AF100 noise

Pertaining to the Black Magic, this actually has nothing to do with the AF100, this is in general. Something is wrong with the Red channel on their card, and others have observed it as well. It is WAY over saturated. In fact, on the LCD I had attached, I had to decrease the Red channel almost completely to get it to match up with a calibrated monitor even a little bit.

The editor was over of our parody the other day and we were mixing audio in my studio, I was playing the footage on that screen. He said "Man, maybe I should reduce the saturation in the reds a bit" to which I had to tell him, NO, there is something wrong with that card. I am going to replace it with either the Motu breakout box or the Matrox one that people seem to have better luck with.

I think because the cameras aren't in all our hands yet, we aren't able to work as a big team to narrow down any potential pitfalls of what people do. People come here to get into the discussion because they are truly interested in helping and have invested in the equipment and are just a little baffled at the time, kind of like, not what they expected. If you don't have the camera and you are bored, of course you are going to pixel peep :) I have nothing to shoot at the moment so I am spending this time researching and such.

What I DO like is when we work together as a group to try and figure out solutions for people, which is when DVXUser shines. Frustration aside, I'm glad there is this much brain power on these things.

It's easy to destroy perfectly good footage :)
 
TheReverend said:
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?
Click to expand...

Leo's video:
1. No. Grays and blacks below probably 40-50 IRE
2. No.
3. Light specks
4. Yes, changes with each frame.
5. On 17" Macbook Pro mid-2010 computer screen: Noise with Vimeo streaming, mp4 download from Vimeo, and 2 different ProRes downloads. On Pioneer Elite PRO-150HD 60" plasma: Noise when running from computer MicroDisplayPort to TV directly, BUT NOT when running from MXO2 Mini to TV.

"Verse...." trailer
1. No noise
2. No noise
3. No noise
4. No noise
5. On 17" Macbook Pro mid-2010 computer screen: NO NOiSE from Vimeo streaming or Vimeo source file download. On Pioneer Elite PRO-150HD 60" plasma: NO NOISE when running from computer MicroDisplayPort to TV directly, AND NO NOISE when running from MXO2 Mini to TV.
 
You start evaluating a Post Workflow for this camera by :

Shooting absolute technically "perfect footage" appropriate to evaluation eg Absolute Black (lenscap on), Grey scale & Black /white cards and suitable subjects such as skintones.

(IMO color is not a requirement for evaluation in the first stage. It could follow along with motion testing - pans tilts etc .)

The footage is shot using the tools ( Waveform and vectorscopes.) and checked back against the same tools.

The footage needs to be established as being 'the control " and then distributed to interested parties willing to test their workflows with it and report with samples.

The footage MUST remain as its Source footage CODEC XXX. This ensures there are no variables introduced to this point. ("Control Footage")

Let the games begin:

The Control files can now be viewed on a calibrated REFERENCE Monitor.
Trick is...play to the Ref monitor from what? ie. what software and what platform and what additional hardware?
Every stage and hardware chain from this point onward introduces variables into the workflow. (including transcoding!)

I would love to be part of the test. Allow me access to the 'Control Footage' from the AF100 and I will run it thru my workflows. eg CS5 Windows, Flame Suite, FCP, Lustre and Film Base CC , Da Vinci.

The CODEC is considered "pure and issue free" and apparently eliminated from the start ...so the tests are for the IDEAL post workflow. Everyone will be able to work out from the control ...what works for theirs..

(I have previously asked in other threads about this so hope this doesnt offend anyone again seeing as others are talking about it as well albeit from the base of "noise")
 
Mike@AF said:
Leo's video:
1. No. Grays and blacks below probably 40-50 IRE
2. No.
3. Light specks
4. Yes, changes with each frame.
5. On 17" Macbook Pro mid-2010 computer screen: Noise with Vimeo streaming, mp4 download from Vimeo, and 2 different ProRes downloads. On Pioneer Elite PRO-150HD 60" plasma: Noise when running from computer MicroDisplayPort to TV directly, BUT NOT when running from MXO2 Mini to TV.

"Verse...." trailer
1. No noise
2. No noise
3. No noise
4. No noise
5. On 17" Macbook Pro mid-2010 computer screen: NO NOiSE from Vimeo streaming or Vimeo source file download. On Pioneer Elite PRO-150HD 60" plasma: NO NOISE when running from computer MicroDisplayPort to TV directly, AND NO NOISE when running from MXO2 Mini to TV.
Click to expand...

When initially looking this information over, it sounds like classic Mac OSX gamma shifting. But notice that the gamma shifting did NOT affect the second video negatively, which means there is something different in the pre-distribution postproduction workflow between these two vids that causes or exacerbates the "noise" issue.

Let's wait for some more info and try to narrow down from there.
 
grimepoch said:
And maybe I should say, this is not a color correction issue in reality, it's a noise suppression issue.
Click to expand...

I agree with everything your write. The need of proper callibration doesn't explain why footage from other cams looks less noisy on the same apparently poorly callibrated monitors.

But I do think it may be a gamma curve issue. All sensors produce noise in the black and if due to the gamma curve the black are too high, noise will be more visible. This can easily be corrected in color correction/grading by correcting the black levels.
 
Barry_Green said:
If your stance is "people are going to shine a flashlight in there no matter what, and I don't want them to see that noise there", then I would guess that the only solution is to crush that region all down to black then. Which really isn't that much of a big deal, as the 0-7.5 IRE range is all supposed to be viewed as "black" anyway. I don't know what else to tell you.
Click to expand...

When people write that "but footage from other cams on same monitor don't look that noisy" comments you can't translate that into shining a flashlight. That's apples and pears. Even if a monitor isn't callibrated properly, if being used for many years and the viewer never saw noise but today he views AF100 footage and he sees a lot of noise, everybody who can think rationally and logically will conclude that there's too much noise in the AF100 footage. The state of the monitor isn't relevant in that.
 
RieniO said:
When people write that "but footage from other cams on same monitor don't look that noisy" comments you can't translate that into shining a flashlight. That's apples and pears. Even if a monitor isn't callibrated properly, if being used for many years and the viewer never saw noise but today he views AF100 footage and he sees a lot of noise, everybody who can think rationally and logically will conclude that there's too much noise in the AF100 footage. The state of the monitor isn't relevant in that.
Click to expand...
And yet anyone who views the AF100 footage on a properly-calibrated monitor will know that the footage looks great and there isn't too much noise, in fact there's hardly any at all. And that's the bottom line. So if someone's setup is designed to magnify and amplify noise in the "invisible" sub-black region of the video signal, well, look -- what exactly are folks supposed to do about it?

If people want to think "rationally and logically" perhaps they should think through the entire production chain, and look at the footage in a proper environment, and realize that when used as intended, it's spectacular. So if there's something out of whack, it's not the footage's fault.

But we've had this debate before, let's not go down this path again.
 
A small correction: "too much noise" in my previous post should be "more noise than in other cams". And the amplification you write about Barry would amplify this extra amount of noise ("out of whack").

But the conclusion of the discussion was that the camera had not been black balanced properly, right?
 
"Nailing down a proper post workflow would indeed be helpful."

So I am wondering why you even proposed that in Post #79 Barry?

"proper post workflows" exist as you already know. Why propose something else?

I agree absolutely with you ..there is no point in judging anything requiring "professional" specs and requirements viewed on an uncalibrated monitor. Fact is peple are going to judge what they see displayed on whatever. They may one day realise and understand why they see , what they see.

There will be certain "users" of these technologies that will never understand a reason for it because their experience and understanding is limited to a computer Screen ( Nuff said).
 
This debate reminds me why SMPTE was established for most all crucial parameters, including the setting of background ambient lighting in the viewing space using calibrated instruments. Top quality work requires established metrics and calibrated tools using statistically valid sampling: http://www.smpte.org/standards/test_materials/

Maybe DVXUSER should establish its own humble standard for members to conform to which will act as a basis for comparing apples to apples instead of these endless and often fruitless exchanges.

Also, there are Spyder calibration tools for laptops, flat screens etc,. These tools can be helpful for setting up a proper system for viewing. Systems that do not conform are simply out of tune systems and should be not taken seriously for mastering work even if the public uses these sub standard systems. Optimizing your product for sub standard system (public) viewing is another conversation. Don't confuse the two.
 
Last edited:
grimepoch said:
Something is wrong with the Red channel on their card, and others have observed it as well. It is WAY over saturated.
Click to expand...

Which card are you talking speifically here? I use the latest decklink extreme card - and have been using blackmagic for a few years without this issue. I run HD-SDI out to a panasonic pro plasma and a JVC DTV CRT monitor and both calibrate perfectly to rec.709 without pulling this issue.
 
RieniO said:
But the conclusion of the discussion was that the camera had not been black balanced properly, right?
Click to expand...

No that isn't correct. I took the original timeline clip, put it back into FCP, and played it out to an HMR10 and recorded it. I then played that back on a 50 inch plasma and it looked great.

Thanks,

Jan
 
Noel Evans said:
Which card are you talking speifically here? I use the latest decklink extreme card - and have been using blackmagic for a few years without this issue. I run HD-SDI out to a panasonic pro plasma and a JVC DTV CRT monitor and both calibrate perfectly to rec.709 without pulling this issue.
Click to expand...

Intensity Pro. I was shocked by the red channel. Went online to do research and sure enough, MANY others with the same problem. Not to mention, the HMDI issues I have with it. EVERYTHING I own OTHER than my IntensityPro works great with my Toshiba flat panel. Definitely going to another card soon.
 
People have guessed at different reasons, so we really don't fully know. I think my Canon supresses the blacks which is why the noise goes away relatively speaking, but, to really know this I just have to wait for my AF100 and then use them at the same time. Soon, we can all get testing and find out the best flows and settings that produce the best results for all our different requirements.
 
Steve Kahn said:
That was the big disagreement with Barry ending up stating it was us, our monitors and not the camera. To many of us this made little sense because other footage we had seen never demonstrated this kind of noise.
Click to expand...

Actually it makes perfect sense. Why? Because THEY ARE DIFFERENT SHOTS.

Shot A, has really crushed blacks (thats where the exposure was to correctly expose whatever the subject was). So we have deep blacks and highlights on subject /s and other areas. This shot wont show any noise on any monitor.

Shot B, has an even gradient from black to white. On an uncalibrated monitor it will show noise, especially if the black (lets call it 0-30ire) on said monitor is being artificially lifted by said monitor.

Bottom line though is the AF100 has noise. Its no worse than XDcam EX - any flavour you like. Its definitely better than my 5DMKII. So we have a large sensor video style camera, that takes any lens you can imagine, all the bells and whistles that panasonic has in its arsenal, solves the DSLR aliasing, moire and mud issues. it costs under 5K and we want to pontificate over marginal noise?

I could understand this thread if we all had cameras and people reported noise under certain shooting conditions / scene file settings and the specific question was, how we can go about improving that? That makes sense. But at the moment to me this thread is a waste of time and effort.
 
Barry_Green said:
And yet anyone who views the AF100 footage on a properly-calibrated monitor will know that the footage looks great and there isn't too much noise, in fact there's hardly any at all. And that's the bottom line. So if someone's setup is designed to magnify and amplify noise in the "invisible" sub-black region of the video signal, well, look -- what exactly are folks supposed to do about it?

If people want to think "rationally and logically" perhaps they should think through the entire production chain, and look at the footage in a proper environment, and realize that when used as intended, it's spectacular. So if there's something out of whack, it's not the footage's fault.

But we've had this debate before, let's not go down this path again.
Click to expand...

This explanation isn't very logical. I understand where you're coming from, but footage is not intended to be watched by pros and calibrated monitors. If people have not been seeing noise in "footage A" and then all the sudden see noise in "footage B", then calibration, pro monitoring, ect, is all irrelevant. The problem is "footage B". How is that so hard for people to understand? All of this "please view on calibrated/proper monitors" needs to stop. That's not the point at all. People are going to watch this footage on all sorts of crap that nowhere near calibrated. If it's noisier than "footage A", it's going to be nosier than "footage A". ... Since any attempt to calibrate and provide an optimum environment for "footage A" will obviously also improve "footage B". This is all very basic logic. That being said, I haven't seen enough AF-100 footage to make a claim either way. I'm just speaking generally about comparing any two sources of footage.

No audience is going to "think logically and rationally" about the production chain when they're not satisfied with the image in front of them. Neither should we. At least not at the viewing stage. I think any talented director, cinematographer, editor, or producer has to be able to work/capture like a pro, but then view like a consumer. You can't justify you'r noisy footage based on the fact that you need a calibrated monitor and all your other excuses using technical jargon. You need to stop that noise from happening at the source. Just like you cannot justify a poor production value with a statement like "well, it's was a quick shoot, we only had two days to do everything ourselves..." Only someone with knowledge of the craft will be able to pick apart your "48 hour film festival submission" and see the talent in a quick production. To an uniformed audience, it just looks low-quality. So if I watch something on an ipad and it has too much noise... then guess what? It has too much noise. It's not my fault for not having a "pro" monitor... It's the camera, workflow, encoding, editing, coloring ect...
 
Last edited:
But at the moment to me this thread is a waste of time and effort.
Click to expand...

Let me put things another way. In a couple of weeks this camera is going to be released into the wild. Wouldn't it be good to have an Internet workflow that is optimal so things look as good as they possibly could?
 
Cbrianchester said:
So if I watch something on an ipad and it has too much noise... then guess what? It has too much noise.
Click to expand...

Agreed, we'll have to adapt to whatever our clients use for watching our videos, you can't expect clients to calibrate their iPods/iPads/laptops/LCD screens before watching our footage. So suggesting that we, the producers should calibrate our monitors, or worse, that those who see noise don't know what they are doing, doesn't help. If the AF100 has more noise than other cams, better just recognize that and come with a work-around, like crushing the blacks to zero a bit more. Then on the other hand, we are not talking about numerous clips showing noise, this is only about one particular clip, right?
 
You must log in or register to reply here.
Back
Top