Digital cameras

[Skii]

So i'd be advised to get myself an external flash card reader then. Got a Cannon camera being delivered some time this week.
While we're on the subject. The camera I'm getting doesn't have any optical zoom, only digital. onsidering I'm only gonna use it for simple pictures am I likly to nitice the difference much, I gather digital zoon just 'zooms' into the picture after its been taken.

I don't know TBH Stubzz

I might be generalising, however when i highlighted this particular problem in a previous thread it was suggested that this is common with Canon cameras.
Worth considering anyway.

[Rave]

While we're on the subject. The camera I'm getting doesn't have any optical zoom, only digital. onsidering I'm only gonna use it for simple pictures am I likly to nitice the difference much, I gather digital zoon just 'zooms' into the picture after its been taken.

Well, a digital zoom just reduces the area of the sensor that's used to capture the picture. If you zoom in by 2x with a digital zoom then only a quarter of the sensor is being used, so for example a 3mp camera would become a .75mp one. The only point of using the digital zoom is to reduce file sizes; otherwise you can just crop the image later in your computer and get the same result.

Rich :¬)

[steve threlfall]

Doesn't show up as a removeable disk in 'my computer' m8, so unless you use the 'easy to use' software to download the images there is no way of accessing the images off the camera, the only other way around it is to take the compactflash card out and use a card reader.

I've always never used the software that comes with cameras, I prefer to go straight into the camera (showing as an additional storage device next to your hard disk) and drag and drop them onto my desktop, I can then chose to open them with photoshop if needs be. I can do that no problem with my Fuji, but not with my Sister-in-laws new Canon.

Very frustrating.

Ah right, thanks for the tip dude.

I think i have narrowed it down to two models now.

http://www.amazon.co.uk/exec/obidos/...15&c=93217&s=1 pentax optio 33LF

Or the minolta dimage x20. http://www.cameras2u.com/products/de...DUCT=V2787101A

Theres only a fiver in it. The pentax is 3.2 and the minolta 2.0. Im leaning towards the minolta from cameras2u as its smaller and ligter.

Cheers for the help evryone

[Rave]

I like the Minolta and considered getting one for a while, but it's a bit of a funny shape to actually hold and use. The Pentax is still a very small camera and there might be less difference in size/weight than you imagine. Your best bet might be to see if Jessops will match those prices, if so you can try them back to back in the shop and see which you prefer.

Rich :¬)

[steve threlfall]

I like the Minolta and considered getting one for a while, but it's a bit of a funny shape to actually hold and use. The Pentax is still a very small camera and there might be less difference in size/weight than you imagine. Your best bet might be to see if Jessops will match those prices, if you you can try them back to back in the shop and see which you prefer.

Rich :¬)

Thats the idea bud

Im not in any huge rush to go out and get a camera or anything. I want to actually hold the camera before buying it and i shall be using the pricematch if possible.

[faisal_uk]

i would say go with the minolta, alot of people have said it takes good quality pics and its nice and small.

just thought i would let you know there is a slightly newer version out now: the Minolta Dimage X21 - here at pixmania: http://www.pixmania.com/dev/gui_web/...&article=41574

not sure what the difference is but it's only £106 inc delivery

[Knoxville]

to be fair dude the pentax is probably the one i'd go for, the minolta's smaller but the optical zoom is imo something really worth having, looking back on the picture's i took at download now some of them could be so much better with a zoom on the camera, next time a reasonably priced digital camera will be going instead of my disposables.

I wanted a small camera when i got my digital but i realised that i'm not really gonna carry it every day so it didn't matter too much, that combined with the fact the most of the small ones around when i bought mine were worse than some camera phones.

[faisal_uk]

to be fair dude the pentax is probably the one i'd go for, the minolta's smaller but the optical zoom is imo something really worth having

yeh that's the beauty of the minolta, its small but still has 3x optical zoom - which most cameras <£100 don't have.

[Knoxville]

didn't think the minolta had the optical zoom

[Butuz]

<Snip> Huge technobabble rant</Snip>

Rich :¬)

Omg!

*hides behind the sofa and waits for Saracen to wake up*



Butuz

[Saracen]

.... a traditional CCD sensor arranges its pixels into square grids containing one red, one blue and two green pixels (or, more precisely, pixels which are sensitive only to blue, green and red light). So, a 3mp sensor will have roughly 750,000 each of blue and red sensitive pixels and 1.5m green sensitive pixels. There's a good reason why there are twice as many green pixels BTW, it's because in practice most of the light from the sun (and from a lot of artificial light sources too) is in the green range, and consequently our eyes are best at detecting subtle differences in green light, so it makes sense to ensure the CCD is most accurate in the green range.

Yes, agreed. It's down to the colour cone cells in our eyes. The cells have poor sensitivity and require a lot of light, unlike the rods which have high sensitivity but poor colour resolution. That is why our sight has poor colour accuracy at night - we are relying primarily on rods not cones. Rods also give most of our peripheral vision, as cones are concentrated in the one relatively small area of the retina. There are, however, three types of cone cells and their differing shapes determine what wavelength of light they are sensitive too (long, medium and short). Rods are all of the same shape and are therefore monochrome.

So far so good, but the thing is that a 3mp camera actualy outputs an image containing 3 million pixels and each of those pixels contains a discrete red, green and blue value, usually in 8 bit precision for each channel. The camera actually has to do a calculation to map a value for the other two colours onto the value for each individual coloured sensor on the CCD. I guess it does this in a similar way to interpolation.

Yes, the camera maps, but no, it isn't like interpolation. Each pixel in the image is calculated from four colour values (two green, one red and one blue).

The camera takes the first two pixels from the first line (green and red) and the first two pixels from the second line (blue and green) and calculates a single IMAGE pixel value from them, i.e. from the readings from four physical sensor elements. It then moves along ONE element, and does the same again. It is using physical values, not made-up ones as in interpolation.

So it's now using red and green from the first line, green and blue from the second line. Note that the red in the first line is used to calculate the first RGB pixel and value and then used AGAIN in the second pixel, as is the green in the second line. When the first line is done, the process moves down a line and starts again on the second line, combined with the third line. So EVERY mono pixel value in the second line gets used AGAIN. Most of the pixel values generated by the actual CCD (or CMOS) sensor elements will get used two or even three times on adjacent pixels. The exception is, as I obliquely referred to above, the Foveon chip which works in a completely different way. The Foveon chip, incidentally, is named after the Fovea, which is the part of the eye directly opposite the pupil, contains mainly cones cells, and is primarily responsible for colour perception.

So, each image pixel is calculated from a matrix of four sensor values, and the matrix overlaps with that used by other pixel elements.

So, the point is that for a 6x4 print at 200dpi, you actually only need about a 1.3mp camera. A 2.1mp camera will give you nearly 300dpi and will look a bit sharper, but 3mp is basically overkill for 6x4 prints.

You've got me there. I shouldn't do that kind of post when tired. The maths on the Digital Darkroom post I linked to, though, is right. I've amended my original post here.

The golden rule, I guess, is to look at the resolution, in terms of image produced by a camera, and to ensure that it's optical not interpolated. Then divide that resolution by 200 (for 200doi) and you'll get the maximum print size (without degradation).

So for 1600x1200, you can print at 8"x6", while for 3000x2000 (6 Megapixel), you can print at 15"x10" (assuming you have a suitable printer). This 200 dpi figure is subject to some debate. Some people claim to get away with 150dpi, whereas others claim 300dpi is necessary. My view is that 150 dpi is pushing it, but 300 dpi is overkill except for very demanding circumstances.

Also, if you get to very large prints (like 15"x10") the odds are they'll be hung on a wall and viewed from a distance anyway, so you can drop from 300doi, even for exhibition quality and get away with it (usually).

Well, not only do they quote double the resolution, they actually interpolate the output image to double the resolution- they have to. The octagonal pixels are arranged in diagonal lines; however no image file format will allow you to store an image with the pixels arranged diagonally so Fuji have to 'fill in the gaps' to give a 'normal' output file. Time for my attempt at a diagram:

o-o-o-o
-o-o-o-
o-o-o-o <--- octagonal orrangement of pixels on CCD (hyphen = gap)
-o-o-o-

ooooooo
ooooooo
ooooooo <--- output file with gaps filled in by camera
ooooooo

I agree with your comments about it being a 50% improvement, but what you actually have is a 6mp file with equivalent sharpness to a 4.5mp file from a normal camera- consequently you pay a penalty in file sizes, although with flash memory so dirt cheap now this is less of a concern than it once was. Ironically, Fuji claim that the octagonal shape of the pixels allows them to be packed closer together which leads to greater sensitivity and in theory lower noise; however in my experience the extra post processing that's necesary makes the images from Fuji cameras some of the noisiest around.

That's not quite how Fuji explained it to me, but you could well be right. The implications are much the same, certainly.

[Stoo]

didn't think the minolta had the optical zoom

The X series cameras don't look big enough for an optical zoom do they.. but it's there, they've done a bit of clever designing to it's actually in the side of the camera.. two secs and I'll do a diagram..

Edit: why bother when Minolta already made one

[acidrainy]

Personally I use a Fuji s602z. I got this after many recommendations for Fuji’s lens quality and indeed the camera model itself.

That aside. I'm very happy with the camera; it surpasses the quality of any of the "point-and-click" 35mm's that I own anyway!
After being so pleased with the camera, and indeed the manufacturer, I recommended the brand to a friend in your position. He bought a Fuji F410 under his budget. Having played about with it, and even other cameras in this price range, I have to admit it’s a very nice camera. The quality of the shots compared to rival manufacturers in this price bracket (Cannon, Nikon etc’) is astonishing. The sharpness, colour accuracy and general appearance of shots do appear better imho.

So on that note, here are 2 recommendations for the Fuji F410. One from my delighted friend, the other from myself. Hope you’re happy on whatever you go with!

[acidrainy]

Just a quick followup to my last post there. When I was looking for a camera the two most influential websites that helped me make my decision were:

http://www.dpreview.com/ (Great reviews of all sorts of kit at many different locations on the budget spectrum)

http://www.digi-darkroom.com/ (Excellent for those "I know this may be a stupid question but" related enquires!)



Note: Another pimp for the digital darkroom there Saracen

[joshwa]

the Optio S/S4/S4i looks a lot smaller than the Minolta X/Xt/Xg in this photo:



http://www.dpreview.com/reviews/pentaxoptios/page3.asp

[Rave]

Omg!

*hides behind the sofa and waits for Saracen to wake up*

ROFL! I've tangled with Saracen before, luckily for me he doesn't seem to take it personally.

The camera takes the first two pixels from the first line (green and red) and the first two pixels from the second line (blue and green) and calculates a single IMAGE pixel value from them, i.e. from the readings from four physical sensor elements. It then moves along ONE element, and does the same again. It is using physical values, not made-up ones as in interpolation.

So it's now using red and green from the first line, green and blue from the second line. Note that the red in the first line is used to calculate the first RGB pixel and value and then used AGAIN in the second pixel, as is the green in the second line.

<slaps forehead> How obvious, and yet the moderate mental leap needed to figure it out eluded me. You learn something new every day.

Most of the pixel values generated by the actual CCD (or CMOS) sensor elements will get used two or even three times on adjacent pixels.

Well, presumably all the pixels except the ones on the very edge get used four times each?

That's not quite how Fuji explained it to me, but you could well be right. The implications are much the same, certainly.

Well, I guess that once it's filled in the gaps it then has to go through the same procedure of creating the image pixels as a normal CCD. What I'm now wondering is why they bother to give a 6mp output image when they could probably get the same results with using a 3mp picture with a higher than normal amount of deatail. I'll have to give the matter some more thought I think, although I guess it could be mainly for marketing purposes, or to try and eliminate jaggies on larger prints.

The exception is, as I obliquely referred to above, the Foveon chip which works in a completely different way. The Foveon chip, incidentally, is named after the Fovea, which is the part of the eye directly opposite the pupil, contains mainly cones cells, and is primarily responsible for colour perception.

I take it back- I've learnt two new things today. That Foveon sensor is amazing, I agree. DPReview seemed to think it had odd clipping behaviour in the Sigma SD9, but if they could sort that out and make a 10mp version which was full frame 35mm size it'd make an incredible camera!

Rich :¬)