[falconeye]

Maybe the word moiré is wrong here ... I doubted that as well

It is called moiré, I call it color moiré to make it distinct from normal moiré. Normal moiré is a sampling aliasing artefact in the luminance channel, color moiré is a sampling aliasing artefact in the chrominance channels.

In this particular case, it is caused by line skipping across a Bayer sensor. The effect is a bit more pronounced than color moiré in a still image from a sensor lacking a Bayer AA filter. Due to line skipping in video, it is almost irrelevant if the sensor has a Bayer AA filter or not.

Still image color moiré locks like a low frequency pattern in regular areas but just like false colors in tiny patches (think of a tiny patch cut out of a larger color moiré pattern). Condition to provoque the pattern is to trigger a very distinct spatial frequency. In the sample image, the vertical bars do just that. If they were twice as thick, the colors would disappear. Examples like this can be used to reverse engineer the sensor readout scheme in video.

---

Last edited by falconeye; 11-13-2013 at 12:09 PM.

[TomGarn]

As a 'pro' Videographer, or Cinematographer, and Ex-Broadcast Camera and Live Director, let me address some points ...

Well said ... but some points need second sight maybe:

1. Like many other blazed professionals you just care about "Canikon" as competitors "above" Pentax.
You seem to have overlooked the advanced quality of Pansonic Lumix GH2 and GH3 completely ?
They have bitrates at 50 and 72 Mbit nowadays ... and they have also taken over Pentax famous
weatherproofed body as well and they are affordable as well compared to fullframe DSLRs

2. In all tv-broadcast camcorders from Betacam to XDCam422 I had audio inputs. I don't get your
point at all. You either insert it via automatic controle, or your sound assistant controles it, but always
via input to the camera ... I am talking about all kinds of broadcast ENG-cameras here.

3. Rigs do look very "professional" ... but to me they are weird, because I need to touch my camera
directly when it needs adjusting - same for broadcast-cameras as for DSLRs. Handles and grips are
distracting you from quick access.
Only in movies and stuff like that, when the cameramen seem to know exactly what scene gets done
and how - the camera-crew will have enough time to fix settings before shooting - or you even have
a (human) focuspuller assisting you - or the boys in the controle room of course, otherwise you will
have to do your settings on your own.
So:
SR as in old Pentaxes was great - since K-30 (at least) it's just on the level of point and shootcameras.
Good SR is a good thing for photos anyway in low light ... For video it helps a lot because of the low
weight and the uncomfortable handling - compared to well balanced schoulder camcorders. It helps
immensely to neutralize micro-shakings. So franklly said, there is no need at all to talk that away in a
blazed way.

[TomGarn]

It is called moiré, I call it color moiré to make it distinct from normal moiré. Normal moiré is a sampling aliasing artefact in the luminance channel, color moiré is a sampling aliasing artefact in the chrominance channels.

In this particular case, it is caused by line skipping across a Bayer sensor. The effect is a bit more pronounced than color moiré in a still image from a sensor lacking a Bayer AA filter. Due to line skipping in video, it is almost irrelevant if the sensor has a Bayer AA filter or not.

Still image color moiré locks like a low frequency pattern in regular areas but just like false colors in tiny patches (think of a tiny patch cut out of a larger color moiré pattern). Condition to provoque the pattern is to trigger a very distinct spatial frequency. In the sample image, the vertical bars do just that. If they were twice as thick, the colors would disappear. Examples like this can be used to reverse engineer the sensor readout scheme in video.

Thanks for talking on that falconeye.
You did such a good job with your K-7 footage on that problem,
It's a shame Pentax didn't pay you or even heard about that at all ...
(Of course I don't see through the technical ways of configurations at all here)

[johnmflores]

The time limit is purely there because the camera us SD cards formatted in FAT32, which has a maximum file size limitation. If any of these cameras could run ExFAT, there would no longer be a file size limitation, and then the camera's could record until the cards space was filled.

FWIW, the GH2 can record for well over an hour. I've done 90 minutes. AFAIK, it rights multiple files to the SD card to get around the FAT32 limit, and then FCPX automagically puts them back together on import.

My understanding of the time limits is related to different tariff levels for still vs. video cameras and issues of overheating.

[tripodquest]

if anything, the sdcard should have an ext2 or ext4 filesystem, where filesize limit is some terabytes or exabytes (thousands or billions of times more than fat32) but Asia tends to do things the Windows way...

It is called moiré, I call it color moiré to make it distinct from normal moiré. Normal moiré is a sampling aliasing artefact in the luminance channel, color moiré is a sampling aliasing artefact in the chrominance channels.

In this particular case, it is caused by line skipping across a Bayer sensor. The effect is a bit more pronounced than color moiré in a still image from a sensor lacking a Bayer AA filter. Due to line skipping in video, it is almost irrelevant if the sensor has a Bayer AA filter or not.

Still image color moiré locks like a low frequency pattern in regular areas but just like false colors in tiny patches (think of a tiny patch cut out of a larger color moiré pattern). Condition to provoque the pattern is to trigger a very distinct spatial frequency. In the sample image, the vertical bars do just that. If they were twice as thick, the colors would disappear. Examples like this can be used to reverse engineer the sensor readout scheme in video.

I had no idea of this issue, thank you for explaining that! I've heard of color moire but i thought the example above was just a lens issue.

[tripodquest]

FWIW, the GH2 can record for well over an hour. I've done 90 minutes. AFAIK, it rights multiple files to the SD card to get around the FAT32 limit, and then FCPX automagically puts them back together on import.

My understanding of the time limits is related to different tariff levels for still vs. video cameras and issues of overheating.

is gh2 the "best" model from panasonic?

[johnmflores]

is gh2 the "best" model from panasonic?

No. It's been superseded by the GH3, and the even newer GX7 produces better video than both but is less capable from a workflow standpoint, i.e., no audio monitoring, etc...

[tripodquest]

No. It's been superseded by the GH3, and the even newer GX7 produces better video than both but is less capable from a workflow standpoint, i.e., no audio monitoring, etc...

I always use an external microhpone & external recorder with real time visual audio monitoring, autogain option etc... and sync in post so that's not an issue for me if you don't mind two separate devices, you'll have the headphones on your head anyway no matter where the cord goes and arms allow you to attach your external devices to your rig at a hand's reach but like i said in the other thread i just can't get behind the idea of shooting without a viewfinder... nothing beats the natural feel of the viewfinder against your eye, looking through the lens, with the option of looking at the monitor when you need to should i learn to live without a viewfinder?

[PiDicus Rex]

I call it color moiré to make it distinct from normal moiré.

That actually makes a signifigant amount of sense to me -> I've always understood 'fringing' to be a result of the De-Bayering, and Colour Moire seems the best description of where the wrong colours are represented in the body of an object, not the edges.

John: I thought the tariff was a Europe only thing. Do they tax cameras differently in the US as well?
Multiple files, yes, that's what harddrive and SD card based video cameras do.

if anything, the sdcard should have an ext2 or ext4 filesystem, where filesize limit is some terabytes or exabytes (thousands or billions of times more than fat32) but Asia tends to do things the Windows way...

ExFat is open format, not Windows, and can be read and written by Linux, PC and Mac, majority of PC's can't do Ext.

You seem to have overlooked the advanced quality of Pansonic Lumix GH2 and GH3 completely ?

I'm in Melbourne, I know a grand total of One person who shoots on GH2, and he brought that back to Oz from New Mexico. It's not a matter of ignoring them, it's a matter of just not ever seeing them anywhere, not even on display in major stores.
My favorite full sized cameras are the HPX-502 and VariCam 3700.

2. In all tv-broadcast .... ... I am talking about all kinds of broadcast ENG-cameras here.

I assume you're talking about recording sound for Film and TV Production?
ENG ( Electronic News Gathering ) set-ups are done in camera, yes. Any production done on DSLR, Red, Alexa, ie, large sensor cameras, the audio is always recorded separately, by an Audio specialist, usually on equipment capable of recording 4 or 8 channels, at 48, 96kHz sample rate, or even higher in some cases.
The only audio recorded on the camera is the guide track to make syncing the audio in post easier.

Rigs do look very "professional" ... but to me they are weird, because I need to touch my camera
directly when it needs adjusting - same for broadcast-cameras as for DSLRs. Handles and grips are
distracting you from quick access.

Only if they are badly designed. A decent rig makes camera operation easier, faster, and more stable. But I do get where the user frustration lies, one mate I work with regularly hates it when I strip half his rig of his camera before shooting because it gets in my way.

Wasn't meant to be 'blazed', just direct statement of the way it is. AF and SR are just not used or wanted in Pro environments.

[Tanzer]

Interesting points-of-view regarding what coulda/shoulda been done. We will never resolve these debates here, since our video needs are likely to be very different, from one person to the next.

As Pentaxians, we are not likely to argue too much that a K-5/K-30/K-3/K-01/Q/whatever meets most of our still photography needs. We are here as a self-selecting population, after all.

But for video, well ... some of us are (or think that they are) filmmakers, some want to shoot sharp video like the evening news, some are just looking for more flexibility/creativity than a typical consumer video camera, some want to leverage their existing collection of Pentax glass, and some simply don't want to carry a separate still camera and video camera. (I'm in every category except the first one.) Anyhow, there just isn't any one solution today that meets all these needs.

[tripodquest]

ENG ( Electronic News Gathering ) set-ups are done in camera, yes. Any production done on DSLR, Red, Alexa, ie, large sensor cameras, the audio is always recorded separately, by an Audio specialist, usually on equipment capable of recording 4 or 8 channels, at 48, 96kHz sample rate, or even higher in some cases.
The only audio recorded on the camera is the guide track to make syncing the audio in post easier.
.

I agree, you only need the in-camera sound track for synching you can't have the same kind of audio in-camera as with an external setup. I have 5 external mics for front, back, and middle/horizontal audio covering 360 degrees, capable of recording true 360' audio as well as audio in any angle configuration from 0' to 360', using 2, 3, 4 or 5 mics. 2 or 4 channels, stereo, mono, up to 96khz, PCM/wav or mp3, raw or compressed with real-time visual monitoring, auto-gain function, low cut, limiter (DRC) with saved settings for different musical instruments, a dedicated (physical) gain dial and so on. I doubt any in-camera audio can do that? Correct me if I'm wrong...

[tripodquest]

Anyhow, there just isn't any one solution today that meets all these needs.

seems that way. i was sure technology would have come far enough for that to be reality, but doesn't seem that way... oh well if you need to buy 2 cameras to get good photos and good video, then you need to buy 2 cameras, what can you do.

[johnmflores]

I always use an external microhpone & external recorder with real time visual audio monitoring, autogain option etc... and sync in post so that's not an issue for me if you don't mind two separate devices, you'll have the headphones on your head anyway no matter where the cord goes and arms allow you to attach your external devices to your rig at a hand's reach but like i said in the other thread i just can't get behind the idea of shooting without a viewfinder... nothing beats the natural feel of the viewfinder against your eye, looking through the lens, with the option of looking at the monitor when you need to should i learn to live without a viewfinder?

Yes separate audio and syncing in post has become much easier with the latest digital tools. I'm often a single person crew though, and seek simplicity where possible.

You're in luck as the GX7 has a built on viewfinder. No IS though, for video at least.

[kadajawi]

Yes, one of my friends has a GH3. That produces pretty nice footage.

...

I agree with the AF. I don't need it, and I'm not doing professional work. However I have tried focus peaking on the Sony a57, and oh my god does it help. Now if you use an HDMI monitor with focus peaking, good for you. If you don't, Pentax leaves you stranded. I suppose the CPU they used just doesn't have the processing power to do h264 AND focus peaking... if only it did MJPEG, maybe then focus peaking were possible? Then again, the encoder is typically taking a seperate portion of the chip for these embedded processors. Hm.

As for SR... try the K-5 WITH SR on. You don't get jello or any other artefacts, you simply get a super stable video (depending on how stable you hold it of course). The reason is simple: What the sensor sees is not moving. It is as if you are using optical image stabilization, it is that good, it has the same effect (though OIS may give you a greater range of movement). That is why we beg Pentax to give us back non-electronic SR, as the K-7 and K-5 had. I don't see any reason why they wouldn't do it, when those old cameras could, and the new one can't. Ideally they would give us the option to select what we actually want. No SR, sensor based SR and electronic SR. And maybe some options for the sensor based system, like how fast it is going to move the sensor to the new position, or if it is trying to hold the frame as long as it can.

Btw., my K-5 sensor does overheat, it gets hot pixels after prolonged usage (warm weather of course greatly accelerates the problem).

Time limit can also come from the intention of selling the camera in Europe. Import tarifs are higher when the camera can record video for more than... 30 minutes or so at a time.

The alpha 57 does 50i at 1080 AND 50p at 1080. Though the bitrate at 50p is a bit low (28 Mbps, 50i 24 Mbps, 25p 24 Mbps). Still, from what I gather here still better than the K-3... and that is with Sony's probably superior encoder.

Btw., if the h264 stream is only consisting of i frames... in what sense is it better than MJPEG, apart from maybe a slightly more efficient i frame encoder?

I don't think Pentax needs to have the ability of shooting ProRes, CinemaDNG, 4K etc. in camera, as that could drive up prices. But say, a battery grip? Something making use of that USB 3.0 port, that includes an 2.5" slot for SSDs for example?

I believe some new Olympus have in body SR for video too, perhaps the OM-D ones?

[osv]

First up, AF and SR are irrelevant to 99% of professional work, and are the second and third items turned off on any camera.

finally! someone with a video background.

Lastly, on the subject of MJPEG. It is not an 'obsolete' CoDec, in fact, everything you watch on TV, shot and broadcast in SD or HD, has been passed through the broadcast system as MJPEG.
MJPEG is the worldwide standard for sending TV signals anywhere and everywhere, and it's not until the signal reaches the transmitter that it becomes the 720x576 (PAL) or 1440x1080 (HD) Mpeg2 or Mpeg4 that you receive and watch at home.

that comment is at least 10 years behind the times.

in the real world, mjpeg is not used for acquisition, editing, or final distribution... whether it may or may not be used in some small portion of a pre-distribution chain is irrelevant to this discussion about acquisition.

I used to edit with the mjpeg codec, back when the holy grail of acquisition was uncompressed data... remember when betacam sp ruled? I still have the tapes to prove it, lol