Miraizon being bought
dapoopta
18 Jan 2015 06:19
http://www.filedropper.com/prores
http://www.filedropper.com/pjpeg
Download both of those, zoom in 150-200% on eyes, and look at the whites and the noise on one vs the other. You do get a gamma shift issue with prores, not sure if it is because of the miraizon codec.
I can see a difference. The Prores is less noisy and smoother... still all detail is there.
The sizes are fairly close too, maybe 10% bigger for prores... which is okay for me. This is also ProRes 422, not HQ.
http://www.filedropper.com/pjpeg
Download both of those, zoom in 150-200% on eyes, and look at the whites and the noise on one vs the other. You do get a gamma shift issue with prores, not sure if it is because of the miraizon codec.
I can see a difference. The Prores is less noisy and smoother... still all detail is there.
The sizes are fairly close too, maybe 10% bigger for prores... which is okay for me. This is also ProRes 422, not HQ.
Mizamook
18 Jan 2015 06:24
Downloading...yeah I know...it's slow.
In the meantime, can you send original?
In the meantime, can you send original?
Mizamook
18 Jan 2015 06:30
Yes - without seeing the original, I can say that this corroborates my claim.
WHY is the PJPEG noisier?
I suggest strongly that the PJPEG appears noisier because the noise is in the original.
If the original is noisy, I want the noise reduction to be because I chose it so, not because the codec removes it arbitrarily. Especially since I work really hard to get the detail I like to see.
I want to see a render that is as close to what I see on the screen before I render as possible.
This is more likely, in my opinion, using PJPEG rather than Mirazion ProRes.
PLEASE show my the error of my method.
It cannot be simply that what I call detail is finer than what everyone else calls detail.
And yes, the gamma shift is disturbing, but from AE CC we get that anyway in PJPEG, and I've already raised that specter.
WHY is the PJPEG noisier?
I suggest strongly that the PJPEG appears noisier because the noise is in the original.
If the original is noisy, I want the noise reduction to be because I chose it so, not because the codec removes it arbitrarily. Especially since I work really hard to get the detail I like to see.
I want to see a render that is as close to what I see on the screen before I render as possible.
This is more likely, in my opinion, using PJPEG rather than Mirazion ProRes.
PLEASE show my the error of my method.
It cannot be simply that what I call detail is finer than what everyone else calls detail.
And yes, the gamma shift is disturbing, but from AE CC we get that anyway in PJPEG, and I've already raised that specter.
dapoopta
18 Jan 2015 06:33
Here is a 3 way compared original on my screen.
It seems the original wasn't as sharp or noisy as the PJPEG, so I'm guessing pjpeg is sharpening a little.
http://snag.gy/ypVGB.jpg
http://snag.gy/ypVGB.jpg
It seems the original wasn't as sharp or noisy as the PJPEG, so I'm guessing pjpeg is sharpening a little.
http://snag.gy/ypVGB.jpg
http://snag.gy/ypVGB.jpg
Mizamook
18 Jan 2015 06:42
Thanks - I can see what you say, and I can also see what I say.
Within the three-pane image, comparing only with the image you linked to at full screen, so I see all three, normal magnification of your already magnified images, I can see the difference. I see blobs with soft edges where there were patches before. I also see increased noise in the PJPEG - wierd.
I guess it depends on what you want.
Right? A beauty shot with human skin might benefit from a softening of detail, whereas a scenic shot where one might want to see individual rocks or tree needles might suffer.
This would explain why some of the comments commending Miraizon come from shooters who work with people. On the other hand, I want to see what I shot, and no less, so I cannot abide by it, and consider my money and time wasted.
Interesting that your PJPEG exhibits more noise. Mine is barely distinguishable from the original (generally) and certainly not adding that much grain or sharpening. I wonder if it is hue-dependent.
Within the three-pane image, comparing only with the image you linked to at full screen, so I see all three, normal magnification of your already magnified images, I can see the difference. I see blobs with soft edges where there were patches before. I also see increased noise in the PJPEG - wierd.
I guess it depends on what you want.
Right? A beauty shot with human skin might benefit from a softening of detail, whereas a scenic shot where one might want to see individual rocks or tree needles might suffer.
This would explain why some of the comments commending Miraizon come from shooters who work with people. On the other hand, I want to see what I shot, and no less, so I cannot abide by it, and consider my money and time wasted.
Interesting that your PJPEG exhibits more noise. Mine is barely distinguishable from the original (generally) and certainly not adding that much grain or sharpening. I wonder if it is hue-dependent.
dapoopta
18 Jan 2015 06:46
The original files were Odyssey 7Q raw images.
I think the prores one most accurately resembles what was shot. I did this test a while ago on a different clip with similar results, and I just grabbed this one tonight to do it again. I did tons of testing in the beginning too and decided that I wanted as close to original as possible, and I feel like that's what I get with this codec.
The PJPEG just worried me because of A) banding, and B) added noise
I also have seen similar results in HD footage, this is all 4k.
I'm just putting everything out now as prores 422, seems like the best option for me at the moment.
I think the prores one most accurately resembles what was shot. I did this test a while ago on a different clip with similar results, and I just grabbed this one tonight to do it again. I did tons of testing in the beginning too and decided that I wanted as close to original as possible, and I feel like that's what I get with this codec.
The PJPEG just worried me because of A) banding, and B) added noise
I also have seen similar results in HD footage, this is all 4k.
I'm just putting everything out now as prores 422, seems like the best option for me at the moment.
Mizamook
18 Jan 2015 06:55
Well at least I'd certainly agree with you that the ProRes looks nicer for this particular subject - distinctly so. I have no hesitation admitting that. And I respect your integrity and opinion. I wish I could say I was happy with 422 or 422HQ for my work. Of course my work doesn't sell lately anyway, so who am I kidding?
If I ever shoot something with flesh in it, I'll use ProRes. I did some work recently that I was really into, and thought I'd use Miraizon 4444XHQ, and the file was HUGE and the detail level was still lacking. I don't get it.
So I'm hoping I'd get a kick in the pants explaining what I was doing (or thinking) wrong. A recent project I shot in would benefit from me transcoding to ProRes, but, being I only have Miraizon, and the other person has the ability to transcode to real ProRes, I can't do anything without hurting the IQ of the footage I have. So he gets the fun job of doing the transcode to ProRes.
Oh well.
If I ever shoot something with flesh in it, I'll use ProRes. I did some work recently that I was really into, and thought I'd use Miraizon 4444XHQ, and the file was HUGE and the detail level was still lacking. I don't get it.
So I'm hoping I'd get a kick in the pants explaining what I was doing (or thinking) wrong. A recent project I shot in would benefit from me transcoding to ProRes, but, being I only have Miraizon, and the other person has the ability to transcode to real ProRes, I can't do anything without hurting the IQ of the footage I have. So he gets the fun job of doing the transcode to ProRes.
Oh well.