Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
03-19-2016, 10:33 PM   #271
Pentaxian




Join Date: May 2009
Location: Ponoka Alberta Canada
Photos: Gallery
Posts: 471
QuoteOriginally posted by dtmateojr Quote
ROFL! Image size is horizontal pixels multiplied by vertical pixels. Nothing more. Don't overanalyse. You really are too funny.
It is important that you understand the size of the image behind the lens, as this is what determines how much signal the sensor is allowed to convert at a given SS and Fstop. So just saying 16mp is 16mp without knowing what size those 16mp represents is a bit foolish as it pertains to how large the signal will be.
As for you argument that the K5 and the D800 capture the same signal ( total light) at the same shutter speed and Fstop let’s look at it from an engineer’s perspective.


We have 2 signals the strength of those2 signals are unknown but we do know that at the same photo site they have the same s/n ratio. We also know that one of the signals is being sampled at a resolution 2.25 times greater 36mp verses 16mp. We also know that both photo sites are exposed with the same Fstop , shutter speed and that both photo sites are the same size .We also know that as we increase the resolution at which we sample a signal the s/n ratio decreases ( very basic principle in analyzing a signal, its bandwidth dependant ). So if both the 36mp and the 16mp photo sites measure the same s/n ratio, how can both signals have equal strength ( a reminder that the signal is everything captured with the limits of the sensor)?

If you can answer this it would give weight to your very argument that the K5 and d800 capture the same signal?

See how I know nothing with regard to sampling I am looking forward to your answer.


Last edited by Ian Stuart Forsyth; 03-19-2016 at 11:01 PM.
03-19-2016, 11:48 PM   #272
Forum Member




Join Date: Jul 2012
Posts: 57
Sorry, my earlier post was inaccurate in language when I said that the DxO normalization process did not 'directly' involve the sensor size (but I went on to show examples to explain how despite that, the Print scores always show that larger sensors have better performance for a common sensor tech/age). The examples are fine, but what I should have said initially was that the DxO normalization process does not APPEAR to directly involve sensor size. But it actually does, because it involves the number of megapixels, which itself is simply the ratio between sensor area and pixel area. With N = number of pixels, A_s = sensor area, and A_p = individual pixel area, N=A_s/A_p. I apologize for glossing over this fact, which may have certainly helped some people to accept how it is that Print SNR truly does reflect gains in performance due to a larger sensor size.

So to summarize, while SNR_screen only takes into account per-pixel noise (which is only related to A_p, not to A_s or sensor size at all), SNR_print certainly does take both A_p AND crucially A_s into account. This is why the normalized Print measures are much more useful, especially when comparing sensors of different sizes.

SNR_print (aka SNR_norm) = SNR_screen + 20 log sqrt(N/N_ref) = SNR_screen + 20 log sqrt( (A_s / A_p) / N_ref)

If you increase sensor size, ie. A_s, then SNR_print increases, EVEN IF keeping the ratio between A_s/A_p constant (ie. maintaining the number of pixels). This is because if you increase pixel area A_p in proportion to A_s in order to maintain # of pixels, then while the second term in the above equation (the + 20 log sqrt... part) doesn't change, the SNR_screen DOES INCREASE, due to the larger pixel area.
03-20-2016, 02:32 AM   #273
Pentaxian




Join Date: Jan 2012
Photos: Gallery
Posts: 587
QuoteOriginally posted by Ian Stuart Forsyth Quote
It is important that you understand the size of the image behind the lens, as this is what determines how much signal the sensor is allowed to convert at a given SS and Fstop. So just saying 16mp is 16mp without knowing what size those 16mp represents is a bit foolish as it pertains to how large the signal will be.
What?!!! Your calculations for IMAGE size was not image size. You were calculating for SENSOR size (sensel pitch x num of pixels). Duh?! Sensor size is NOT equal to image size. Fact.

QuoteQuote:
As for you argument that the K5 and the D800 capture the same signal ( total light) at the same shutter speed and Fstop let’s look at it from an engineer’s perspective.
I NEVER said that. I never mentioned total light. You equivalence fu proponents should stop using that total light crap.

I said K5 and D800 have same SNR and dxomark SCREEN snr is PROOF of that.


QuoteQuote:
We have 2 signals the strength of those2 signals are unknown but we do know that at the same photo site they have the same s/n ratio. We also know that one of the signals is being sampled at a resolution 2.25 times greater 36mp verses 16mp. We also know that both photo sites are exposed with the same Fstop , shutter speed and that both photo sites are the same size .We also know that as we increase the resolution at which we sample a signal the s/n ratio decreases ( very basic principle in analyzing a signal, its bandwidth dependant ). So if both the 36mp and the 16mp photo sites measure the same s/n ratio, how can both signals have equal strength ( a reminder that the signal is everything captured with the limits of the sensor)?

If you can answer this it would give weight to your very argument that the K5 and d800 capture the same signal?

See how I know nothing with regard to sampling I am looking forward to your answer.

FFS, I never said they capture the same total light. Learn to read.
03-20-2016, 02:41 AM - 1 Like   #274
Loyal Site Supporter




Join Date: Mar 2009
Location: Southern Indiana
Photos: Gallery
Posts: 15,310
I think we are heading for this thread being shut down for good soon. Oh, well...

I guess I would say that I don't really understand the angst over this whole thing. Clearly at many print sizes (not all) a 36 megapixel sensor is going to produce significantly better results than a 16 megapixel APS-C sensor. And yes, if you have a larger sensor, your high iso performance does improve -- as long as you maintain a given print size.

I am planning to get a full frame camera. There will be some improvement in performance. My lenses will see a wider angle of view. I'll have to stop down a little to get more depth of field. But the end result (hopefully) will be nice images that I can share with others. Images of kids and family and landscapes. And honestly, all of these cameras are capable of taking nice images of those things in most settings. Even entry level cameras have nice image quality and amazing specs.

I am glad that Pentax is releasing a full frame camera and while I'll get one, I imagine my images will look pretty similar between the two, since the big factor is sitting about six inches behind the viewfinder.

03-20-2016, 02:45 AM   #275
Pentaxian




Join Date: Jan 2012
Photos: Gallery
Posts: 587
Jack Hogan of DPR has seen the light

03-20-2016, 03:31 AM   #276
Pentaxian




Join Date: Sep 2011
Location: Nelson B.C.
Posts: 3,226
QuoteOriginally posted by crussellsprout Quote
Sorry, my earlier post was inaccurate in language when I said that the DxO normalization process did not 'directly' involve the sensor size (but I went on to show examples to explain how despite that, the Print scores always show that larger sensors have better performance for a common sensor tech/age). The examples are fine, but what I should have said initially was that the DxO normalization process does not APPEAR to directly involve sensor size. But it actually does, because it involves the number of megapixels, which itself is simply the ratio between sensor area and pixel area. With N = number of pixels, A_s = sensor area, and A_p = individual pixel area, N=A_s/A_p. I apologize for glossing over this fact, which may have certainly helped some people to accept how it is that Print SNR truly does reflect gains in performance due to a larger sensor size.

So to summarize, while SNR_screen only takes into account per-pixel noise (which is only related to A_p, not to A_s or sensor size at all), SNR_print certainly does take both A_p AND crucially A_s into account. This is why the normalized Print measures are much more useful, especially when comparing sensors of different sizes.

SNR_print (aka SNR_norm) = SNR_screen + 20 log sqrt(N/N_ref) = SNR_screen + 20 log sqrt( (A_s / A_p) / N_ref)

If you increase sensor size, ie. A_s, then SNR_print increases, EVEN IF keeping the ratio between A_s/A_p constant (ie. maintaining the number of pixels). This is because if you increase pixel area A_p in proportion to A_s in order to maintain # of pixels, then while the second term in the above equation (the + 20 log sqrt... part) doesn't change, the SNR_screen DOES INCREASE, due to the larger pixel area.
That works comparing two sensors of the same pixel density. If you change the density of one the function favors the higher density.

I'm not entirely convinced that it is representative.
03-20-2016, 05:24 AM - 1 Like   #277
Veteran Member
jsherman999's Avatar

Join Date: Dec 2007
Photos: Gallery
Posts: 8,228
QuoteOriginally posted by crussellsprout Quote
Sorry, my earlier post was inaccurate in language when I said that the DxO normalization process did not 'directly' involve the sensor size (but I went on to show examples to explain how despite that, the Print scores always show that larger sensors have better performance for a common sensor tech/age). The examples are fine, but what I should have said initially was that the DxO normalization process does not APPEAR to directly involve sensor size. But it actually does, because it involves the number of megapixels, which itself is simply the ratio between sensor area and pixel area. With N = number of pixels, A_s = sensor area, and A_p = individual pixel area, N=A_s/A_p. I apologize for glossing over this fact, which may have certainly helped some people to accept how it is that Print SNR truly does reflect gains in performance due to a larger sensor size.

So to summarize, while SNR_screen only takes into account per-pixel noise (which is only related to A_p, not to A_s or sensor size at all), SNR_print certainly does take both A_p AND crucially A_s into account. This is why the normalized Print measures are much more useful, especially when comparing sensors of different sizes.

SNR_print (aka SNR_norm) = SNR_screen + 20 log sqrt(N/N_ref) = SNR_screen + 20 log sqrt( (A_s / A_p) / N_ref)

If you increase sensor size, ie. A_s, then SNR_print increases, EVEN IF keeping the ratio between A_s/A_p constant (ie. maintaining the number of pixels). This is because if you increase pixel area A_p in proportion to A_s in order to maintain # of pixels, then while the second term in the above equation (the + 20 log sqrt... part) doesn't change, the SNR_screen DOES INCREASE, due to the larger pixel area.
Not understanding this leads to interesting claims, like "simply dumping more pixels into a smaller sensor will beat or match a larger sensor in the final image."

Or, "screen tab proves the OMD E-M5 matches the FF Canon 5dS"


.
03-20-2016, 05:25 AM   #278
Pentaxian




Join Date: Dec 2010
Location: Ontario
Photos: Gallery
Posts: 2,375
QuoteOriginally posted by derekkite Quote
That works comparing two sensors of the same pixel density. If you change the density of one the function favors the higher density.
His last paragraph is important. Changing the pixel density should change the SNR_screen result and, at least in practical examples of sensors from roughly the same generation, this mostly results in a wash for the SNR_print when combined with the change in downsampling.

---------- Post added 03-20-16 at 08:33 AM ----------

QuoteOriginally posted by normhead Quote
It is kind of surprising though that a Nikon 300 ƒ4 weighs less than the DA* 200 ƒ2.8. The Canon 200 ƒ4 however weighs quite a bit more. So even that is somewhat dependant on the lens design.
That Nikon is a very different design using a fresnel element to keep it crazy small- A Closer Look at Nikon's New Phase Fresnel (PF) Lens Technology. There are more trade offs than just cost, but it looks like a pretty cool option when space is a premium (this has nothing to do with your post, I just think it's a neat lens design).

As you said they do have a more standard design similar in paper specs to the Canon - Nikon AF-S NIKKOR 300mm f/4D IF-ED Lens 1909 B&H Photo Video


Last edited by BrianR; 03-20-2016 at 05:34 AM.
03-20-2016, 05:48 AM   #279
Loyal Site Supporter




Join Date: Apr 2010
Location: Tennessee
Posts: 5,550
QuoteOriginally posted by MRCDH Quote
K-1 So What Is Full Frame Going To Provide Over A Crop Frame DSLR
More money for Ricoh.
03-20-2016, 06:09 AM   #280
Veteran Member
jsherman999's Avatar

Join Date: Dec 2007
Photos: Gallery
Posts: 8,228
QuoteOriginally posted by dtmateojr Quote
Jack Hogan of DPR has seen the light
Pasting things out of context to try to help yourself? I see why you stopped there :



Good try, though!

(Also, I think Jack Hogan would be surprised to find that he's an equivalence denialist, which I assume you're also implying. )

.
03-20-2016, 06:37 AM   #281
Pentaxian




Join Date: Jan 2012
Photos: Gallery
Posts: 587
K-1 So What Is Full Frame Going To Provide Over A Crop Frame DSLR

QuoteOriginally posted by jsherman999 Quote
Pasting things out of context to try to help yourself? I see why you stopped there :







Good try, though!



(Also, I think Jack Hogan would be surprised to find that he's an equivalence denialist, which I assume you're also implying. )



.

Keep believing that lol! Equivalence will not change facts.

moving_comfort wrote:

Correct? Print tab accounts for total light differences (not just MP differences) and gives an accurate comparison of output from different sensor sizes?
I don't know what total light or its units are but the print tab simply puts all sensors on a level playing field by assuming all images are resized to 8MP. So the D7k 'screen' values would be adjusted by sqrt(16/8) and the D800's by sqrt(36/8) to become 'print'.

Jack
03-20-2016, 06:46 AM - 1 Like   #282
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 24,317
QuoteQuote:
Clearly at many print sizes (not all) a 36 megapixel sensor is going to produce significantly better results than a 16 megapixel APS-C sensor. And yes, if you have a larger sensor, your high iso performance does improve -- as long as you maintain a given print size.
Cleary at many print sizes a 36 MP sensor makes no difference at all. So, without defining what that size is, you're not saying anything. The many print sizes you discuss could all be over 60 inches wide... and completely irrelevant to the average photographer. Personally, I'd say large print sizes are relevant only to the accomplished photographer with sterling reputation and a couple of galleries under his thumb. The people at the craft shows I've been to don't buy huge three to six thousand dollar images, and don't have the wall space to display them if they did. If you can do an opening and have three or 4 hundred millionaires drop by, then you need to worry about supersize printing. But people were selling those prints long before 36 MP camera came along.

Based on an image I once ended up printing at 50 dpi before upscaling, people are really underselling the current 20-28 MP cameras, compared to just a few years ago they were top of the line. Now they aren't good enough. Not because of a lack of quality, but just because there is more.

We are truly in the era of 350 HP cars. People are buying what on some metrics is the biggest, based on the fact it's the biggest. Whether or not it's the best at anything is pretty much irrelevant, to this type of purchaser.
03-20-2016, 07:04 AM - 1 Like   #283
ACG
Forum Member




Join Date: Apr 2014
Photos: Gallery | Albums
Posts: 55
QuoteOriginally posted by Rondec Quote
I think we are heading for this thread being shut down for good soon. Oh, well...

I guess I would say that I don't really understand the angst over this whole thing. Clearly at many print sizes (not all) a 36 megapixel sensor is going to produce significantly better results than a 16 megapixel APS-C sensor. And yes, if you have a larger sensor, your high iso performance does improve -- as long as you maintain a given print size.

I am planning to get a full frame camera. There will be some improvement in performance. My lenses will see a wider angle of view. I'll have to stop down a little to get more depth of field. But the end result (hopefully) will be nice images that I can share with others. Images of kids and family and landscapes. And honestly, all of these cameras are capable of taking nice images of those things in most settings. Even entry level cameras have nice image quality and amazing specs.

I am glad that Pentax is releasing a full frame camera and while I'll get one, I imagine my images will look pretty similar between the two, since the big factor is sitting about six inches behind the viewfinder.
I agree. It is all about enjoyment. If you think that you are going to enjoy jumping to the FF format, if you have the money to do so, if you wish to resurrect your FF lenses from the film era, if you have the patience for the added attention to technical detail required when using the FF system and if you want to own the latest technology for a while then go for it and enjoy the whole experience.
Will you notice a significant difference in the end product on a computer screen? No. Will you notice a significant difference at A2, probably not (I have done comparisons between sony a7r (36mp) with pentax k28 and sony Rx100 mk3 (20 MP 1 inch sensor) at A2 and can"t spot any significant difference).
So for those pentaxians in the northern hemisphere, spring has sprung and summer will be upon us so get out there and get shooting!!
03-20-2016, 07:17 AM   #284
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 24,317
QuoteQuote:
Will you notice a significant difference at A2, probably not (I have done comparisons between sony a7r (36mp) with pentax k28 and sony Rx100 mk3 (20 MP 1 inch sensor) at A2 and can"t spot any significant difference).


There are two kinds of people in these discussions, people who have done actual comparisons, and say there is next to no difference, and people who haven't done any side by side printing, and think there's a big difference.

The simple fact is, if you have a great image, people don't spend a lot of time worrying about resolution. If you have a bad image, high resolution may make it look worse.
03-20-2016, 07:24 AM   #285
Veteran Member
jsherman999's Avatar

Join Date: Dec 2007
Photos: Gallery
Posts: 8,228
QuoteOriginally posted by dtmateojr Quote
Keep believing that lol! Equivalence will not change facts.
Selectively quoting again? Why don't you quote some things from this thread, in which your... EDIT "interesting"... theories were not-so-gently shown the door (big D is the 'unknown member' in that thread, becuase he was later banned as a troll.) Link to major fail

Here's how this thread/argument has progressed, btw:

Gallon and quart full of water sit next to each other.

Guy1: As you can see, an Oz of water taken from each of these containers has the same mass and density, so the containers are equally capable of carrying water.

Guy2: What do you mean equally capable? The Gallon container holds more total water.

Guy1: What is this concept of "Total Water?!" Can't you see the individual ounces are EXACTLY THE SAME?

Guy2: Of course, but if you weigh both containers, the Gallon weighs more. It's got more volume, same quality of water inside, means more total water.

Guy1: It weighs more simply because it has MORE OUNCES!

Guy2: Huh?

Guy1: For example if you simply put MORE OUNCES IN the QUART CONTAINER, it would match the gallon in weight! LULZ ROFL! It's about the # of ounces, and the quality of each ounce! You Total water guys kill me!

Guy2: How would that work, exactly, putting more ounces into a quart?

Guy1: Well if such a beast existed maybe denser water maybe, that's all they would have to do and then they would weigh the same!

Guy2: OK... Setting aside the fact that an Oz is volume... couldn't they just put this 'denser water' into the gallon then too?

Guy1: ROFL! You guys!

..etc...

An actual example of that last part:

QuoteOriginally posted by dtmateojr:
An 50Mp E-M5, if such a beast exists, will match the performance of a full frame 5DS at any print size. Sensor size does not matter.

Therefore, equivalence is wrong.
... ^ this imagines a magical or at least yet-to-be-invented pixel with the current efficiency and performance of the E-M5's current pixel (or maybe even better efficiency, who knows what's being imagined,) but made smaller to fit 50MP of them into that same micro-four-thirds sensor area.

Why not take these magical yet-to-be-invented pixels and put them into a larger sensor, thus maintaining the same advantage that exists today?

.

Last edited by jsherman999; 03-20-2016 at 08:08 AM.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
24x36mm, 300mm, advantages, app, aps-c, body, camera, cost, crop, dslr, f/2.8, ff, field, focus, frame, full-frame, k-1, k-3, k3, lens, ovf, pentax, pentax k-3, quality, sensor, sigma, view
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Crop Sensors vs Full Frame :: Crop Or Crap? i83N Photographic Industry and Professionals 44 07-30-2014 06:00 AM
No Pentax Full Frame DSLR, What to Do Now Then? RiceHigh Pentax DSLR Discussion 95 07-19-2014 10:57 AM
The Full Frame NEX is going to be announced in October ! jogiba Non-Pentax Cameras: Canon, Nikon, etc. 30 09-15-2013 03:23 PM
So.... Pentax Full Frame is a sure thing? LFLee Photographic Industry and Professionals 25 11-02-2012 12:55 PM
So It's True....a Full-Frame 'Professional' Samsung Dslr Coming !!!! Confused Pentax DSLR Discussion 9 05-02-2008 04:12 PM



All times are GMT -7. The time now is 01:52 AM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top