Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
04-26-2020, 03:30 AM   #1
Pentaxian
Scout's Avatar

Join Date: Dec 2018
Location: western France
Photos: Gallery
Posts: 568
Technical question about RAWS from different camera bodies

Hello,
sorry for these strange questions, but it remains in my mind :

If we take different camera bodies from different brands, but we imagine they have the same sensor (Pentax K5, Nikon D7000, Sony Nex ... for example) and same caracteristics, same settings, same lenses,

To what extent the RAW files and results are different and why ?
On the other hand
to what extent the raw files and results can be similar ?
Thank you for sharing your knowledge !



Last edited by Scout; 04-26-2020 at 03:35 AM.
04-26-2020, 03:50 AM - 1 Like   #2
Site Supporter
Site Supporter




Join Date: May 2019
Photos: Albums
Posts: 3,116
The raw files will be - to some extent - different, as the sensor is just one part of the light-gathering media: microlens arrangements and color filter arrays are, AFAIK, determined by each manufacturer. The image processors will also affect the file (ahem baked-in lens corrections ahem Star-Eater). If there is on-sensor PDAF that, too, will have an impact (as some pixels are half-covered). The software profiles also have completely different behaviour depending on the source, at least in Lightroom.

As to the extent of the difference, I think the only way to quantify it would be to use the exact same lens (not even the same model; one copy of a given lens adapted for both systems) on two bodies, and then process the raw file in exactly the same way... I'm not even sure it is possible.
04-26-2020, 04:58 AM - 1 Like   #3
Pentaxian
Lowell Goudge's Avatar

Join Date: Jan 2007
Location: Toronto
Photos: Gallery | Albums
Posts: 16,234
QuoteOriginally posted by Serkevan Quote
Snip....

As to the extent of the difference, I think the only way to quantify it would be to use the exact same lens (not even the same model; one copy of a given lens adapted for both systems) on two bodies, and then process the raw file in exactly the same way... I'm not even sure it is possible.
It is entirely possible with a Fully manual macro bellows lens, on a close focus shot where The registration distance Does not come into play.

If someone makes all the adaptors for the back end, I will gladly lend my bellows and 135mm enlarging lens for the project
04-26-2020, 05:16 AM - 1 Like   #4
Site Supporter
Site Supporter




Join Date: May 2019
Photos: Albums
Posts: 3,116
QuoteOriginally posted by Lowell Goudge Quote
It is entirely possible with a Fully manual macro bellows lens, on a close focus shot where The registration distance Does not come into play.

If someone makes all the adaptors for the back end, I will gladly lend my bellows and 135mm enlarging lens for the project
Oh no, I mean if it's entirely possible to get the image processing equal. I should have specified . As for the lens, I would just suggest slapping a Pentax 67 lens on the desired cameras (and, for bonus points, watch how the tiny MILCs become ergonomic hell ).

04-26-2020, 05:26 AM - 2 Likes   #5
Loyal Site Supporter
Loyal Site Supporter
BigMackCam's Avatar

Join Date: Mar 2010
Location: North-East of England
Posts: 15,462
QuoteOriginally posted by Serkevan Quote
The raw files will be - to some extent - different, as the sensor is just one part of the light-gathering media: microlens arrangements and color filter arrays are, AFAIK, determined by each manufacturer. The image processors will also affect the file (ahem baked-in lens corrections ahem Star-Eater). If there is on-sensor PDAF that, too, will have an impact (as some pixels are half-covered). The software profiles also have completely different behaviour depending on the source, at least in Lightroom.

As to the extent of the difference, I think the only way to quantify it would be to use the exact same lens (not even the same model; one copy of a given lens adapted for both systems) on two bodies, and then process the raw file in exactly the same way... I'm not even sure it is possible.
For different reasons (related to camera profiling against colour charts), two of our members - @Digitalis and @Dartmoor Dave; - have used pinholes. Since they're readily available for all mounts - or easily made using a body cap - this would be a useful common "lens" with which to compare sensors between different cameras, at least so far as colour and noise are concerned (it wouldn't be much help regarding detail, due to diffraction)...

Last edited by BigMackCam; 04-26-2020 at 05:36 AM.
04-26-2020, 05:35 AM - 3 Likes   #6
Pentaxian
Lowell Goudge's Avatar

Join Date: Jan 2007
Location: Toronto
Photos: Gallery | Albums
Posts: 16,234
QuoteOriginally posted by Serkevan Quote
Oh no, I mean if it's entirely possible to get the image processing equal. I should have specified . As for the lens, I would just suggest slapping a Pentax 67 lens on the desired cameras (and, for bonus points, watch how the tiny MILCs become ergonomic hell ).
Getting the image processing equal is the part of this that each camera maker adds to the equation, otherwise, we would have one camera maker per chip manufacturer, and a standard across platform lens mount, and all other companies would be reduced to making optics.

Itíll never happen.
04-26-2020, 05:42 AM - 1 Like   #7
Site Supporter
Site Supporter




Join Date: May 2019
Photos: Albums
Posts: 3,116
Honestly there are so many variables in between subject and print (or screen) that the differences between the raw file differences are never going to really matter in practical shooting. It's an interesting thought exercise but I certainly wouldn't spend a lot of my own time on it
04-26-2020, 06:20 AM - 1 Like   #8
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 36,218
QuoteOriginally posted by Serkevan Quote
Honestly there are so many variables in between subject and print (or screen) that the differences between the raw file differences are never going to really matter in practical shooting. It's an interesting thought exercise but I certainly wouldn't spend a lot of my own time on it
When I buy a new camera, I create a few processing one click adjustments for use right after import. These profiles do up to 7 or 8 steps of adjustments which I find are always needed on that sensor.

My K-3 needs twice as much sharpening as my K-1 and less contrast. My Lumix ZS100 needs no sharpening, and if I use a profile with sharpening there will be sharpening artifacts.

Using my K-3 profile on my K-1, I cut the sharpening in half.

Anyway, you should do this with every camera. Take an average kind of shot. Work it to perfection (with no brushed corrections, global corrections only) get it perfect, save it as a preset. You will still have to make adjustments but you will be doing minor adjustments to the existing preset.

04-26-2020, 06:58 AM - 3 Likes   #9
Loyal Site Supporter
Loyal Site Supporter




Join Date: Feb 2018
Posts: 309
Other folks have raised great points, here's a few other thoughts for you.

QuoteOriginally posted by Scout Quote

If we take different camera bodies from different brands, but we imagine they have the same sensor (Pentax K5, Nikon D7000, Sony Nex ... for example) and same caracteristics, same settings, same lenses,

To what extent the RAW files and results are different and why ?

Besides the obvious file format differences and manufacturer-specific content, to make a comparison, the CMOS Active Pixel Sensors have to be the same specification to be comparable.
There's a handy Wikipedia listing of most of the Sony Semiconductor EXMOR/Pregius/Starvis families here: Exmor - Wikipedia
For example, the Pentax K1, Nikon D810, and Sony A7R all have the Sony Semiconductor IMX094 CMOS APS sensor with virtually the same characteristics.
Individual sensors vary; some have hot pixels/warm pixels/weak pixels/dead pixels. Each has a certain amount of inherent noise and amplifier glow.
The sensor is only one piece of the puzzle.

The power supplies in each camera are different as are the firmware that controls the operation of the sensor. Read-out speed also sometimes influences noise levels - slower readout usually means cleaner images.
The sensors also generate heat; that heat has to be dissipated, or it creates more noise in the image. Each manufacturer has their own way of getting the heat out of the chip.
The firmware in the different cameras also does some amount of calibration and image processing and control of the bias levels, gain, and black-level offset compensation.

The gain, offset, and bias levels are calibrated to the ISO number differently by each manufacturer.

In many cases, the individual camera is calibrated to deal with the variations as no two CMOS APS sensors are 100% identical. Manufacturing variations over the lifetime of the sensors also have a big impact. A sensor made in a batch from 2016 may behave differently than one made in 2019. Plus electronic components sometimes become obsolete or are substituted, and this causes variations.

Optical filters used to block IR and UV light are also different between manufacturers, and this is why some cameras (Pentax) tend to show deeper reds/magentas than others, as they let more deep red light through.
The quantum efficiency varies from sensor to sensor, (eg how much e- signal you get for each photon that heads in to pixel). This is partly related to variations in the sensor semiconductor, the color dyes that mask each pixel (RGGB), and the microlenses over each pixel. Add the UV/IR blocking optical filter, (no two are perfectly identical), and thus the colour balance is different from camera to camera.

So, despite having the same sensor, there's a lot of differences from model-to-model.

Someone else mentioned that the RAW images aren't necessarily truly the digital values of each pixel as read by the electronics. They sometimes are processed to remove dead/hot pixels etc. eg Image Processor/Accelerator.

You might find it interesting to compare raw images from two cameras of the same model, shot with the same lens, subject, lighting, and tripod mounted.



QuoteOriginally posted by Scout Quote
On the other hand
to what extent the raw files and results can be similar ?
Thank you for sharing your knowledge !
QuoteOriginally posted by Scout Quote
The dimensions of the active pixels will be identical. eg 7390 x 4912 pixels, 4.88um square.
Quantum efficiencies will be very similar, so RGGB colour balance will be similar, but not identical.
Results depend on how you process the RAWs, the accuracy of the calibrations, and what is in common between them (calibration targets, calibrated light source, use of a common lens, etc.)
No two images will be 100% identical but they can come close.

Great questions!
04-26-2020, 12:17 PM - 1 Like   #10
Otis Memorial Pentaxian
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 37,855
QuoteOriginally posted by Scout Quote
to what extent the raw files and results can be similar ?
They all contain capture metadata (information about camera, settings, and such) and also minimally processed sensor capture data. Often, they also include preview JPEG images. They never contain data organized as pixels (image data).

QuoteOriginally posted by Scout Quote
To what extent the RAW files and results are different and why ?
With one exception, all RAW file types are proprietary. That answers the "why" question. How the data and metadata are stored within the file is specific to the type and not published by intent. The one exception is DNG where the specification is published by Adobe. Even that does not mean that all DNG are data equivalent and broadly compatible across all tools. For example, a DNG created on export from Lightroom will not be readable by Pentax Digital Camera Utility or by a Pentax camera.*

Why results vary even with similar sensors depends on the low level image processing (Pentax Prime vs. Nikon EXPEED, for example) as well as the quirks of the RAW processor software used to create the image itself.


Steve

* At least that was the case the last time I looked.

Last edited by stevebrot; 04-26-2020 at 12:23 PM.
04-26-2020, 06:26 PM - 1 Like   #11
Junior Member
richard_b's Avatar

Join Date: Oct 2012
Location: Massachusetts
Posts: 35
QuoteOriginally posted by Scout Quote
Hello,
sorry for these strange questions, but it remains in my mind :

If we take different camera bodies from different brands, but we imagine they have the same sensor (Pentax K5, Nikon D7000, Sony Nex ... for example) and same caracteristics, same settings, same lenses,

To what extent the RAW files and results are different and why ?
On the other hand
to what extent the raw files and results can be similar ?
Thank you for sharing your knowledge !
I used the same sensor in three cameras - the Sony SLT A55, the Nikon D7000 and the Pentax K-5iis. I never did strict comparison testing, but based on experience, the lenses used make a much bigger difference than the body. I used the CZ16-80 on the Sony, this was a Zeiss design with Zeiss T* coatings, but manufactured by Sony. it was an excellent all round performer, but crucially it delivered great contrast. On the Nikon I used the Nikkor 16-85. This had very good resolution and control of aberrations etc, but it just couldn't deliver the contrast that the CZ16-80 could provide.

It was by chance I discovered the Pentax Ltd's and switched to Pentax to regain the contrast/rendition I had lost on the Nikkor 16-85. With the D7000, it was not the case of some other camera body difference causing the lack of contest, as I also had a Nikkor 70-300 (don't remember the exact designation) and this provided much better contrast than the Nikkor 16-85.
04-27-2020, 01:56 AM   #12
Pentaxian
Scout's Avatar

Join Date: Dec 2018
Location: western France
Photos: Gallery
Posts: 568
Original Poster
QuoteOriginally posted by richard_b Quote
I used the same sensor in three cameras - the Sony SLT A55, the Nikon D7000 and the Pentax K-5iis. I never did strict comparison testing, but based on experience, the lenses used make a much bigger difference than the body. I used the CZ16-80 on the Sony, this was a Zeiss design with Zeiss T* coatings, but manufactured by Sony. it was an excellent all round performer, but crucially it delivered great contrast. On the Nikon I used the Nikkor 16-85. This had very good resolution and control of aberrations etc, but it just couldn't deliver the contrast that the CZ16-80 could provide.

It was by chance I discovered the Pentax Ltd's and switched to Pentax to regain the contrast/rendition I had lost on the Nikkor 16-85. With the D7000, it was not the case of some other camera body difference causing the lack of contest, as I also had a Nikkor 70-300 (don't remember the exact designation) and this provided much better contrast than the Nikkor 16-85.
Hello and thank you all for your answers, I am wandering in your case if the lack of contrast is not due to the AA filter on the D7000, I think there is not on the K5IIs.
And I am sure it's not possible to compare prime limiteds with standard zooms

---------- Post added 04-27-20 at 09:12 AM ----------

QuoteOriginally posted by ProfessorBuzz Quote
Other folks have raised great points, here's a few other thoughts for you.


Besides the obvious file format differences and manufacturer-specific content, to make a comparison, the CMOS Active Pixel Sensors have to be the same specification to be comparable.
There's a handy Wikipedia listing of most of the Sony Semiconductor EXMOR/Pregius/Starvis families here: Exmor - Wikipedia
For example, the Pentax K1, Nikon D810, and Sony A7R all have the Sony Semiconductor IMX094 CMOS APS sensor with virtually the same characteristics.[/LEFT]
Individual sensors vary; some have hot pixels/warm pixels/weak pixels/dead pixels. Each has a certain amount of inherent noise and amplifier glow.[LEFT]
The sensor is only one piece of the puzzle.

The power supplies in each camera are different as are the firmware that controls the operation of the sensor. Read-out speed also sometimes influences noise levels - slower readout usually means cleaner images.
The sensors also generate heat; that heat has to be dissipated, or it creates more noise in the image. Each manufacturer has their own way of getting the heat out of the chip.
The firmware in the different cameras also does some amount of calibration and image processing and control of the bias levels, gain, and black-level offset compensation.

The gain, offset, and bias levels are calibrated to the ISO number differently by each manufacturer.

In many cases, the individual camera is calibrated to deal with the variations as no two CMOS APS sensors are 100% identical. Manufacturing variations over the lifetime of the sensors also have a big impact. A sensor made in a batch from 2016 may behave differently than one made in 2019. Plus electronic components sometimes become obsolete or are substituted, and this causes variations.

Optical filters used to block IR and UV light are also different between manufacturers, and this is why some cameras (Pentax) tend to show deeper reds/magentas than others, as they let more deep red light through.
The quantum efficiency varies from sensor to sensor, (eg how much e- signal you get for each photon that heads in to pixel). This is partly related to variations in the sensor semiconductor, the color dyes that mask each pixel (RGGB), and the microlenses over each pixel. Add the UV/IR blocking optical filter, (no two are perfectly identical), and thus the colour balance is different from camera to camera.

So, despite having the same sensor, there's a lot of differences from model-to-model.

Someone else mentioned that the RAW images aren't necessarily truly the digital values of each pixel as read by the electronics. They sometimes are processed to remove dead/hot pixels etc. eg Image Processor/Accelerator.

You might find it interesting to compare raw images from two cameras of the same model, shot with the same lens, subject, lighting, and tripod mounted.




The dimensions of the active pixels will be identical. eg 7390 x 4912 pixels, 4.88um square.
Quantum efficiencies will be very similar, so RGGB colour balance will be similar, but not identical.
Results depend on how you process the RAWs, the accuracy of the calibrations, and what is in common between them (calibration targets, calibrated light source, use of a common lens, etc.)
No two images will be 100% identical but they can come close.

Great questions!
Thank you for this exhaustive answer, I understand there are a lot of parameters to take into account,

That data and metadata are differently treated by the software, that heat, light and camera bodies live and evolve differently.
after all I would be very interested in a comparison study, like BigMackCam said, maybe results would be very very close !
04-27-2020, 03:20 AM   #13
Junior Member
richard_b's Avatar

Join Date: Oct 2012
Location: Massachusetts
Posts: 35
QuoteOriginally posted by Scout Quote
Hello and thank you all for your answers, I am wandering in your case if the lack of contrast is not due to the AA filter on the D7000, I think there is not on the K5IIs.
And I am sure it's not possible to compare prime limiteds with standard zooms

---------- Post added 04-27-20 at 09:12 AM ----------

I think the lack of AA filter only makes a slight difference in resolution, I don't think it makes much, if any, difference to the contrast. This site did comparison testing between the K-5ii and K-5iis with the same lens. There was a slight difference in resolution, but none in contrast that I could see. The primary difference between the K-5ii and K-5iis was that one had an AA filter and the other did not. As others have pointed out there are many factors affecting final image quality, but it certainly possible to compare the final output from zooms where they overlap in range or by setting a zoom to the same focal length as a prime you want to compare.
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!
bodies, camera, files, heat, light, nikkor, pentax, pentax help, photography, pixel, question about raws, sensor, sensors, troubleshooting, variations
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
New 645z owner. What computing power am I going to need to handle these raws? TPR Pentax Medium Format 40 03-07-2020 09:57 AM
k30 raws vs k5iis raws everydaylife Pentax DSLR Discussion 17 01-14-2016 10:48 PM
Does the in camera noise reduction apply to Raws or just Jpegs? coreyhkh Pentax K-5 8 08-04-2012 06:53 AM
Very technical question about in-camera stabilization olenl Pentax DSLR Discussion 29 11-22-2010 02:29 PM
bug in Camera Utility 4.10 with K-x RAWs? noou Digital Processing, Software, and Printing 6 12-28-2009 07:50 AM



All times are GMT -7. The time now is 01:07 PM. | 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