Slashdot Mirror
Open Source Brings High-End Canon Camera Dynamic Range Closer To Nikon's
PainMeds writes "Magic Lantern is an open source 'free software add-on' that 'adds a host of new features to Canon EOS cameras that weren't included from the factory by Canon.' One of ML's newest features is a module named Dual ISO, which takes advantage of the sensor in some of Canon's high-end cameras (such as the 5D MK II and MK III) to allow the camera to capture an image in two different ISOs, greatly expanding the dynamic range of the camera, and bringing its dynamic range closer to Nikon's popular D800 and D4."
Magic Lantern is fucking awesome. It turned my Rebel T2i (550D) into something that I definitely wouldn't have been able to afford. If you own a Canon DSLR, you owe it to yourself to give it a try. You'll be amazed, confused, and then even more amazed.
That being said, the cameras mentioned in the summary cost more than all my lenses combined. Sounds like an awesome feature, but not one that will be available to casual photographers.
Chuuch. Preach. Tabernacle.
Alternate line exposure is not new, it is in a lot of current generation sensors. Omnivision, Sony and Toshiba all have sensors out with this capability.
The underlying issue is that when doing alternate line exposure you are getting only half the resolution for each range. DSP and image processing techniques can help smooth out the issues, but you are fundamentally dealing with a half-height dark and a half-height light image. Depending on the alternate-line approach, you also get other funky color fringing issues due to the underlying bayer pattern. As the article notes, there are color fringing issues
A good generalized approach is to output a 1/2 resolution image in both dimensions, otherwise you will get a vertical stretch if you keep the horizontal width at full resolution. So it means for a 16 MP camera, you will get only 4 MP HDR images. In a lot of cases this will be more than good enough... But it makes it really difficult to sell and explain to users.
There is usually a good reason that advanced features aren't release/published. A lot of the time it comes down to features be sub-optimal on what is supposed to be a highly polished product.
Isn't Magic Lantern the name of an FBI spyware program designed to snarf passwords from suspects?
If I can be modded down for being a troll, can I be modded up for being an orc, or a balrog?
It could of been on purpose. Consider how Pigdog DeCSS (a program that removes style markup from HTML) was named after the DeCSS program for DVD-Video.
I was a Canon man. But they have seriously dropped the ball in the image sensor department. They continue to use ancient sensors that simply cannot compete in today's market. I sold all my Canon cameras and lenses and moved to Nikon and could not be happier. I do not have to resort to hacks to get an image that is close to a Nikon image.
"One of ML's newest features is a module named Dual ISO, which takes advantage of the sensor in some of Canon's high-end cameras (such as the 5D MK II and MK III)"
Nope. Dual ISO works on *most* Canon DSLRs, not just the 5DII/5DIII. More info here. Technical paper from the developer here.
Oh yeah, Magic Lantern is one of the most amazing and useful pieces of software I've ever come across. If you own a Canon DSLR, you owe it to yourself to check it out.
I really like the feature list it brings to the table and have thought about trying it out with my T4i for a while now. One question that isn't clear though, is there a "simple mode" for those times that I just want to pick the camera up and just go shooting or recording with basic automatic settings? I know there are times that I would like/need the features that Magic Lantern exposes, but I also know that I would miss shots or video moments if I had to configure 14 settings every time I wanted to shoot something. Can a current ML user educate me here?
Nikon is still a step ahead! First post!
My crappy old Nikon D70s, consigned to the junk heap, says they could do better on build quality and components.
A feeling of having made the same mistake before: Deja Foobar
Magic Lantern was the name of the first image projection system which was developed in the 18th century. This firmware was originally developed in order to exploit some of the untapped capability of the 5D2 video system, and since has expanded into enhancing still features.
Brilliant work overall.
So what does the article try to mean?
Dynamic Range is not same thing as Exposure Range.
Exposure Range is calculated in stops (EV) and it tells what can be difference between darkest and brightest parts of image where any data can be recorded, example D800 gives with a base ISO about 14.5 stops of exposure range. What means your landscape image can have details in bright clouds and then shadows under trees. Olympus E-M1 camera has 13.8 stops exposure range in base ISO. That is only 2/3 stops lower exposure range than what Nikon top model offers.
Dynamic Range tells how much information can A/D processor detect from sensor in specific energy consumption and it can be higher or lower but not result to lower or better Exposure Range. Dynamic Range informs more about Linear Resolution of the sensor, how much data you can pull and push from brightest and darkest part, but it isn't still same thing as Exposure Range is.
It is like people believe that F-stops is the light gathering capability meter. While it is only meter to light passing trough aperture, while T-stops is the meter to light gathering. The T-stop value is always lower than F-stop, meaning example F/1.4 lens has about T/1.3-1.35.
Nikon is still a step ahead! First post!
My crappy old Nikon D70s, consigned to the junk heap, says they could do better on build quality and components.
Consumer line = consumer quality.
Should have opted for the D700.
Two different ISOs?
Hasn't one International Standards Organization caused enough design-by-committee standards harm? Do we really need another?
"National Security is the chief cause of national insecurity." - Celine's First Law
A shootout of a 24MP APS-C sensor on a Pentax K-3 against a 24MP full frame Nikon D600.
"We figured the Pentax would do a good job, but we never imagined the results would be so outstanding. Our testing shows that the Pentax K-3 swept the Nikon D600 in almost every image we took. Even at high ISOs the Pentax held its own against the full frame sensor!"
http://www.digitalcamerareview...
"This test clearly shows the full frame Nikon D600 does not have image quality that is nearly as appealing or accurate as the cropped sensor Pentax K-3. Plus, the Pentax K-3 is currently $500 less expensive than the Nikon D600."
Maybe they need to keep updating the software.
Procrastination; I'll think of a sig tomorrow.
While I thought the results were cool, I was annoyed by the bloggers use of the word 'chromatic aberration' instead of color noise.
Chromatic aberration means the lenses bend different colors of light differently resulting in color fringes around the edges of object. Color noise which is observed in low-light conditions here is not an aberration effect of the lens, but pixel counting noise on the CMOS detector.
Are as good as a quality film camera loaded with quality film. Maybe one day..
Hey KID! Yeah you, get the fuck off my lawn!
Nikon is still a step ahead! First post!
My crappy old Nikon D70s, consigned to the junk heap, says they could do better on build quality and components.
Consumer line = consumer quality.
Should have opted for the D700.
Odd that. I could drive nails into a board with my old FE II and it would still work like a Swiss watch. It's been banged up and struck rocks while hopping to a better vantage (and slipping) and all I ever lost as a lens cap. Consumer end back then was still quality. Now consumer is an excuse to use cheap parts. So much fail there.
A feeling of having made the same mistake before: Deja Foobar
I dunno, I abuse my D90 pretty badly, and it just keeps on working perfectly. Maybe I'm just lucky. Or unlucky - if I could actually break it my wife wouldn't mind me buying a new body.
Firstly this kind of technique can be applied in post processing with better results (not wanting to advertise but for example using photomatix ) than can be achieved in the camera. In post processing this technique can be applied equally to Canon or Nikon or Sony or Panasonic (pick your favourite manufacturer) images, and really the only reason it hasn't reached Nikon/Sony/others in camera yet is that there isn't such a big firmware modding community following with Nikon/Sony/others. So Canon certainly has that advantage (a bigger modding community) - (but then again as the meme goes; "Canon is the camera designed by engineers for engineers and Nikon the camera designed by photographers for photographers").
OK - so now try using this technique to take pictures of objects that are moving; you instantly see the benefit of a sensor with inbuilt increased dynamic range. HDR photography relies on multiple exposures being combined together. This means that they are forced to be static, either that or they suffer from side effects such as ghosting , where anything moving across the image will be present in one of the exposures - and not necessarily in the same place in the other, leading to some weird looking end results. Alignment of the image is another issue - you need the camera on a tripod to make this work really well.
So really this technique does not replace a superior
A shame I cannot edit (on new slashdot) to correct my own posts - as a few words appear truncated...
Your single source is a review of subjective qualities *perceived* in JPG conversions from RAW files made by lightroom. All the review is about is what base curves and algorithms LightRoom applies to the RAW files, not about the actual quality of the RAW files itself, or the ability to make a decent image of said file with manual adjustments. Analogy: you are comparing the quality of JPG images an automatic scanner software generates from two different brands of 35mm film in a film scanner. These cameras have different requirements for making adjustments to things like colour balance, saturation, sharpening and such to get an optimal image output.
As soon as images get rated on how "pleasing" they are, I'm out when it comes to a *technical* review. Give me objective measurable repeatable tests and results and I'm willing to take you seriously, but not this.
I was promised a flying car. Where is my flying car?
We are indeed listening and making changes based on feedback, and at a faster pace now that most of the underlying engineering work is done. We'll put up another post in a week or two explaining everything we've changed.
In the meantime, if you want evidence, here's one example: load up the beta and look at the comments on one of the stories. One of the biggest complaints was that the comments field was too narrow because of the right rail (which was empty once you scrolled down a bit anyway). So we removed the right rail entirely on story pages, letting comments flow all the way across the page.
Thanks for the update, now that fix so we can actually return to an earlier post and engage in.....discussion!
"The likes of Facebook and WhatsApp are free to those whose privacy is of zero value."
As you observed the mechanical parts are fine, it's the electronics that need time to stabilise and transfer the picture.
"The likes of Facebook and WhatsApp are free to those whose privacy is of zero value."
There's a Preview and then an Edit button, just like the ol' ways :)...
"The likes of Facebook and WhatsApp are free to those whose privacy is of zero value."