m4/3 sensor technical specification

Discussion in 'This or That? (MFT only)' started by st3v4nt, Jun 25, 2011.

  1. st3v4nt

    st3v4nt Mu-43 Veteran

    317
    May 26, 2011
    Jakarta, Indonesia
    Friend of mine an avid user of Nikon system ask me the question about 4/3 and m4/3 sensor specification. While I can answer some of it, it's a bit difficult to answer specific things like RAW compression bits, pixel density and so on. He also ask the question about why many picture example in four-third.org have a bit loss of gradation in highlight part. I know that Olympus use 12-bit RAW compression but how about Panasonic what kind of raw compression did they use? Also I have read in other forum that m4/3 tough use smaller sensor size have more pixel density compared to some APS (entry level) sensor. They even have some sort formula to count the pixel density? Anyone can clarify this?
     
  2. mauve

    mauve Mu-43 Top Veteran

    892
    Mar 9, 2010
    Paris, France
    On the specific subject of gradation, there's a firmware parameter on Olympus pens that let the user select between 'auto', 'normal', 'high key' and 'low key' depending on which zone (as in Ansel Adams' zone system) you want the best optimization. So in effect, many samples on the 'net can appear suboptimal in highlights or lowlights depending on the photographer's personal settings. It doesn't mean much about true sensor capabilities.

    The other questions may be of interest to people who prefer to brag about gear rather than taking pictures. They are mostly irrelevant to :43: users. When you come to :43: system, you know there are trade-offs. What you get is 'almost' DSLR quality in half the size and a quarter of the price of a full blown pro DSLR system. So pat your friend in the back, tell him kindly he has already chosen the best setup available and be done with it.
     
    • Like Like x 1
  3. Hikari

    Hikari Mu-43 All-Pro

    Nov 26, 2010
    Sounds like a camera snob. I remember talking to our camera reps about the camera shows they would attend. They said owners of different brands would come over and question them about the specs of our cameras, in particular about things we did not have but their model did. Our reps would just smile and tell them what a great camera they had bought
     
    • Like Like x 4
  4. Ned

    Ned Mu-43 Legend

    Jul 18, 2010
    Alberta, Canada
    Not really. The :43: sensor is merely cropped squarer than an APS-C sensor, which is wider on the sides. Since you only require the full resolution of the image in print (pixel resolution is irrelevant for web use), that means that you will need to crop your images for print stock. All standard print stock uses a squarer aspect ratio than APS-C, meaning a LOT of the image is cropped off the side for print and thrown away. The :43: sensor is already cropped to an aspect ratio more relevant to the print industry.

    What that means, for instance, is that a 14MP sensor on an APS-C sensor only produces 12MP out of the actual 5:4 or 4:3 center area that will be used in printing. A 12MP :43: sensor will use almost the entire 12MP captured to print with.

    When you equate that to sensor size, that means that the portion of the sensor which will actually be used to capture the part of the image that is kept, is nearly identical in size with :43: and APS-C. APS-C is not really "bigger", it's just wider. It captures extra stuff on the side you don't even need, making the technical specifications seem larger but giving no real-world benefit at all.

    Full Frame on the other hand, is actually a bigger sensor in both dimensions.
     
    • Like Like x 4
  5. Grant

    Grant Mu-43 Veteran

    Most of this information is academic for me as I have both systems but out of curiosity I would like to know the bit depth for the GF1.

    I do have a Nikon that has both 12 and 14 bit depth and I must say that even at 12 bit the D300 out preforms my GF1 in certain areas. Because of this I do all my single shot HDR using my Nikon, in fact I do all multi exposure HDR with my Nikon but there are other reasons for that.

    Please don't take this as a slag against my GF1 as it is a wonderful camera, just there are different tools for different jobs.
     
  6. Hikari

    Hikari Mu-43 All-Pro

    Nov 26, 2010
    Bit depth has nothing to do with dynamic range. DR is related, but not limited to, pixel size. Bit depth just defines the numbers of luminance levels that DR is divided into.

    And a m4/3 sensor is not a crop of an APS-C sensor, it is both narrower and shorter than APS-C.
     
    • Like Like x 2
  7. mauve

    mauve Mu-43 Top Veteran

    892
    Mar 9, 2010
    Paris, France
    The GF1 sensor is the same one Olympus uses, with a 12 bit depth by color channel.

    I have no problem with people choosing the right tool for the job at hand, but "I fear men of a single book", and when I read 'avid Nikon user' instead of photographer, I fear the worst. Hence my answer.

    Cheers,
     
    • Like Like x 1
  8. RDM

    RDM Mu-43 All-Pro

    i would tell your friend not to worry about those things unless he really wants to buy a ยต4/3 camera, and to just go have fun with his Nikon.
     
    • Like Like x 1
  9. Hikari

    Hikari Mu-43 All-Pro

    Nov 26, 2010
    8 bit images have 256 luminance levels per color channel
    12 bit images have 4096 levels
    14 bit images have 16384 levels
    16 bit images have 65536 levels.

    8 bit is the magic number because for the human visual system to see a stepless gradient go from black to white, it take approximately 200 levels of gray/grey. The higher bit depth allows you to make greater changes during processing and not introduce banding--the point where there are too few luminance levels to define a smooth gradient causing bands.

    What the practical difference between the other bit depths, I don't know. 4000 levels of gray is a lot of information. The gradients that you want to define would have to be extremely low contrast for a difference to be seen.
     
    • Like Like x 3
  10. Ned

    Ned Mu-43 Legend

    Jul 18, 2010
    Alberta, Canada
    The difference in height between a Four-Thirds sensor and a Canon APS-C sensor is about the same as the variance between a Canon APS-C sensor and a Nikon APS-C sensor (the Nikon is slightly larger). I'm not about to argue about the difference a mm in size makes, lol. That's why I used terms like "nearly" and "almost" in my post, because there's always bound to be some nitpickers around. :rofl:

    I think differences in build and technology can more than make up for a mm in size.
     
    • Like Like x 1
  11. Hikari

    Hikari Mu-43 All-Pro

    Nov 26, 2010
    The difference is twice as much: 15.7mm, 14.8mm, 13mm...

    I am a m4/3 user so obviously I am not trolling. There is a difference. Size does matter, but then many things matter. I am just not sure I can say they are almost the same size...
     
  12. Ned

    Ned Mu-43 Legend

    Jul 18, 2010
    Alberta, Canada
    Okay, how many people even know that there is a difference between Canon and Nikon APS-C sizes? If the difference is that small, then twice that difference is still hardly anything.

    I've never once heard anybody say that Canon APS-C is a worse performer than Nikon, Pentax, and Sony APS-C sensors because Canon's is so much smaller. So why do people say exactly that when comparing Four-Thirds with APS-C? It just sounds like something stupid to pick at, if you ask me.

    Mind you, sometimes Nikon, Sony, and Pentax sensors are reputed to have better ISO sensitivity and Dynamic Range than similar Canon sensors. Perhaps we could blame that on the 0.9mm size difference, lol? I'd love to see what would happen if that subject came up on a Canon/Nikon forum, and somebody answered with, "That's because the Canon APS-C specification is smaller than everybody else's". I'm sure they would be instantly shot down by the same people who shoot down Four-Thirds under the same pretense.

    Now the improvements made to our Four-Thirds sensor performance over the years have actually been tremendous. That improvement has nothing to do with size, as they have been made with the same sized sensors and in many cases the exact same physical sensor build. APS-C sensors also continue to grow and change in performance, and shaving off 1.8 mm on one end will certainly not make it suddenly turn into a poor performer.

    I think our sensors were behind the competition for a while in various ways... and people were blaming that 1.8mm on technological limitations which are totally irrelevant to sensor size. I guess they didn't want to admit that the roadblock was something which could easily be put out of the way, not a limitation put on by our hard specifications. :) Performance of recent Four-Thirds and Micro Four-Thirds sensors however, have more than caught up to its similar sized cousins in APS-C. Making a jump from Four-Thirds or APS-C sized to Full Frame however, gains various advantages and disadvantages, and reducing the disadvantages will take more time. People are more than happy to use APS-C cameras for professional work though, and some are very adamant about its advantages in the Canon and Nikon communities. But there will always be those Full Frame users who try to convince them that their only advantages is price, and that APS-C will always be inferior to Full Frame because of physical size. I think we both know there's more involved than that. ;)
     
    • Like Like x 2
  13. photoSmart42

    photoSmart42 Mu-43 Top Veteran

    628
    Feb 12, 2010
    San Diego, CA
    Don't confuse sensor performance with processing performance. They're quite different. In terms of actual sensor performance, not a whole lot has changed since the original G1 and GH1 sensors, respectively. Most of the 'advances' have been made in processing, and not all of them positive IMO. The GH2 sensor is higher resolution, but by most accounts performance at the sensor level is no different than the GH1 sensor.
     
  14. Ned

    Ned Mu-43 Legend

    Jul 18, 2010
    Alberta, Canada
    No, this is true. The physical sensors in the Olympus cameras have actually remained essentially unchanged since the G1! That's more or less the main idea behind what I'm saying though, is that there's so much more surrounding the sensor to produce an image than just the physical sensor chip, even if you don't account for lenses and optics. The increase in performance in so many ways of the E-5 and E-PL2 by weakening the AA filter is a good example.

    Increasing the pixel resolution on the same sensor I don't think was such a great move either. The GH1 sensor is different than the G1 or the Olympus 12.3MP LiveMOS sensors, and it was a good performer before the 16MP resolution change. But at least they stuffed those extra MP into a sensor that could handle it. However, when the GH1 came out only the Micro Four-Thirds community noticed the difference... the general populace thought it was the same as the G1 until the GH2 upped the resolution. To most people it's all in the numbers, but the real make-up of sensor performance is normally subtle and behind-the-scenes, often in areas that we don't measure with numbers. How many of us still love the images that were produced by the old Kodak CCD sensors of early Olympus DSLRs? The numbers show them to be worse sensors in every way, but our eyes don't agree.

    But hey, if taking the same GH1 sensor and making it 16MP will now show the consumer market that it's a better sensor and encourage them to buy it, then why not? So what if it has the same performance, lol? Some changes are all political, but are deemed necessary for that fact.
     
    • Like Like x 2
  15. Pan Korop

    Pan Korop Mu-43 Veteran

    479
    Mar 31, 2011
    Phare Ouest
    Is it Karl Marx who said "History doesn't repeat itself; it just stutters".

    So this roostery cock-fight, or vice-versa, reminds me of those who once blurted that 35 mm was not professional, because so much smaller than 2 1/4". And I read that Gene Smith got once fired from Time-Life for sticking to this stupid small square format instrad of good merkin 4x5 in... May well be why he ended fronting for the Pen ads in the same magazine--atta boy!

    But I haven't yet met one expert who'd examine one of my 11x14 digital prints and determine what it comes from: 8x10", Rolleiflex, Pen-F, Leica, D3x, Oly E-520 or Fuji S3, given a dozen such choices...

    Many people talk about wine. Very little can face a blind test...
     
    • Like Like x 2
  16. Grant

    Grant Mu-43 Veteran

    Thanks Mauve. I suspected as much.

    I wasn't concerned with you answer at all as I though "So pat your friend in the back, tell him kindly he has already chosen the best setup available and be done with it." Was a great response, all be it a bit tongue in cheek.

    I have always been told never argue with a fool as they will drag you down to their level and beat you with experience.
     
    • Like Like x 2
  17. Ned

    Ned Mu-43 Legend

    Jul 18, 2010
    Alberta, Canada
    lol. That's a good one. :rofl:
     
  18. RobWatson

    RobWatson Mu-43 Hall of Famer

    I find that just giving a great deluge of technical facts often shuts down the conversation which is a good indication that the person asking really doesn't know enough to actually know the difference but at least one has the satisfaction of actually fully answering the question.

    Of course, it is more fun to say 'yeah, it's a crap camera, but my photos are way better. Go figure!'
     
    • Like Like x 1
  19. phigmov

    phigmov Mu-43 Hall of Famer

    Apr 4, 2010
    :rofl:

    Yeah I hear this was taken with a 50 mega-pixel Hasselblad H4D-200MS in RAW and post-processed for that grainy effect -

    [​IMG]
     
    • Like Like x 1
  20. Ned

    Ned Mu-43 Legend

    Jul 18, 2010
    Alberta, Canada
    Some of the worst photos I've had to edit were taken with 50MB Hasselblads, lol. You should see the severe Chromatic Aberrations which looked like an American flag along the edges of one product photo I had to edit from an in-house studio taken with their Hasselblad and Phase One digital back. Large size, high resolution, and high ticket price aren't what makes a good photo...
     
    • Like Like x 1