IT and then sell it to Nikon or AF...or both (but use us as guinea pigs for free before)
#3 In the meanwhile could someone write a "D90 Stair-Stepping Removal for Dummies" type of tutorial? I mean a VERY BASIC in your face step by step tutorial for (technically speaking) mentally delayed individuals such as my self. Or maybe do like one of those Creative Cow tutorials where they film the computer screen while narrating what they're doing (and then post it on vimeo).
I agree with #3. While many of the people in this forum are experienced in video editing, I for one have no clue about anything when it comes to FCP, video formats, compression of video, etc. I'm sure there are plenty of viewers that come in here who arent experienced looking for a step by step guide.
rawfa
Veteran
You're a mean bad man
P.S - The request by no means was directly aimed at you. It might surprise you but there are more than 40 thousand members
Let's hope some have a more positive attitude when it comes to helping people with less technical knowledge...after all that is the spirit a forum should have and the reason most of us are here.
Emanuel
Cross-Examiner
At times, a removal does seem obvious, sometimes does not. The hassle is there, just not so noticeable or noticeable at all. There's a margin where the flaw can be accepted for the job or not (with the due respect -- like a woman :laugh
. That's what I am referring to. Quoting again the original poster: «probably Neat Video will blur the 5D more than a reconstruction of the lost information».Or then, getting a somehow method in order to create such pattern which could be more effective and take away the non-pattern blocks too. But on that, only one or two of you are saying this might not be working in any way other than extracting the add-on waste. Returning to my previous post, I've seen a lot of solutions which, not one or two but, a lot stated as impossible straight out of the box. And later, it could be possible to attest they were not, after all.
yes, i see what you're saying but it's impossible. it's just basic signal theory. even if you succeed in removing it you will have removed information too unless you've figured out a way to differentiate between signal and noise, which brings us right back to the original problem. you don't even need to assign any properties to the "intermediate codec" for this logic to hold. i'm not explicitly stating that neat video can't remove h.264 artifacts, merely extrapolating the consequences based on that assumption.
/matt
that must be why i developed a plugin that takes about 5 seconds to download, 2 more to apply, and doesn't even have any setting. did i mention that i'm giving it away for free? did i also mention that i've posted around half the posts in the stairstepping thread. i'm one of the fcking founding father for crying out loud. without me you wouldn't even know what stairstepping was.
good bye.
/matt