S5 high iso and DR

I hope someone around here can help me with this. I’m trying to figure out if the S5 might be a good camera for my needs.

I currently us a D70s and I’m still pretty satisfied with it. I don’t think I’m the typical S5 shooter. I don’t do portraits or weddings or things like that (check the gallery in my sig if you want to get an impression). The only people I photograph are my family. But it is especially those shots I would like to improve. I’d like to have better high iso results and better skintones.

In fact, I don’t think the D70s high iso’s are that bad, but I really have to expose to the right. And unfortunately most of the time a want to get a brighter exposure but can’t because I would severely blow the highlights. Apart from the overall better high iso performance of the S5 I think that I might be able to push the exposure a bit more and still be able to recover highlights. (Apparently some people overexpose up to 1 stop to get better shadows with the S5?) Am I right so far?

I recently found out however that the R-pixels (?) are not active above ISO1000. Does that mean that the S5 no longer has a DR advantage and that my reasoning wouldn’t be true for anything over ISO1600?

--
My gallery: http://www.flickr.com/photos/klaastuin/

de_klaas said:

I hope someone around here can help me with this. I’m trying to
figure out if the S5 might be a good camera for my needs.
I currently us a D70s and I’m still pretty satisfied with it. I don’t
think I’m the typical S5 shooter. I don’t do portraits or weddings or
things like that (check the gallery in my sig if you want to get an
impression). The only people I photograph are my family. But it is
especially those shots I would like to improve. I’d like to have
better high iso results and better skintones.

Click to expand...

First....I want to congratulate on your images. You have a good eye and some lovely images. I have owned a few Nikons including your D70. If you are happy with the D70 you will be even more happier with the S5. It has better skintones and tonality. And it has better noise performance than the D70. ALso you have a better viewfinder, LCD, body and way more color parameters to choose from.

de_klaas said:

In fact, I don’t think the D70s high iso’s are that bad, but I really
have to expose to the right. And unfortunately most of the time a
want to get a brighter exposure but can’t because I would severely
blow the highlights. Apart from the overall better high iso
performance of the S5 I think that I might be able to push the
exposure a bit more and still be able to recover highlights.
(Apparently some people overexpose up to 1 stop to get better shadows
with the S5?) Am I right so far?

Click to expand...

Yes. There is a lot more latitude with the S5.

de_klaas said:

I recently found out however that the R-pixels (?) are not active
above ISO1000. Does that mean that the S5 no longer has a DR
advantage and that my reasoning wouldn’t be true for anything over
ISO1600?

Click to expand...

There is a drop off in DR but if you shoot with any other camera you will have a drop off as well. And I don't know if the R pixels are not active at above ISO1000.... but all I do know is that it is still the king compared to other dslrs especially in jpeg even at these high ISOs. If you look at the review of the S5 on DPR, it has better DR at ISO1600 than the D80 has at ISO100! The S5 has 9.8ev range at ISO1600 in jpeg whereas they were only able to squeeze 8ev out of the D80 at ISO100!

*****************************************
Packy

While I don't have any personal experience with a D70 and only have an S3, I have a friend who uses a D70 and I think you'll be pleased with the difference in IQ, including high ISO performance that the s5 offers.

From test analyses I've seen on the web (do a search of S5 reviews), the high DR advantage holds up into the high ISO range. It may not have quite the DR in the upper ranges, but it's still an advantage over the other brands.

Regardi
--
With kind regards,

Robert05 (AKA Fleming)
http://www.pbase.com/robert_in_sc

IMHO it keeps DR advantage above ISO 1000 but loose the extra resolution.

R pixel are still used but only instead or in combination of clipped S pixels, like for a 6MP output.

Didier

copied from a previous post (by myself):

"even the S5 manual says that R pixels are not used in hi iso, but, i have noticed that when changing the Dynamic range setting on a raw file at hi iso (even 3200) you see differences.. also on S7RAW rendering iso3200 raw files with custom S+R mix makes difference as well, so it seems that the R pixels are in use even at HI ISO..."

I upgraded from d70s to s5. one of the things i use to shoot is hi iso (music and theatre) the s5 is well worth the upgrade, i loved my d70s but i think the s5 is a much better camera, is on another level.

the new images with the s5 are definitely better, so if you think the d70s is holding you back (like i did) then the s5 will cover that gap and leave room for you to improve...

you have to consider that the raw files on the d70s are 5mb while on the s5 are 25mb so thats 5 times more memory you will need...

all best

But many feel that the S5 jpegs are so good that you may not wish to use
S4 RAW shooting. Like Robert aka fleming I am an S3 shooter, and find
that the jpegs are amazingly accurate for what i do . . .

Keith

--

cast your dancing spell my way,
I promise to go under it . . .

I am not sure about that..as I recall the S5 file size drops to 12mb which would indicate the little pixels are turned off??

Can someone help here? Certainly the S3 retains it's full file size up to and including 1600asa.

Another question which keeps coming up..resolution...It is claimed that by "mapping the S+R pixels" to the same location makes the S5 "a 12mp camera"??..How??

And how is this different to the S3. I understand the S5 eeks out a little more resolution on test charts but I could believe this is because of the new filter design and better processor.....?

As ever, there never seems to be clear explanations of these aspects and Fuji seems reticent to talk about the design changes between the models in any detailed way...just interested, that's all!!!

R-pixels are not used above ISO1000 in the S5,

The extra spatial resolution in the S5 is not due to the R-pixels,
but the octagon-shape of the pixels,
8-9MP (or more precise 6MP x square root of 2, if you like mathematics)

I don't mean this as a challenge - I just am honestly interested and would be happy for your answer to hold...

So, because of the honeycomb pixel layout does it really yield 6 times square root of 2 or is this possibly some Fuji P.R."fuzzy-math"?

The people at Phase One for instance, at least when talking about S2 data, were adamant that there was only 6 mp of data in the RAW files. DxO also defaults to 6 mp output.

ACR/Lightroom and HU of course both default to 12 mp.

So is there really some truth to the claims that it's more than 6 mp (even if it isn't a full 12 - which I think no one really believes)? Is there a doc or post somewhere you could point to that really substantiates that?

Just a personal interest if you're so inclined.

Thanks!

--
Matt Fahrner
http://www.boinkphoto.com

no its not PR,

look at the resolution measurements att dpreview.com,
or any other serious camera review with resolution figures,

S5 resolutions matches or slightly exceeds the 8,2MP canon 20D,

in the end, its the amount of information the pixels hold that is
interesting, not the exact number ,

cheers,
=)

scathew said:

I don't mean this as a challenge - I just am honestly interested and
would be happy for your answer to hold...

So, because of the honeycomb pixel layout does it really yield 6
times square root of 2 or is this possibly some Fuji P.R."fuzzy-math"?

The people at Phase One for instance, at least when talking about S2
data, were adamant that there was only 6 mp of data in the RAW
files. DxO also defaults to 6 mp output.

ACR/Lightroom and HU of course both default to 12 mp.

So is there really some truth to the claims that it's more than 6 mp
(even if it isn't a full 12 - which I think no one really believes)?
Is there a doc or post somewhere you could point to that really
substantiates that?

Just a personal interest if you're so inclined.

Thanks!

--
Matt Fahrner
http://www.boinkphoto.com

Click to expand...

Thanks for all the input!

Unfortunately I won't be buying a S5. I wrecked our car yesterday and will have to buy a new one. Cameras will have to wait...

As for resolution: I think I read something in either Phil's review or in Thom Hogan's review about how the diagonal layout gives better vertical and horizontal resolution.

--
My gallery: http://www.flickr.com/photos/klaastuin/

its totaly clear in the dpreview-review that the resolution figures for S5 is
comparable to a 8+MP camera,

http://www.dpreview.com/reviews/fujifilms5pro/page27.asp

and the older S2 (that have no R-pixels) is almost in the same league:

http://www.dpreview.com/reviews/FujiS2Pro/page19.asp

so its all about the octagon pixel shape and the overlapping pixel layout,

and you need 8,5MP file to express the resolution inherent in the origiinal 6MP file,

so forget about the R-pixels for spatial resolution, but for DR they do miracles.

interpolating the fuji-files to 12MP compared to 8MP wont increase resolution,
but its Fujis choise, and with some sharpening, it looks good in print,
though jaggies in diagonal lines is the weak spot of the S5,
but this is hardly an issue in portraits,

=)
cheers

de_klaas said:

Thanks for all the input!

Unfortunately I won't be buying a S5. I wrecked our car yesterday and
will have to buy a new one. Cameras will have to wait...

As for resolution: I think I read something in either Phil's review
or in Thom Hogan's review about how the diagonal layout gives better
vertical and horizontal resolution.

--
My gallery: http://www.flickr.com/photos/klaastuin/

Click to expand...

. . . Is that when I was using a D80 last year, the images ( talking jpegs only )
were much more fragile than Fuji S3 files in PP. The robustness of Fuji files
was fascinating in comparison to the way that resizing ruined D80 images.

I realise that RAW files would have been different, but one of my reasons for

drinking the Fuji Koolaid is that the cams produce such great jpegs, my preferred
image choice.

Keith
--

cast your dancing spell my way,
I promise to go under it . . .

It is true that the Fuji resolution is at max in the vertical and horizontal..being that their pattern of pixels is "angled" at 45%..and when interpolated is also the reason why the resolution may be expressed as root2 x pixels apparently...or that figure we all remember from school, 1.414!!

If you multiply the pixel dimension by 1.414 you get 4242 x 2828=12mp...nothing to do with small pixels, which on the S5 are also mapped to the same location as the big ones. True also if you speak to Capture One or DxO they will state the Fuji is a 6mp sensor, which it is...but their programmes use all the file information from the camera..25mb, which of course includes the output from the R pixels as well. All a bit of a black science if you ask me which nobody to date has given me an explanation that I can understand. I do know that the DxO programme, despite it's 3000x2000 output still produces the best resolution and sharpness of any..and easily scales up to match and indeed better both HU and ACR for instance..

At high iso's it also resolved more information with less noise (a lot less!) and sharper too than the D300..so something is going on!!! Can anyone actually extract the facts from the years of myths....Maths please...not Myths!!

But here is a trick that I discovered in Photo Zoom,

After processing and saving from your RAW processor, you run the file through Photo Zoom S-Spline, keeping the same image/file size and turning off their USM.

The 45 degree steps disappear through the S-Spline interpolation algorithm....you keep all the original sharpness but it just takes the image just one step nearer that smooth detailed film look.

It would be interesting to discover if re-interpolating the file in any programme takes out the jaggies and steps, cos it certainly does in Photo Zoom..just another step closer the the full 12mp look??

OldMaster said:

The 45 degree steps disappear through the S-Spline interpolation
algorithm....you keep all the original sharpness but it just takes
the image just one step nearer that smooth detailed film look.

Click to expand...

OldMaster said:

It would be interesting to discover if re-interpolating the file in
any programme takes out the jaggies and steps, cos it certainly does
in Photo Zoom..just another step closer the the full 12mp look??

Click to expand...

interpolation can be done more or less precise,
but youll never get 12MP resolution from a S5,

I would like to add my observations on this. It is possible to make a lower pixel image appear to be of a higher resolution by working it in an editing program. I have done this with a low res image by first uprezing to whatever size and then doing an unsharp mask. So combining an uprezed S2,3,5 image (from 6MP to 12MP) and doing a good internal camera or HU treatment of the image, it could create a fairly good advantage over a straight 6MP image. That doesn't make it a 12MP sensor but just good internal editing.
Will

Correct of course, but I was saying by removing the tendency the Fuji has of "stepping" 45 degree detail due to it's interpolation method brings it closer to the look of a higher resolution camera..which it already has in vertical and horizontal information.

There is no more detail, except in the sense that picture elements that were stepped are now smoothed out giving the impression of more resolution..and an overall improved image....which after all is what it is all about??

After all, as we know, having too many pixels on a small chip brings with it other problems..I would rather have fewer pixels and better IQ personally...lower noise, better per pixel sharpness and less of a test on the lenses.

Interestingly the latest report coming out of Luminous Landscape concur. Once again the Nikon D3 showing itself capable with 12mp of resolving more detail and being sharper than the 21mp Canon...no surprises there I have to say!!

Snoddas Date/Time 11:00:10 PM, Friday, December 07, 2007 (GMT)

"R-pixels are not used above ISO1000 in the S5, "

oh yeah you know so much man!

so how can you expain what i said? have you tried this?

or you just don't read and reply repeating what you have read before somewhere else?

did you get 25MB filesize from above ISO1000?

no one else does, including my self,

so where are the data from the R-pixels in the RAW-file?

this information also can be found in the fuji s5 manual,
so i guess you could say i have read it somehere else,

cheers,
=)

cs_foto said:

Snoddas Date/Time 11:00:10 PM, Friday, December 07, 2007 (GMT)

"R-pixels are not used above ISO1000 in the S5, "

oh yeah you know so much man!

so how can you expain what i said? have you tried this?

or you just don't read and reply repeating what you have read before
somewhere else?

Click to expand...

at higher iso:s, that may be true

http://photobusinessforum.blogspot.com/2007/12/nikon-vs-canon-megapixel-issue.html

its same situation when comparing D200 and S5,
noise erases the resolution advantage for D200 above ISO800

OldMaster said:

Interestingly the latest report coming out of Luminous Landscape
concur. Once again the Nikon D3 showing itself capable with 12mp of
resolving more detail and being sharper than the 21mp Canon...no
surprises there I have to say!!

Click to expand...