4k Inspire footage problems: noise + moiré

[crazyshannon]

Hello all, I'm an assistant editor at a post-production studio. We have a television movie that keeps failing international QC because of problems with our quad footage. Every single shot (we used about five) has failed QC due to noise and moiré.

The shots in question are over a grassy field with sky in background. We gave up fixing the sky and ended up cropping it out, so now it's just grass. After many denoise attempts using different software, it's still failing.

The footage was shot on an Inspire 1, 4K at 60mb/s. We requested that the operator shoot 1080p @ 60mb/s as we thought this might prevent some of the problems we're having, perhaps this is not possible given the Inspire's camera settings, at any rate the operator shot 4K.

Anyway, I was wondering if anyone here has had problems getting Inspire footage to pass broadcast standards, and if there are any specific workflows unique to this camera type.

I'm also interested to hear any commentary on the Inspire's onboard camera vs gopro. My post supervisor seems to believe that the gopro camera would have been better, as we have gopro shots for non-quadrocopter footage that all passed QC fine (although our colorist is not a big fan of them). I'm wondering if that's not just a case of the post team being more familiar with the gopro vs the relatively unknown Inspire camera.

Any help is much appreciated, thanks.

 

[The Editor]

Hello all, I'm an assistant editor at a post-production studio. We have a television movie that keeps failing international QC because of problems with our quad footage. Every single shot (we used about five) has failed QC due to noise and moiré.

The shots in question are over a grassy field with sky in background. We gave up fixing the sky and ended up cropping it out, so now it's just grass. After many denoise attempts using different software, it's still failing.

The footage was shot on an Inspire 1, 4K at 60mb/s. We requested that the operator shoot 1080p @ 60mb/s as we thought this might prevent some of the problems we're having, perhaps this is not possible given the Inspire's camera settings, at any rate the operator shot 4K.

Anyway, I was wondering if anyone here has had problems getting Inspire footage to pass broadcast standards, and if there are any specific workflows unique to this camera type.

I'm also interested to hear any commentary on the Inspire's onboard camera vs gopro. My post supervisor seems to believe that the gopro camera would have been better, as we have gopro shots for non-quadrocopter footage that all passed QC fine (although our colorist is not a big fan of them). I'm wondering if that's not just a case of the post team being more familiar with the gopro vs the relatively unknown Inspire camera.

Any help is much appreciated, thanks.

Hi, and welcome to Inspire Pilots.
Unfortunately you have come across the major failing of the Inspire - Its camera!
First of all it isn't a 4k camera. Well, it pretends to be but at a bit rate topping out at 60mbps it can't really be called a serious contender for anything serious in the post production pipeline.
Add to that the dreadful GOP structure of the codec that causes intraframe pulsing every 8 frames and you are beginning to see the problem.
If you are intending to push out the footage interlaced, you will lose around 30% more resolution due to the the Kell factor and by that time you just give up.
Moiré is marginally better using 1080 but still a problem.
The best results are probably 1080/60p but still not stellar.
People have really just learnt to put up with the quality and use it for what it is - not exactly great to go through broadcast QC!
How are you ingesting the footage and what NLE are you using?
Give me your pipeline/workflow and I will try to help.

 

[ADSSER]

I didn't think the Inspire was ever intended to be a player in the high end field of movie production, was it?

At best, it's a mid-level platform for the "serious hobbyist" or similar. To get the quality you are looking for, it seems to me, you'd have to increase your budget by a few factors.

 

[crazyshannon]

Hi, thanks very much for your response! Not really the one I was hoping for, I was rather hoping for a magic solution

Our final conform is in Resolve. We've tried denoising primarily in Resolve and After Effects using different settings, and I think a denoiser from Boris. We have access to all NLEs if there's a particular plugin that works better with the Inspire issues - Avid, Premiere, FCP7, and X.

Footage is going interlaced at some point, but we're not the ones doing the conversion. We hand off to QC as 2k progressive.

Do you have any thoughts on GoPro vs the Inspire camera? Would the GoPro have been better? As I said we haven't had too many problems with our Gopro stuff, but it's not aerial and the camera operator may have used 1080p instead of the 4K option.

As for budget, well when you're on a low-budget made-for-tv movie, there's a lot of cost-cutting, and when you have two guys bidding and they both have "drones with cameras" and they both submit demo reels that showcase flashy flying shots that the PM is viewing on his laptop - and one's drastically cheaper than the other - well, then you get to where we are now. Post-production nightmare. Hah.

 

[Machoman]

The shots in question are over a grassy field with sky in background.

It uses the same Sony chip which Sony uses in the CHEAP compact cameras for years which are known for washed trees. So thats the quality you can get even worse in 4K with 30p instead of 60p and 60 Mbit/s instead of the 100 Mbit/s Sony uses in their own products.

 

[The Editor]

It uses the same Sony chip which Sony uses in the CHEAP compact cameras for years which are known for washed trees. So thats the quality you can get even worse in 4K with 30p instead of 60p and 60 Mbit/s instead of the 100 Mbit/s Sony uses in their own products.

If you are talking about the imaging sensor it has nothing to do with that and everything to do with how the raw image is debayered, processed and compressed using the codec, sampling rate and GOP structure.
The quality of the CMOS is actually quite good given it's size, it is whats going on behind the scenes that is the problem.

 

[Machoman]

If you are talking about the imaging sensor it has nothing to do with that and everything to do with how the raw image is debayered, processed and compressed using the codec, sampling rate and GOP structure.
The quality of the CMOS is actually quite good given it's size, it is whats going on behind the scenes that is the problem.

So you think the noise in the pictures and video has nothing to do with that its only the codec.

The opposite is the case people waited for years for small cameras with 1" sensor and using the same codec like cameras with smaller chips there is a big difference.

With no codec you get the noise out of a 1/2.3" chip the sony codecs reduce noise already very much at the cost of detail but more is not possible with this chip size.

 

[The Editor]

So you think the noise in the pictures and video has nothing to do with that its only the codec.

The opposite is the case people waited for years for small cameras with 1" sensor and using the same codec like cameras with smaller chips there is a big difference.

With no codec you get the noise out of a 1/2.3" chip the sony codecs reduce noise already very much at the cost of detail but more is not possible with this chip size.

Ahh yes of course!

Do you not think though it might have something to do with DJI's firmware conspiracy and how they are hiding things in their code and not telling people they are changing compression parameters? Maybe a lot of this can be put down to how DJI are gathering information on usage of their products and selling it to the Chinese Secret Service? I have wrapped tin foil round my X3 to make sure the signals don't leak out and someone doesn't pick them up on their iPhone and email it to the Chinese Government.

Its all very scary and I know they are out to get us!!!

 

[Firuzan]

Hi there ,
Does anyone have an idea about compass error caused by a lightning rod ? Well on top of my roof there is a lightning rod and when my inspire 1 gets close to it it causes compass error. Might that be the cause ?
Thanks

 

[m00se]

Hi there ,
Does anyone have an idea about compass error caused by a lightning rod ? Well on top of my roof there is a lightning rod and when my inspire 1 gets close to it it causes compass error. Might that be the cause ?
Thanks

it might not be a bad idea to ask this question in another place

 

[m00se]

Maybe a lot of this can be put down to how DJI are gathering information on usage of their products and selling it to the Chinese Secret Service?

It wouldnt surprise me if they actually do this. They do pull a few screengrabs of each flight when the ipad is connected to the internet. That pisses me off no end.

 

[Firuzan]

it might not be a bad idea to ask this question in another place

Why is that

 

[m00se]

Why is that

perhaps the help section would be a better place...

 

[Firuzan]

I asked this question in so many places no one seems to have an answer . Than I made a little research , turns out lightning rods carry a lot of static electric even if the weather is sunny and beautiful . That's a free information for you too

 

[m00se]

I asked this question in so many places

Not on here by the looks of it.

 

[The Editor]

I asked this question in so many places no one seems to have an answer . Than I made a little research , turns out lightning rods carry a lot of static electric even if the weather is sunny and beautiful . That's a free information for you too

You should have come here first - you would have got a sensible answer.
Although off topic (this thread is about noise/moiré) to answer your question, it wouldn't have anything to do with static since the magnetometer in the Inspire measures magnetic flux not static charge.
The likely cause of your issue is that lightning rods are usually magnetic or have a strong magnetic field due to the intense magnetic field that is induced during a lightning strike.
This field remains after the storm has gone so its basically a big spikey magnet on the roof.
I hope that answers your question.

 

[Firuzan]

You should have come here first - you would have got a sensible answer.
Although off topic (this thread is about noise/moiré) to answer your question, it wouldn't have anything to do with static since the magnetometer in the Inspire measures magnetic flux not static charge.
The likely cause of your issue is that lightning rods are usually magnetic or have a strong magnetic field due to the intense magnetic field that is induced during a lightning strike.
This field remains after the storm has gone so its basically a big spikey magnet on the roof.
I hope that answers your question.

Thank you what took you so long .

 

[The Editor]

Thank you what took you so long .

Oh yes I apologise - if you had posted a new thread with a descriptive title instead of burying off-topic related questions in a camera related thread I would have pounced like the caped crusader and answered it interfrastically!

 

[m00se]

Oh yes I apologise - if you had posted a new thread with a descriptive title instead of burying off-topic related questions in a camera related thread I would have pounced like the caped crusader and answered it interfrastically!

I personally saw a big bat projected onto clouds and realised something must be up.

 

[The Editor]

I personally saw a big bat projected onto clouds and realised something must be up.

Exactly - and the red phone was ringing on the commissioners desk!