Technical Question

gman

gman

Messages
11,574
Name
Graham
Edit My Images
Yes
Had a conversation with a guy a few nights ago who says he doesn't shoot at full resolution but shoots at a slightly lower one from the settings within the camera, kinda like instead of shooting at Super Fine he'll shoot at Fine.

He says he gets better shots by doing this but I can't get my head around how and wondering if I'm missing something or if it's just his perception? :thinking:

This got me thinking hehe...wait for it! lol

Ok, let's take the Canon EOS400D for an example. It's got a 10.1 million effective pixels CMOS sensor which is 22.2 x 14.8mm in physical size.

Obviously the physical dimensions of this sensor can never change.

So as I understand it a 10MP sensor which is of a larger dimension to the EOS400D should handle noise better because the photodiodes (or something?) are larger because they are less prone to noise as more light can hit each one - or something like that?

So going back to the EOS400D sensor, if I reduced the quality so it's shooting at say 5MP does that mean the photodiodes are operating at twice the size and there less prone to noise or is the camera simply cropping the area of the sensor used?

I suspect the latter as I presume the photodiode's size can't change either but am still curious about all this!

Also whilst on this topic, am I right in understanding a 10MP sensor of a larger physical size can give a better dynamic range than a 10MP sensor on a smaller physical size because the photodiodes are larger and can therefore take in a greater range of light so to speak before they overflow and cause overexposure?

Or am I going mad with all this? lol
 
btw, apologies in advance if this is a stupid thing to ask but I'm kinda working my way through a case of Stella provided by work! hehe
 
I'm waiting for the answers to this too, was interesting because I heard that conversation and was intrigued as to how that was.
 
I'm thinking this might be an urban myth but I will stand corrected if someone can give a proper scientific explaination. Eitherway if this is a way to gain quality in one area I suspect there would be a trade off in another.
 
The output file size in Jpeg is not related to the number of pixels used. If you take a shot of a black or white surface you'll have a smaller output file than you would with a complex composition.

Bob
 
Whether it's shooting at a higher compression or shooting at a lower resolution, I'd always prefer to pick the highest quality (least JPEG compression, largest file size) in camera and deal with it later on the computer. Using a lower resolution setting on the camera only means that it downsizes the image after it's taken. That's something one could deal with on the computer more effectively and try different methods of downsizing without being stuck with it. The same goes, IMO, for digital zoom on a P&S.
 
dcash29 said:
I'm thinking the conversation was a waste of time?
Click to expand...

I'll remain civil and simply say go somewhere else then :cuckoo:



@ Canon Bob, your explaination makes sense so I take it the camera always uses its full capability it's just the processing which is changed by altering the in-camera settings?

But then how does shooting in crop mode allow a higher fps (for certain cameras)? I would have though it's the actual shutter itself in hardware terms that determines the fps and not any processing that happens - which I understand would only effect the duration which you can shoot at continuous for, not the actual speed.
 
gman, most of your musings are correct.

Other things being equal, larger pixels produce higher dynamic range and lower noise.

Basically, what's going on is that each pixel is like a little bucket for capturing and counting photons. The bigger the bucket, the more photons it can hold before it's full, and therefore the more different output signals it can generate - this is where the dynamic range comes from. Also, electrical and thermal noise acts as a sort of random element of +/- a few photons on each count. So the bigger the bucket, the higher the photon count and the less is the relative effect of noise. (This also explain why noise shows up in dark areas of photos; there aren't many photons there anyway, so a small random error makes a big difference.)

Note however those weasel words other things being equal. One important factor is how modern the technology is. Newer sensors will have more efficient and accurate photon counting machinery and suffer from less noise. So whilst bigger is better when you're comparing sensors of the same vintage, it doesn't necessarily hold true if you're comparing a modern high-density sensor with an older low-density sensor.

Now back to your friend. If his camera is anything like my 350D he has two image parameters to play with; the resolution and the JPEG compression. I have 6 different image quality settings available: three different resolutions (Large = 8MP, Medium = 4MP, Small = 2MP) and two compession levels (Normal and Fine).

Choosing the Small or Medium image size does improve the quality compared to the Large size. Each pixel in the image is created by combining 2 or 4 pixels from the sensor, so it behaves as if the pixels were larger. On the other hand you could do that in PhotoShop - reduce the image in size from 8MP to 4MP or 2MP - and you'd probably get the same benefits. So to my mind the value of this is questionable - other than to give you more shots on each memory card, if you know you'll never want to print them large.

Choosing the Normal setting instead of the Fine setting, however, degrades the image. It simply applies a higher level of compression to the JPEG, and this is destructive and non-reversible. Again, it has the benfit of giving you smaller files and more images per card, but there is no conceivable argument in favour on the grounds of image quality.
 
I'll remain civil and simply say go somewhere else then [/QUOTE

Gman, as merely putting across that i didnt think what the guy told you was correct

But take it how you want
Click to expand...
 
gman said:
But then how does shooting in crop mode allow a higher fps (for certain cameras)? I would have though it's the actual shutter itself in hardware terms that determines the fps and not any processing that happens - which I understand would only effect the duration which you can shoot at continuous for, not the actual speed.
Click to expand...

Once the shutter/mirror has been eliminated from the equation then the buffer transfer rate becomes the limitation. Improved throughput can be achieved by lowering the ISO or reducing the quality (increasing the compression) which results in less data being transferred.

Bob
 
dcash29 said:
I'll remain civil and simply say go somewhere else then [/QUOTE

Gman, as merely putting across that i didnt think what the guy told you was correct

But take it how you want
Click to expand...

Dcash - your statement that the conversation was a waste of time led us to think that you were saying THIS conversation was a waste of time. It wasn't immediately apparent that you were referring to gman's conversation with the other person.
Click to expand...
 
gman said:
Also whilst on this topic, am I right in understanding a 10MP sensor of a larger physical size can give a better dynamic range than a 10MP sensor on a smaller physical size because the photodiodes are larger and can therefore take in a greater range of light so to speak before they overflow and cause overexposure?
Click to expand...

Here's a Nikon APS-C sensor (in this case a D100) against the sensor from a point and shoot.

2748503640_91923d54fc_o.jpg
 
Ah, thanks Stewart and Bob. It's making sense now and also explains why when you reduce a noisy image in ps it instantly looks a little better - as well as an ISO1600 shot looking perfect in the view finder only to look horrible on screen! lol

Dcash, yes it was the way you put it. Wouldn't say it was a complete waste of time as it's resulted in me having a better understanding of how the camera sensors work! :)
 
I guess quoting the section may have helped.
But if i'd have thought the thread was a waste of time i'd have said the thread. But thanks for replying:thumbs:

If i remember correctly this noise and pixel conversation came up years ago concerning the Fuji prosumer cameras on another site.
This is where the cams used to interpolate 6mp to 12mp i think then the outcome was that the cam produces the maxium and reduces as cams quality setting.
 
Hmmm, I done a fair bit of searching on google and am struggling to find a really good explaination or guide about all this sensor stuff and photodiodes etc. Does anyone know of anywhere that shows how all this works? Don't know why I'm suddenly so interested, guess I'm not all that happy with my ISO performance what with seeing all this D3 stuff! hehe
 
It's good to know about these kind of things, knowledge and all that :) Although it does seem very technical.
 
StewartR said:
Each pixel in the image is created by combining 2 or 4 pixels from the sensor, so it behaves as if the pixels were larger. On the other hand you could do that in PhotoShop -
Click to expand...

Good to have a clear explanation. :thumbs:..

So would combining the photon buckets help reduce the signal to noise ratio?...or is it the same... I'm thinking maybe theres some noise reduction going on whilst it being combined or somesuch. :thinking: ...hence the myth.



Toothie said:
but am i right in thinking this is not at all useful if you are using RAW?
Click to expand...

Yes your right Ruth, nowt difference to raw.
 
You must log in or register to reply here.
Back
Top