Canon 7D Zone Plate Shots

[Darren Orange]

Hey Graeme,

Is it safe to say that the Panasonic GH1 does the same thing as the 7D?

 

[Graeme Nattress]

I don't know what the GH1 does precisely....But the aliasing is ugly. From what I could see though, it looked like the 1080p was almost like upsampled from the 720p as there were two aliasing patterns in the image. Resolution was low, and aliasing, especially vertical aliasing was ugly.

Graeme

 

[jbeale]

shooting 1080p

shooting 1080p

Very educational. In some respects, it looks like a claim of "shoots 1080p" on a box is becoming similar to the claim "records CD-quality sound".

If you know what you're doing, and are shooting the right kind of material (to avoid aliasing on detail, and jello effects), and making SD resolution or below deliverables, I think the 7D can produce attractive images. Seems pretty clear it can't be relied upon for every shooting situation.

 

[Tom Lowe]

The point of this is to make sure everyone knows exactly what they have... you can shoot with anything you want, including Super 8, 16mm, 35mm film. Those were pretty easy to understand. The difference was clear. Line-skipping and other digital tricks of the trade have made it easy to assume... maybe the wrong thing.

Jim

Jim, haven't you heard? Apparently it's no longer politically correct to actually analyze Canon's so-called "1080p" on its merits. If you dare to post zone plates or screen captures - or god forbid, side-by-sides with Red One - several things will happen:

1. You will be called an elitist.
2. You will be reminded of Canon's price. Again, and again, and again.
3. You will be lectured about "horses for courses."

Nevermind that all of us here already understand the prices, and already understand "horses for courses." Nevermind that most of us here already own and shoot on the Canons. Nevermind that some of us actually care deeply about image quality and like to analyze and pixel peep different cameras and formats. You will still be labeled an elitist and subjected to these tiresome lectures.

 

[Gunleik Groven]

Been over ' dpreview Tom?

 

[morgan_moore]

Graham..

Can you comment on this ..

http://maxmax.com/nikon_d200hr.htm

Showing filter removal of a D200 with increase moiree but better detail resolution

This is a demo of why testing with a zone plate MAY be a poor way of understanding how nice (or not) a camera will perform in the real world

In the zone plate test the filtered D200 looks better but in the real world test the De-Filtered D200 looks better

While I dont dispute that 7d is way worse than Red I am a little suspicious of relying on zone plates to prove that..

I have had two filterless stills cams Kodak SLRN and Hassy 22mp - both can be shown to be poor performers using a zone plate but are know to produce exceptional real world images that are on occasion bugged by moiree if you photograph something where the pixel pitch matches the repetetive frequency of the subject matter

SMM

 

[Gunleik Groven]

Hm... On most of the images (leaving "reading text" out)I sorta prefered the filtered.

I must have been infected by the RED juice...

 

[Roberto Lequeux]

I hope this is only coming up in such a direct way because of the whining mentioned above at other forums, and not because of a major sensitivity shortcoming of M-X when compared to the 7D/5D/1D. (all I want is close to 7D sensitivity and I'll be s-t-o-k-e-d)

Personally believe that it is just to disclaim some of the noise that is out there. To do so in such an objective and clear way is commendable.

I decided to look at my 7D as a 720p camera from day one, especially since whatever you get above 720p you will lose with the filter you'll need to get rid of most of the moire. But hey, I sold an HVX for it, and I am glad. It fits my personal-use needs and then some. Of course I would love a pro (clean) image, just can't justify it or afford it for personal use.

Tom, I don't think there is anything wrong with #2 though. Same thing for Red One vs. the other top shelf cameras, and with the upcoming Epic and Scarlet.

 

[morgan_moore]

Additionally I just saved a timelapse of a building a set (shot on my D3)

I downsized all the images to 1080 and saved them at jpg 8

The file size of the still images vary from 148k to 400k, the 148k images are when there was less detail in the whole image (because some of the lights were off on the set making half of the image black) the 400K images are where the house lights are on and we can see the seating in the the arena

This 'means' you can represent a simple image with a 148k file while needing 400k to represent a complex image

Now I understand little of compression but it does suggest that better performance MAY be attained on simpler images

ie the 7d may perform well on simple images (the classic interview head with blurred background)

My observations are that my 7 performs OK with a 50 at F2 on a head while looking crap on detailed wide shots which I am now using my EX1 for

Dependant on the method of compression used by the 7d it would therefore also be a valid test (maybe) to do some tests where a portion of the plate is obscured lowering the data required to represent the whole image

Again Im not denying the superiority of the Red but Im thining that to be truely fair in testing such things need to be understood

Because the DSLRs (particularlty the 5d) may actuallly be ok performers for some simple images while being appaling at capturing complex scenes of the nature of the helicopter shot previously linked

Emperical evidence back this up - there is nice looking footage out there from the cams - a zone plate is thier worst scenario and to view that in isolation can be misleading

I feel Im getting pretty good performance shooting with a 5d and EX1 by matching the camera to the shot

Of course I dont want to do this, I want a scarlet, but seeing a broader vision of how the cams work would be more useful than one test maybe

For example I feel that for some shots (1000ISO, simple image composition, 1080 final delivery) the 5d could be a better camera choice than a Red

That would be a useful undertanding even for red owners..

S

 

[Mark Toia]

So in short...

Great HD results $1500 and weighing in at 3 pounds +
Exceptional 4k results $50,000. weighing in at 20 pounds +++

We watch everything sent through the airways compressed to about 5 % of its true resolution. Most Movies we watch are mastered 2k then put out to film and in that transfer to film, the original sharpness is compromised again due to this duplication.

Image Quality Junkies like myself will go with RED.
Price per pixel junkies will go with Canon.

Easy math.

Moving on everyone. nothing to see here.

 

[morgan_moore]

Not so simple - if you want great HD you need and EX1 or something plus a canon at least unless you are going to seriously compromise the conditions you shoot under

Obvously you can skip the canon if you dont care about lowlight or DOF or ultrawides

You could skip the sony if you are just doing soft look art/indy

To me the question is can RED produce a cost effective solution (for decent HD not 4k) that will out do some canon/sony hotchpotch of compromises

I hope so

If you need 4k obviously your choices are restricted into the $50k+ region (for a useable kit) of course that kit may be usefully expanded with a canon for tight spaces and ultra dark situations in the same manner that kit can be expanded with a GoPro for POV

S

 

[Sanjin Jukic]

Not so simple - if you want great HD you need and EX1 or something plus a canon at least unless you are going to seriously compromise the conditions you shoot under

Obvously you can skip the canon if you dont care about lowlight or DOF or ultrawides

You could skip the sony if you are just doing soft look art/indy

To me the question is can RED produce a cost effective solution (for decent HD not 4k) that will out do some canon/sony hotchpotch of compromises

I hope so

If you need 4k obviously your choices are restricted into the $50k+ region (for a useable kit) of course that kit may be usefully expanded with a canon for tight spaces and ultra dark situations in the same manner that kit can be expanded with a GoPro for POV

S

Morgan,

but there is always possible a sort of solution (in solution).

Have a look carefully on this post

LINK>>>

 

[Roberto Lequeux]

Sanjin, just to make sure I understood your post correctly: You said that the HDMI is possible, just haven't been able to test it yet, correct?

I was going to check on that later, but if it is already a possibility I'd love to know for sure.

If so, I do believe that 720p from a 7D via HDMI will be something I'll be doing for a few night exteriors if we can't get M-X upgraded Red Ones in time.

 

[christopher witzke]

When the world seems to shine like you've had too much wine
That's a moiré

 

[David Rasberry]

Because the DSLRs (particularlty the 5d) may actuallly be ok performers for some simple images while being appaling at capturing complex scenes of the nature of the helicopter shot previously linked

Emperical evidence back this up - there is nice looking footage out there from the cams - a zone plate is thier worst scenario and to view that in isolation can be misleading

I feel Im getting pretty good performance shooting with a 5d and EX1 by matching the camera to the shot

Very good observation. A zone plate test at least helps define the limitations. A skilled shooter will take those limits into consideration when choosing the appropriate tool for a particular shot, or given the tool, plan shots to minimize the weaknesses.

 

[Sanjin Jukic]

Sanjin, just to make sure I understood your post correctly: You said that the HDMI is possible, just haven't been able to test it yet, correct?

I was going to check on that later, but if it is already a possibility I'd love to know for sure.

If so, I do believe that 720p from a 7D via HDMI will be something I'll be doing for a few night exteriors if we can't get M-X upgraded Red Ones in time.

Already there is

User Guide: Canon EOS 7D HDMI capture Version 1.0

by

Syndicate Entertainment AB www.syndicate.se

and they did it with brand new

Matrox MXO2 Mini.

"Introduction

The Canon EOS 7D is a great camera to film full HD material with. However,
the material you get on the Compact Flash memory card is compressed using H.264
at a bitrate that might be to low for certain scenarios and with 4:2:0 chroma encoding,
sometimes resulting in somewhat blocky primary colors.
To the rescue is the HDMI output of the camera. The HDMI signal is a uncompressed 1080i
signal at 60 Hz with 4:2:2 chroma encoding, giving a superior picture quality to the compact flash material.
However, to extract the real 25p or 24p image from the HDMI stream requires some special software.
This guide describes the process of capturing input footage so it is ready to process by the Syndicate 7D Tool.

What you'll get

Due to the nature of the HDMI output of the 7D the actual usable resolution of the uncompressed footage
will be 1620 x 910 at 24p or 25p with a reconstructed 4:4:4 chroma and a red dot in the corner.
The red dot can be replaced with the corresponding pixels in the compact flash recorded material in post production.
Future versions of the firmware modification Magic Lantern might support 7D and allow for removal of the red dot or perhaps even 1080p uncompressed.
The final output file will be a QuickTime with each frame stored with lossless compression, which means huge files.

What you need

To follow this guide to capture video from the Canon EOS 7D you need the following:
• Canon EOS 7D (duh!)
• Matrox MXO2 Mini
• A computer running Mac OS X 1.0.5 or later. Must be one of the following.
◦ MacBook Pro 15” / 17” with 2.66 Mhz minimum (faster recommended) and ExpressCard slot
◦ Mac Pro with free PCIe slot
• Final Cut Pro 6 or later
• A computer running Windows XP/Vista/7 with QuickTime 7.6 or later installed
• Syndicate 7D Tool Software (demo version available)
• Something to film"

 

[Lauri Kettunen]

Nevermind that all of us here already understand the prices, and already understand "horses for courses." Nevermind that most of us here already own and shoot on the Canons. Nevermind that some of us actually care deeply about image quality and like to analyze and pixel peep different cameras and formats.

Tom, I always like your comments and find this one also quite to the point.

Guess, everybody has their view and engineering is always about compromizes. As one colleague once said, everybody is able to make a bridge if they are given enough trucks, catepillars and stone and sand. But engineering skills is about making a cheap bridge that takes the planned load.

In the same spirit, one may say the line skipping is (imagewise) a poor choice but as well from Canon managers' point of view this may be (costwise) a miracle and the engineers behind it are considered as genius. And usually engineers respect other engineers as everybody knows there's always limitations that have to be taken into account.

The akward thing is that marketing wants to hide from the consumers what engineers have invented, and thus, consumers may get disappointed when they found that the marketing phrases do not reach reality.

 

[Sanjin Jukic]

Capture via BM Intensity Pro is still cheapest solution:

According to the latest discussions about Canon 7D HDMI out possibility

and image quality I contribute with one "photo simulation" about all process like this>>>

You could capture 7D HDMI out with BM Intensity Pro card where you can get something like this:

HD Format Support 1080i50, 1080i59.94, 720p50 and 720p59.94.
HDMI Video Sampling 4:2:2.
HDMI Color Precision 4:2:2.
HDMI Color Space YUV 4:2:2.
HDMI Audio Sampling Television standard sample rate of 48 kHz and 24 bit.
Card type PCI Express 1 lane, compatible with 1,4,8 and 16 lane PCIe slots.


So you feed your drive with 1080i YUV 4:2:2 footage from 7D.

Shot with Canon 7D with Leica Elmarit R Macro 60mm f/2.8 @ f/3.4, ISO 1600, shutter 125, 25 fps,
Canon 7D HDMI out on Panasonic 1080 plasma shot with Leica D-LUX 4.

Canon 7D HDMI out 1080 download>>>

Then in Apple Motion, Adobe AE, etc,... crop your footage at 720p and you get something like this:

Shot with Canon 7D with Leica Elmarit R Macro 60mm f/2.8 @ f/3.4, ISO 1600, shutter 125, 25 fps,
7D footage cropped from 1080 to 720.

Then 720p up-sample to 1080p 4:2:2, add a bit of sharpness and you could get a quality that can almost match with R1 footage down-sampled from 4K.

With this BM Intensity Pro capture solution for 7D you could pretty good intercut 7D and R1 footage without any H264 compression artifacts.

As you can see I still didn't try out this in real but that right now exists as an idea that I will do the test "realization" as soon as I can.

Shot on R1 with Leica Elmarit R Macro 60mm f/2.8 @ f/3.4, ISO 1600, shutter 1/50, 25 fps.



Below you get 1080p files from the "7D HDMI out by photo simulation" to compare with R1 1080p (4:2:2):

7D HDMI out 1080i (photo from 7D HDMI out to 1080i plasma) for download>>>

R1 shot with ISO 1600 4K down-sampled to 1080 for download>>>



One more thing:

Don't forget that Anthony Dod Mantle, BSC, DFF, British cinematographer, the ASC and Academy Award® Winner captured some shots on Slumdog Millionare

with SI-2K tethered to to MacBook Pro.

What does it mean?

You could do it almost a similar type of shooting with this kit:

Magma Express Pro Box with Black Magic Intensity Pro card inside

connected to MacBook Pro 17" using Magma ExpressCard/34 card and cable.

Lightweight higher ISO camera that can get 1080p 4:2:2 in post is in your hands just for $1700.

720p as a 1080i HDMI out to capture card BM Intensity Pro HDMI in, crop to 720p and then 720p upscaled and sharpened to 1080p is a way to go!!!

Even you could do the same for less money (just for $649.95) with Pentax K-x that is somehow intentionaly painted in red color

PENTAX K-x digital SLR in red

 

[Graeme Nattress]

Graham..

Can you comment on this ..

http://maxmax.com/nikon_d200hr.htm

Showing filter removal of a D200 with increase moiree but better detail resolution

This is a demo of why testing with a zone plate MAY be a poor way of understanding how nice (or not) a camera will perform in the real world

In the zone plate test the filtered D200 looks better but in the real world test the De-Filtered D200 looks better

While I dont dispute that 7d is way worse than Red I am a little suspicious of relying on zone plates to prove that..

I have had two filterless stills cams Kodak SLRN and Hassy 22mp - both can be shown to be poor performers using a zone plate but are know to produce exceptional real world images that are on occasion bugged by moiree if you photograph something where the pixel pitch matches the repetetive frequency of the subject matter

SMM

There's a vast difference between a camera that performs as a movie camera and one that does stills. If you get some moire in your shot it might be somewhat fixable, or you might just ignore it.

In a movie it's different because moire moves, and it moves in the opposite direction to the thing causing the movement. This not only adds to it's ability to distract the viewer, makes it nigh on impossible to effectively remove and screws up an motion compensated compression being used in the delivery chain, eating up bits from the bit-rate that you'd prefer to be using for actual image content, not aliases.

Lastly, for all the talk here about Canon, on the stills side their engineering is usually pretty darn good. Same at Nikon, Sony, etc. etc. and they all know that it's correct to put in an anti-alias filter, because the benefits do outweigh the losses.

Large format cameras tend to ignore the AA filter for a couple of reasons - 1, it's darn expensive, especially in the sizes they need. 2, the DOF with large format cameras is so thin, it's harder to get enough in focus in many situations to causes aliasing. Aliasing is an issue though, or Hassy would not be selling their Phocus software with magic moire removal tool (which I tried, though very clever, but it left the image with enough artifacting to be, IMHO, an unusable solution).

Of course, the one utterly correct answer to any aliasing issues is "more pixels please". (along with proper filtering).

Lastly, there's a vast order of magnitude difference between the levels of aliasing you get on a stills camera in stills mode that you may or may not remove with an OLPF, and the vastly greater levels of aliasing that we see in vDSLRs in video mode, which are often very visible without the aid of a diagnostic chart.

Another thing to note is the chart I use is made up or sinusoidal waves, no actual sharp square-wave edges at all. I have used square-wave charts and they will provoke aliasing even worse than my test chart does.

Graeme

 

[Sanjin Jukic]

Visible moire on 7D video shot (cropped and up-scaled detail).