/p/ - Photography » Thread #3569169

[26 / 4 / ?]

Anonymous Fri 15 Nov 2019 01:34:44 No.3569169 View View Reply Original Report

Quoted By: >>3569251 >>3569285 >>3569545 >>3569797

Could you even tell the difference between a digital mirrorless medium format vs. digital mirrorless full-frame camera?

It's definitely easy to spot film medium format vs. 35mm or digital, though I feel as if there is hardly any difference on a modern camera visually other than the obvious larger sensor advantages.

Anonymous

Anonymous Fri 15 Nov 2019 05:37:30 No.3569251 Report

Quoted By:

>>3569169
Not when you're just posting on Facebook or Instagram, but if you're cropping, pixel peeping, or making a large print to hang on the wall medium format does make a difference.

Anonymous

View Same Google ImgOps iqdb SauceNAO Digital-sensor-sizes.jpg, 122KiB, 700x326

Anonymous Fri 15 Nov 2019 08:36:31 No.3569285 Report

Quoted By: >>3569411 >>3569487

>>3569169
The only reason I could think of why the difference is less obvious in digital vs film is because digital medium format is smaller in size than film medium format.
The largest medium format sensor (the ones Phase One makes) are almost as big as 645, the smallest medium format frame size in film.
Compared to 6x7 they are still dwarfed.

Anonymous

Anonymous Fri 15 Nov 2019 20:00:02 No.3569411 Report

Quoted By:

>>3569285
Doesn't really matter when you don't have any MF lenses that are faster than 1.2 equivalent on 35mm. Yeah sure if you had like a 150mm f/1.4 for 6x7 it would have a completely different look compared to any existing 70-75mm FF lens, but it doesn't exist. Also you don't even shoot wide open that often because it gets gimmicky.
I think the "look" mostly comes down to composing differently with a wlf and a different aspect ratio in mind, the aspect ratio itself, and the look of film vs digital but with many of the limitations of film "removed". You're used to comparing 35mm film to FF digital and seeing grain and less detail on film, but 6x7 essentially removes those and leaves only the other characteristics of film.
I think all of this could be reproduced with using a wider lens and cropping to 6x7, and digital post processing if you spent enough time on it. Despite what boomers would have you believe, digital does have more dynamic range and higher effective resolution at the same size so you do have more data to work with and you could work backwards by removing very specific little bits of that data. The new sony A7r mark whatever has >60mpx of resolution and you can use pixel shift for even more. At the same time it would be like trying to create a completely photorealistic image in MS paint, technically possible but extremely impractical, just take a fucking photo at that point.
The depth of field and field of view can be perfectly reproduced by fast modern FF lenses, so that's definitely not it.

Anonymous

Anonymous Sat 16 Nov 2019 00:22:26 No.3569487 Report

Quoted By: >>3569535 >>3569634 >>3569795

>>3569285
With film, the larger the format, the better image you get. No questions about it. On digital sensor, performance stays the same, no matter how large it is. At same iso, you'll get same exact noise on same sensor tech, no matter how large the sensor is. Only real difference will be pixel count.

Anonymous

View Same Google ImgOps iqdb SauceNAO Screen-Shot-2019-11-15-at-5.50.0 (...).jpg, 395KiB, 1326x1314

Anonymous Sat 16 Nov 2019 01:50:55 No.3569535 Report

Quoted By: >>3569539

>>3569487
>At same iso, you'll get same exact noise on same sensor tech, no matter how large the sensor is.
LOL

Anonymous

Anonymous Sat 16 Nov 2019 01:59:14 No.3569539 Report

Quoted By: >>3569549

>>3569535
Those aren't the same tech.

Anonymous

Anonymous Sat 16 Nov 2019 02:05:55 No.3569545 Report

Quoted By:

>>3569169
Depends on what you are doing with the images really. If printing by format size digital has the advantage, as film has limitations on how much you can blow up an image for the format size. Digital does as well, but it is dependent upon the file size, as where film is on the format size.

Anonymous

Anonymous Sat 16 Nov 2019 02:15:26 No.3569549 Report

Quoted By: >>3569836

>>3569539
No they're not. In terms of sensor tech age from newest to oldest they are: 90D > GFX > E-M1X > 5DsR

Yet in terms of high ISO performance best to worst they are: GFX > 5DsR > 90D > E-M1X. Funny how that matches sensor size regardless of tech.

Anonymous

View Same Google ImgOps iqdb SauceNAO Screenshot_20191116-124920_Samsu (...).jpg, 543KiB, 1337x1386

Anonymous Sat 16 Nov 2019 04:52:46 No.3569634 Report

Quoted By:

>>3569487
>On digital sensor, performance stays the same, no matter how large it is
Can confirm, my 1/3" sony BSI smartphone sensor produces the exact same noise at ISO 6400 as the sony A7III. It makes you wonder why they even bother putting such large sensors into "professional" cameras, just make it like 0.3mm x 0.2mm and be done with it. It's probably just a marketing ploy to sell you more useless silicon.

Anonymous

Anonymous Sat 16 Nov 2019 10:10:52 No.3569795 Report

Quoted By: >>3569803 >>3569840 >>3570004

>>3569487
Noise depends mostly on the level of sensor technology and photosite size.
However higher resolution means you can use more noise reduction before the image goes to shit.
Therefore a larger sensor will give you a better image quality for the light hitting it.

However lens development in MF is pretty much nonexistant and the sensors are a rather small niche market.
Anyway, just like film they use the same basic technology just at a larger scale for larger formats.
Meaning the photosites of these sensors will probably be the same as used in some other camera with a smaller sensor.

TL;DR
MF digital should have the same noise at a 100% crop as a smaller equivalent sensor at the same ISO but a higher resolution making it less visible.

Anonymous

Anonymous Sat 16 Nov 2019 10:15:57 No.3569797 Report

Quoted By:

>>3569169
I will when you give me the source of that pic

Anonymous

Anonymous Sat 16 Nov 2019 10:22:39 No.3569803 Report

Quoted By: >>3569808

>>3569795
>Noise depends first on the total sensor surface area and second on the technology used to fabricate the sensor. Individual pixel size is almost irrelevant to the equation.
FIFY

Anonymous

Anonymous Sat 16 Nov 2019 10:34:31 No.3569808 Report

Quoted By: >>3569815 >>3570004 >>3570587

>>3569803
That is not correct.
The noise depends on the signal/noise ratio of the individual photocell, not how many photocells you have.
The larger the photocell, the stronger the signal and the better the signal/noise ratio.
However the higher resolution will allow for more noise reduction before the image quality suffers too much.
As an example:
If you use an APS-C crop sensor camera and a FF camera with identical sensor technology and architecture, both will have the same noise but different resolutions.

Anonymous

Anonymous Sat 16 Nov 2019 11:02:57 No.3569815 Report

Quoted By: >>3569817

>>3569808
>That is not correct.
It absolutely is correct.

>The noise depends on the signal/noise ratio of the individual photocell, not how many photocells you have.
The number of photocells is mostly irrelevant. The physical size of the array however is very relevant.

>However the higher resolution will allow for more noise reduction before the image quality suffers too much.
True but irrelevant. You don't need to apply NR to see that a FF camera has the expected SnR gain over an APS-C sensor with the same pixel pitch.

>If you use an APS-C crop sensor camera and a FF camera with identical sensor technology and architecture, both will have the same noise but different resolutions.
Observably false. Compare the 5Ds to the 7D mark II as one example.

Anonymous

Anonymous Sat 16 Nov 2019 11:07:42 No.3569817 Report

Quoted By: >>3569828 >>3570004

>>3569815
Explain to me how on earth the same photosites should produce less dB of noise when there are just more of them.

Anonymous

Anonymous Sat 16 Nov 2019 11:24:46 No.3569828 Report

Quoted By: >>3569831 >>3569874

>>3569817
To be clear the larger sensor has higher SnR because it's collecting more light which is signal. But the reason pixel size is mostly irrelevant is because with gapless mircolenses there's no loss of light/signal when you divide the silicon into more pixels. Smaller pixels do have more read noise, but 4 less clean samples average out to 1 cleaner sample. You don't even have to use software to scale the image down because at the same view size your eye/brain are averaging out the individual elements.

Anonymous

Anonymous Sat 16 Nov 2019 11:30:39 No.3569831 Report

Quoted By: >>3569834

>>3569828
If you average out 4 pixels to one, that's allready a rather rudimental noise reduction.

Anonymous

Anonymous Sat 16 Nov 2019 11:36:00 No.3569834 Report

Quoted By: >>3569840

>>3569831
In a sense. But you don't have to engage NR software. The fundamental point is that pixel size really doesn't matter at this point in time. Sensor surface area does. At one time it did matter because there weren't gapless microlenses. The smaller the pixels, the greater the light loss.

I keep saying "mostly" because at some extreme read noise would scale out of proportion to the pixel size.

Anonymous

Anonymous Sat 16 Nov 2019 11:42:53 No.3569836 Report

Quoted By:

>>3569549
This is correct

Anonymous

Anonymous Sat 16 Nov 2019 11:57:24 No.3569840 Report

Quoted By: >>3569874

>>3569834
But a software scaling down images by averaging 4 pixels to one performs noise reduction.
As I said in my original post:
>>3569795
>MF digital should have the same noise at a 100% crop as a smaller equivalent sensor at the same ISO but a higher resolution making it less visible.

Anonymous

Anonymous Sat 16 Nov 2019 13:14:31 No.3569874 Report

Quoted By:

>>3569840
>>3569840
>But a software scaling down images by averaging 4 pixels to one performs noise reduction.
And as I said in this post: >>3569828
>You don't even have to use software to scale the image down because at the same view size your eye/brain are averaging out the individual elements.
Software NR is not required.

giannis

giannis Sat 16 Nov 2019 18:57:52 No.3570004 Report

Quoted By: >>3570113

>>3569795
>>3569808
>>3569817
Don't bother with him.
Your understanding is perfectly correct.

The fallacy the other anon usually falls onto, is outputting at the same size (not 100% crop each).
Outputting at the same size, means combing pixels on the large camera by downsizing, which is effectively very close to using larger photosites in the first size, because you're adding/averaging the signal of, say 4 photosites, when you downscale 2x in each dimension.

I'd say you won't change his mind no matter what, because he keeps repeating "larger sensor area produces better signal", which is effectively treating the whole sensor as a huge single photosite.
To make it clear, yes, if you use the whole sensor to output a single pixel, the bigger sensor will have better snr. But at the photosite level the snr is the same.
Nowhere in the formal equations does "array size" appear, it's all about photosite size and technology (which determines dark current, shot noise, quantum efficiency, etc. etc., all factors in the equation).

Anonymous

Anonymous Sat 16 Nov 2019 22:54:54 No.3570113 Report

Quoted By: >>3570564

>>3570004
>it's a fallacy to print to a common size because people always and only print their photos at 96ppi and 90"

>which is effectively treating the whole sensor as a huge single photosite.
No, it's not. (Though for the sake of signal detection you could treat it as such using software.)

>>3570004
>Nowhere in the formal equations does "array size" appear
You've very clearly never worked with the equations as collection device size is part of any SnR equation involving signal capture.

>because he keeps repeating "larger sensor area produces better signal"
Photons are the signal. All other factors being equal, a larger sensor collects more photons. It's really not that hard.

Anonymous

Anonymous Sun 17 Nov 2019 17:02:56 No.3570564 Report

Quoted By:

>>3570113
by your logic a bigger sensor captures even more noise.... that we then use to average out. did you even read what you replied to?

Anonymous

Anonymous Sun 17 Nov 2019 17:37:19 No.3570587 Report

Quoted By:

>>3569808
>If you use an APS-C crop sensor camera and a FF camera with identical sensor technology and architecture, both will have the same noise but different resolutions.
This. But for normal shutter exposures. Long exposures will act differently since larger sensor will allow for shorter exposures, and they will even rely on image processor technology.
Two cams with same sensor size, with same sensor technology, but different image processor, may give up different noise readings.

Capcode	All Only User Posts Only Verified Posts Only Moderator Posts Only Manager Posts Only Admin Posts Only Developer Posts Only Founder Posts
Show Posts	All Only With Images Only Without Images Only Spoiler Images Only Non-Spoiler Images
Deleted Posts	All Only Deleted Posts Only Non-Deleted Posts
Ghost Posts	All Only Ghost Posts Only Non-Ghost Posts
Post Type	All Only Sticky Threads Only Opening Posts Only Reply Posts
Results	All Grouped By Threads
Order	Latest Posts First Oldest Posts First

On these archives

Your latest searches