AF100 noise
- Thread starter Steve Kahn
- Start date
Yes, but I have a black cat. He will be the downfall of the AF100. (evil laugh)
Ticheli
Well-known member
We've no carnival, although we've been called Carny-trash, and our local waterfalls are a bit boring unless you're standing right there; I might just do some nighttime in the city, but my first "test" of the production AF100 when it arrives will be a little piece for the belly dance troupe that provided the dancer for our initial tests for Barry. Since the emphasis will be on art, not evaluation, we'll light it much more dramatically. Yes, we'll post a link to download for those who wish to see.
Following that, I'll put it to work earning it's keep on jobs that for whatever reason are not best suited for the Red. I've got a feeling it's going to be a busy little bee.
Good shooting and best regards,
Leo
Following that, I'll put it to work earning it's keep on jobs that for whatever reason are not best suited for the Red. I've got a feeling it's going to be a busy little bee.
Good shooting and best regards,
Leo
alpi69
Veteran
I am sorry Barry. But like many others here I have seen noise in the studio-test-shoot on several monitors. The same monitors that do NOT show any noise in Tmurs ISO1600 test, in any of the F3 footage, in many of the HVX vimeo-films and also nothing in the Verse-trailer.
Yes, "uncalibrated monitors" can create even more noise. But what we are talking about is noise that shows up on many monitors that are used by our clients: I saw it on Macbooks, IMacs, Panasonic Plasmas, Acer LCD screens.
Somewhere must have been a flaw in the compression. If your shot looked great and your edit looked great....it can only be the compression, because it is not vimeos fault. there are many trailers without this sort of noise.
I am happy "Verse" looks so good, but it shows there was a technical problem in the other test. We only need to find out what it was so we can avoid it.
bgundu
Veteran
Chasing a cat at F0.95! But seriously, I have a fun idea for my first project with this camera. Think Levis, think hot girl, think loft....
Well at least it's only a couple weeks now!Alright, who is the first person that will be shooting (1) carnival, (2) nightime in a city, (3) slow motion water fall.
No one at Panasonic is comparing the AF to an Alexa or F35. IMO the cams that fit that, low price high performance mark are the F3 and REDone. In fact, no one is comparing it to the RED one either. The closest competitor the AF100 has is the F3 and its 3 times the price.
You should care about price. You cant lump all Digital cameras together into one giant competitive pool, thats not how the real world works. You will wind up disappointed by everything that gets released because it doesnt live up to..... What? whats the perfect camera system? None of them. I got GREAT spotless images out of the AF100, you can too. Learn to use it and light for its strengths and it will reward you.
Well at least it's only a couple weeks now!
i think i got one of the firsts down already....
Agree again Tim. There are grey areas in cost-benefit analysis. If every example of Af100 footage looked as noisy as the Models/studio test, it might give me pause. But I've seen another test on a chart up to very high ISO, and it looked quite clean.
And Timur, speaking of clean, your NYC quick test on the streets, daylight, on Vimeo, looked good terms of image quality. May I ask what lens you used, and generally where you were with ISO and f stop? Thanks in advance for your info.
Last edited:
It was about 85% 14-140, all the slomo long lens stuff, and the shot of the trees covered in Xmas lights was the 21mm F2.8... i looked at my notes.
I think once more AF100's gets "out there" this will become a moot issue.
Go visit the GH2 "Blue Banding" issue....folks expect an $800 still camera
to be the end all when it comes to video. Granted I have seen a few
of the bands they are talking about - but like Barry said "Don't do that".....
Go visit Reduser - It gets really crazy when you see the 89,000 different
RED workflows. They have so many programs with gamma, gamma2, LUT's
for every viewing situation.....It makes my eyeballs hurt.
What is "normal"? - Even the trailers at Apple.com can be inconsistent.
Some look fantastic and some are downright ugly.
Go visit the GH2 "Blue Banding" issue....folks expect an $800 still camera
to be the end all when it comes to video. Granted I have seen a few
of the bands they are talking about - but like Barry said "Don't do that".....
Go visit Reduser - It gets really crazy when you see the 89,000 different
RED workflows. They have so many programs with gamma, gamma2, LUT's
for every viewing situation.....It makes my eyeballs hurt.
What is "normal"? - Even the trailers at Apple.com can be inconsistent.
Some look fantastic and some are downright ugly.
Last edited:
I don't think so. I don't think everyone was loading the video into FCP so you can't say it was something with FCP on everyone's computer. I saw the noise in every version of the footage I saw. The Vimeo streaming, Vimeo mp4 download, and the 2 versions of ProRes that were made available for download. The noise was visible in Quicktime X, Quicktime Pro 7, VLC, Toast Video Player, and Final Cut Pro. In Quicktime it showed whether the FCP compatibility was turned on or off, although there was the usual gamma issue you expect with Quicktime.
With the "A Verse Before Dying" trailer there is no noise when viewed the exact same way as Leo's video. There's also no noise from other AF100 test footage I've watched in the same way, or with GH1 or GH2 footage (since they have similar sensors and codecs). This tells you it's not any issue with the way we're viewing it. It was an issue with the camera, the way it was set up (I doubt this one though), or with something in the post workflow. Since Timur is saying he saw noise at first, and then none after he black balanced, then maybe there was an issue with the black balance function on the camera that was used for Leo's shoot. I know Barry said they black balanced, but maybe there was a bug in the camera with that function.
Too many people were saying there was excessive noise vs other AF100 footage on the same monitors to just dismiss it as monitor calibration.
EDIT: I also highly doubt it was the post workflow, as what I know about the post workflow that was used I don't see how that could have possibly added noise.
Last edited:
Duke M.
Veteran
Timur said it was noisy until he black balanced it, and it black balances in an unusual way (1/2 push of the button instead of a second push.)
Its probably as simple as others thinking they both white balanced and black balanced, but hadn't really because of slightly different operation methods. Everyone should read their manual when it arrives. Its an education issue, problem solved.
Also remember, the camera i had was a hand built, model, that isnt a full representation of the final version. It will probably be more straightforward in the release version.
Sorry guys, but to back Barry G up... And to sort of add as to why his claim makes sense... Your monitor definitely plays a part in what you're seeing... Because it's connected to your graphics card and trust me when I say that they are not all created equal.
I had a look at RED footage on my netbook (alienware) and switching between the high end GPU and the intel one, footage is night and day. That also goes for everything i watch online and offline. Your GPU definitely adds noise to every video and photo you glance at if it's not up to task.
It has to do with decoding video on the fly if it's GPU intensive. And now-a-days that is mostly the case.
I doubt the AF100 is noiseless, but neither is a RED MX image and that costs three to four times as much.
Noise definitely won't keep you from shooting something great, nor will the AF100 be incapable of producing footage that could cut with film... If the atrocious image from a 5 or 7D can then the a f100 IS film.
Added: and I am not saying the footage is noiseless, just that I'm fairly sure anyone without a more modern gpu is seeing more noise than they should.
I had a look at RED footage on my netbook (alienware) and switching between the high end GPU and the intel one, footage is night and day. That also goes for everything i watch online and offline. Your GPU definitely adds noise to every video and photo you glance at if it's not up to task.
It has to do with decoding video on the fly if it's GPU intensive. And now-a-days that is mostly the case.
I doubt the AF100 is noiseless, but neither is a RED MX image and that costs three to four times as much.
Noise definitely won't keep you from shooting something great, nor will the AF100 be incapable of producing footage that could cut with film... If the atrocious image from a 5 or 7D can then the a f100 IS film.
Added: and I am not saying the footage is noiseless, just that I'm fairly sure anyone without a more modern gpu is seeing more noise than they should.
Rick Burnett
Section Moderator
No one is saying that anything should be completely noiseless, if you read all the posts. The question came into play because footage from other cheaper cameras on the SAME equipment exhibited less noise under the same sub-optimal display settings. That was the concern. Also, myself included, no one thought it was a huge problem, just wanted to get to the bottom of the difference and what was happening between the sensor and my uncalibrated monitor as to why. Now that we have some idea of what could have caused that (black balancing, hand built camera, unfinished firmware, not crushing the blacks as much below a certain IRE) we have something to work with in knowing how to get the footage to look the way we want.
Much of my stuff does end up online, and therefore for me, given that is ONE of the medium that it will be displayed (as well as projected) I am trying to cover multiple bases to get the best image that I can in the majority of situations. I don't have this problem with my 7D footage, it holds up well in both. Therefore, I feel confident that I can get this with the AF100 as well, especially after looking at the trailer for 'A Verse Before Dying' on multiple displays and seeing footage that exceeds some of the best 7D footage I have seen. It's just a matter of understanding what i need to do where to get there.
And you want to talk about image quality, hooking a black magic intensity card through the component connectors to a display looks HORRIBLE. I couldn't believe how bad that card is on playback. On capture it is much better, but i was shocked.
Steve Kahn
Well-known member
True.
I recalibrated my monitors as the noise does disappear. It was just very strange how I had never personally seen so much noise at ISO 400 on all of those monitors as when I looked at Leo's footage with the multi-candle lit female.
Again, we just wanted to know why. And, I think it was a reasonable question even if it were just a post issue which it was not as Timur reported problems with noise at 400 until he did a proper black balance.
And, yes it is an education issue. We must learn all aspects if this camera. Sometimes that comes with asking not such politically correct questions. Sometimes that comes with rocking the boat a bit.
In some cases it can be black balance, and in some cases it can be the post flow. There is a lot that can go wrong. I just watched a rough cut mp4 of "verse".
Looked horrid on VLC player, and perfect on WMP and quicktime. Go figure. Sometimes one little thing sets off a downstream chain of events.
Looked horrid on VLC player, and perfect on WMP and quicktime. Go figure. Sometimes one little thing sets off a downstream chain of events.
Steve Kahn
Well-known member
Right. So when better then now to nail down these issues and create ironclad post workflows? Even the biggest budget features show their trailers on the net. It would behoove us to optomize for that distribution outlet as well.
Last edited:
Barry_Green
Moderator
Nailing down a proper post workflow would indeed be helpful.
There was once upon a time when someone would say "why does this look so awful when I play it in VLC player?" and we could then give an answer. But today, it's much more common to say "why does the footage from that camera look so awful?" And when dozens of us look at it and it looks friggin' amazing, you can see where the frustration comes in.
Here's the full-stop end-all be-all: the AF100 footage can look utterly fantastic, if the image fidelity is maintained all the way through. And if somebody's got something "off" somewhere along the way, it can make it look awful. So yes, some effort needs to be put into figuring out what those things are. The only one I've absolutely identified, is proper monitor calibration. Timur has now identified VLC player and Kholi has said his intel GMA was an issue. And grimepoch has identified that component output from a BlackMagic intensity looks horrendous (although I would also ask to what kind of display?)
These are issues that need to be worked through, that is agreed. I just want people to spend their energy looking for the right things, and not pointing fingers at something that's already been exonerated.
There was once upon a time when someone would say "why does this look so awful when I play it in VLC player?" and we could then give an answer. But today, it's much more common to say "why does the footage from that camera look so awful?" And when dozens of us look at it and it looks friggin' amazing, you can see where the frustration comes in.
Here's the full-stop end-all be-all: the AF100 footage can look utterly fantastic, if the image fidelity is maintained all the way through. And if somebody's got something "off" somewhere along the way, it can make it look awful. So yes, some effort needs to be put into figuring out what those things are. The only one I've absolutely identified, is proper monitor calibration. Timur has now identified VLC player and Kholi has said his intel GMA was an issue. And grimepoch has identified that component output from a BlackMagic intensity looks horrendous (although I would also ask to what kind of display?)
These are issues that need to be worked through, that is agreed. I just want people to spend their energy looking for the right things, and not pointing fingers at something that's already been exonerated.
TheReverend
Well-known member
I'm pretty sure you've been right about the IRE values all along. Different players (VLC, Quicktime, software with GPU accelerations) can be a train wreck when it comes to codecs and gain/gamma. We know the AF100 images are clean, but they aren't "noiseless", and many players can show the noise exaggeratedly. The big question is, is there a step in the Post-workflow that makes the noise more prone to abuse, or is there a step that we can add to the Post-workflow to reduce the potential problem? To figure this out we need to start with initial observations about where/when we see noise on these (improper monitor calibration, VLC player, intel GMA, BlackMagic intensity) devices.
If you've experienced the "noise" anomaly, answer the following questions:
1. Is it ONLY in the blacks?
2. Is the "noise" observed blocky coming from compression?
3. Is the noise colored, or black/white specks?
4. Is the noise as fast as the frame rate?
5. Is there a specific file type (mp4, mov) or codec (h264, prores) that exhibits the anomaly?