Ad
Ad
Ad
Pages: « 1 2 [3] 4 5 »   Bottom of Page
Print
Author Topic: Sony A900 vs Hassy H3 vs Mamiya AFDiii  (Read 26460 times)
Steve Hendrix
Sr. Member
****
Offline Offline

Posts: 1122


WWW
« Reply #40 on: November 03, 2008, 08:40:31 AM »
ReplyReply

Quote from: rainer_v
steve:
i would like to hear something about my thoughts i wrote above about the default sharpening of p30 files in p1 and the strong bayer artefacts the cropfile shows.
thanks...

Rainer:

I tried downloading that raw file and I get a "page not found" error.

I was only able to download the original test shot under sunnier conditions and I do not see any artifacts in those files. If you know another way to get the raw file I will look at it, or perhaps since you have it you could send it to me.

sth@phaseone.com


Thanks,
Steve Hendrix
Phase One
Logged

Steve Hendrix
Sales Manager, www.captureintegration.com (e-mail Me)
MFDB: Phase One/Leaf-Mamiya/Hasselblad/Leica/Sinar
TechCam: Alpa/Cambo/Arca Swiss/Sinar
Direct: 404.543.8475
Steve Hendrix
Sr. Member
****
Offline Offline

Posts: 1122


WWW
« Reply #41 on: November 03, 2008, 11:48:25 AM »
ReplyReply

Quote from: EPd

Thanks EPD, I'll take a look.


Steve Hendrix
Phase One
Logged

Steve Hendrix
Sales Manager, www.captureintegration.com (e-mail Me)
MFDB: Phase One/Leaf-Mamiya/Hasselblad/Leica/Sinar
TechCam: Alpa/Cambo/Arca Swiss/Sinar
Direct: 404.543.8475
Steve Hendrix
Sr. Member
****
Offline Offline

Posts: 1122


WWW
« Reply #42 on: November 03, 2008, 12:02:56 PM »
ReplyReply

Quote from: rainer_v
steve:
i would like to hear something about my thoughts i wrote above about the default sharpening of p30 files in p1 and the strong bayer artefacts the cropfile shows.
thanks...

Rainer:

I do see some coloration in the palmetto fronds. Applying a combination of the color noise removal tool and the moire removal tool eliminated nearly all of that coloration in the fronds. I still see a little pattern weirdness, but it's possible that there was some wind producing movement of the fronds.


Steve Hendrix
Phase One



Logged

Steve Hendrix
Sales Manager, www.captureintegration.com (e-mail Me)
MFDB: Phase One/Leaf-Mamiya/Hasselblad/Leica/Sinar
TechCam: Alpa/Cambo/Arca Swiss/Sinar
Direct: 404.543.8475
rainer_v
Sr. Member
****
Offline Offline

Posts: 1131


WWW
« Reply #43 on: November 03, 2008, 12:24:25 PM »
ReplyReply

Quote from: Steve Hendrix/Phase One
Rainer:

I do see some coloration in the palmetto fronds. Applying a combination of the color noise removal tool and the moire removal tool eliminated nearly all of that coloration in the fronds. I still see a little pattern weirdness, but it's possible that there was some wind producing movement of the fronds.


Steve Hendrix
Phase One
i am asking at first about the default "zero" sharpening in p1 sw. i havent handled any file with zero sharpening which accepted so little usm than this "unsharpened"
p30 file converted in p1. do you use in your algorythm some default usm or sharpening which is applied to phase backs only? and which let them appear sharper if you compare these files with other backs or the same file with other softwares? i dont have much other explanation than this, because its not logical that an unsharpened file cant be treated with usm in post anymore without looking strongly oversharpened.

my second point are the nyquist artefacts ( not so the color defects which are easy to soften but there are many of little crosses, patterns and stairsteps in zones where the sensor resolution is surpassed). clearly all backs without aa filters may show similar patterns, but they look very strong in the p30/p1 file.
my theory is that one thing ( default zero sharpening which seems to be everything but zero with the p30 file) hangs together with the other thing ( artefacts ).
« Last Edit: November 03, 2008, 05:44:41 PM by rainer_v » Logged

rainer viertlböck
architecture photographer
munich / germany

www.tangential.de
Steve Hendrix
Sr. Member
****
Offline Offline

Posts: 1122


WWW
« Reply #44 on: November 03, 2008, 07:46:03 PM »
ReplyReply

Quote from: rainer_v
i am asking at first about the default "zero" sharpening in p1 sw. i havent handled any file with zero sharpening which accepted so little usm than this "unsharpened"
p30 file converted in p1. do you use in your algorythm some default usm or sharpening which is applied to phase backs only? and which let them appear sharper if you compare these files with other backs or the same file with other softwares? i dont have much other explanation than this, because its not logical that an unsharpened file cant be treated with usm in post anymore without looking strongly oversharpened.

my second point are the nyquist artefacts ( not so the color defects which are easy to soften but there are many of little crosses, patterns and stairsteps in zones where the sensor resolution is surpassed). clearly all backs without aa filters may show similar patterns, but they look very strong in the p30/p1 file.
my theory is that one thing ( default zero sharpening which seems to be everything but zero with the p30 file) hangs together with the other thing ( artefacts ).

Rainer:

The "no sharpening" setting in C14.5 or the sharpening amount slider pushed all the way to the left is the least sharp you will see. In my experience, Sinar and Hasselblad raw files are at default least sharp, Leaf and Phase One files at default are sharper. I should point out that they "start" sharper with the setting set to no sharpening. I'm not saying because of this our files will wind up being sharper, simply referring to the beginning state itself. This has been the case with Phase One files going back quite a while.

I do see the little weird artifacts you're talking about. I'm not sure what they are, whether they are bayer artifacts, movement-produced artifacts, or it is possible something is not quite right with that particular unit. It almost appears as if there is a calibration error with the sensor calibration file itself. It looks similar to what I've seen with other DB's when there have been corrupted calibration files. Hard to say...


Steve hendrix
Phase One
Logged

Steve Hendrix
Sales Manager, www.captureintegration.com (e-mail Me)
MFDB: Phase One/Leaf-Mamiya/Hasselblad/Leica/Sinar
TechCam: Alpa/Cambo/Arca Swiss/Sinar
Direct: 404.543.8475
thsinar
Sr. Member
****
Offline Offline

Posts: 2066


WWW
« Reply #45 on: November 04, 2008, 06:33:18 AM »
ReplyReply

Dear Steve,

Just to make things clear: Sinar does not apply ANY sharpening by default at any stage of the raws, but I think you know this. It is always a free user choice to sharpen or not, in the capture soft or later in CS or whatsoever.

Therefore, if the Sinar files look less sharp than Phase One files, or better said Phase One files look sharper than Sinar ones, that implies that there is some sharpening taking place somewhere with this Phase file.

Could you clarify this?

I have made the same experience, by sharpening PO files with a basic sharpening of 240/0.4, and it does simply look over-sharpened, as compared to Sinar files with the same settings.

Thanks for your time and best regards,
Thierry

Quote from: Steve Hendrix/Phase One
The "no sharpening" setting in C14.5 or the sharpening amount slider pushed all the way to the left is the least sharp you will see. In my experience, Sinar and Hasselblad raw files are at default least sharp, Leaf and Phase One files at default are sharper. I should point out that they "start" sharper with the setting set to no sharpening.
Steve hendrix
Phase One
Logged

Thierry Hagenauer
thasia_cn@yahoo.com
Steve Hendrix
Sr. Member
****
Offline Offline

Posts: 1122


WWW
« Reply #46 on: November 04, 2008, 07:38:38 AM »
ReplyReply

Quote from: thsinar
Dear Steve,

Just to make things clear: Sinar does not apply ANY sharpening by default at any stage of the raws, but I think you know this. It is always a free user choice to sharpen or not, in the capture soft or later in CS or whatsoever.

Therefore, if the Sinar files look less sharp than Phase One files, or better said Phase One files look sharper than Sinar ones, that implies that there is some sharpening taking place somewhere with this Phase file.

Could you clarify this?

I have made the same experience, by sharpening PO files with a basic sharpening of 240/0.4, and it does simply look over-sharpened, as compared to Sinar files with the same settings.

Thanks for your time and best regards,
Thierry


Um, I think I already did.

There is little to no sharpening applied in the base conversion algorithm with Sinar files, while Phase One does provide some amount of sharpening in the base conversion itself.

It might be added that at the actual raw capture stage, all digital capture devices have some sharpening in the mix (as well as a number of other elements). At least this has been my understanding as explained to me by several raw development programmers.


Steve Hendrix
Phase One


Logged

Steve Hendrix
Sales Manager, www.captureintegration.com (e-mail Me)
MFDB: Phase One/Leaf-Mamiya/Hasselblad/Leica/Sinar
TechCam: Alpa/Cambo/Arca Swiss/Sinar
Direct: 404.543.8475
thsinar
Sr. Member
****
Offline Offline

Posts: 2066


WWW
« Reply #47 on: November 04, 2008, 07:44:11 AM »
ReplyReply

Steve,

Thanks and sorry, I just wanted to make sure I understood it right.

I confirm again: there is NO sharpening in Sinar RAW files, not even a little.

Best regards,
Thierry

Quote from: Steve Hendrix/Phase One
Um, I think I already did.

There is little to no sharpening applied in the base conversion algorithm with Sinar files, while Phase One does provide some amount of sharpening in the base conversion itself.

It might be added that at the actual raw capture stage, all digital capture devices have some sharpening in the mix (as well as a number of other elements). At least this has been my understanding as explained to me by several raw development programmers.


Steve Hendrix
Phase One
Logged

Thierry Hagenauer
thasia_cn@yahoo.com
douglasboyd
Newbie
*
Offline Offline

Posts: 31


WWW
« Reply #48 on: November 06, 2008, 06:26:16 PM »
ReplyReply

All,

I have just done some more testing in my comparison of Sony a900 and Hasselblad H1-P30.  It is here: http://forums.dpreview.com/forums/read.asp...essage=29934382 .  I compared a cheap "Beercan" zoom against the Hasselblad HC 150mm Prime.  The Beercan needed some CA correction, but after correction there is very little if any difference in result.  The raw files can be downloaded at the link.

==Doug
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7487


WWW
« Reply #49 on: November 07, 2008, 12:27:41 AM »
ReplyReply

Hi!

I have checked out both images in LR 2.1. The observations I have made:

- I don't see the "halos" on the Hassy image that were visible on the original posting.
- I didn't see color fringing in the corner on the A 900 but a lot on the Hassy.
- The beercan seems to be sharper especially in the corners
- I see much more moiré on the Hassy than on the A900

I enclose a screendump from LR with "yellow/blue" fringing corrected on Hassy and no correction on the A900. No sharpening (amount = 0) on both.

[attachment=9504:Bild_9.jpg]
Quote from: douglasboyd
All,

I have just done some more testing in my comparison of Sony a900 and Hasselblad H1-P30.  It is here: http://forums.dpreview.com/forums/read.asp...essage=29934382 .  I compared a cheap "Beercan" zoom against the Hasselblad HC 150mm Prime.  The Beercan needed some CA correction, but after correction there is very little if any difference in result.  The raw files can be downloaded at the link.

==Doug
Logged

design_freak
Sr. Member
****
Offline Offline

Posts: 1074



« Reply #50 on: November 07, 2008, 01:32:59 AM »
ReplyReply

Quote from: douglasboyd
All,

I have just done some more testing in my comparison of Sony a900 and Hasselblad H1-P30.  It is here: http://forums.dpreview.com/forums/read.asp...essage=29934382 .  I compared a cheap "Beercan" zoom against the Hasselblad HC 150mm Prime.  The Beercan needed some CA correction, but after correction there is very little if any difference in result.  The raw files can be downloaded at the link.

==Doug

Please add to dpreview that it is H1 with P30. Because if sombody read "Hasselblad" think H3DII...
Thank you

Design Freak

-------------------------------------
Work hard and be nice to people
-------------------------------------

Logged

Best regards,
DF

-------------------------------------------
WORK HARD AND BE NICE TO PEOPLE
-------------------------------------------
Dustbak
Sr. Member
****
Offline Offline

Posts: 2370


« Reply #51 on: November 07, 2008, 02:31:50 AM »
ReplyReply

Furthermore. I consider the 150 the weakest lens of the HC lenses right after the 80 (the 80 is the weakest). At least of all the HC lenses I own these 2 have the least performance. The 150 has excellent AF speed though (for MF standards that is ).

Having said that I think the Sony has great performance but we already knew the 24MP DSLR's are great tools.
« Last Edit: November 07, 2008, 04:44:32 AM by Dustbak » Logged
rainer_v
Sr. Member
****
Offline Offline

Posts: 1131


WWW
« Reply #52 on: November 07, 2008, 07:15:26 AM »
ReplyReply

Quote from: Steve Hendrix/Phase One
Rainer:

I do see the little weird artifacts you're talking about. I'm not sure what they are, whether they are bayer artifacts, movement-produced artifacts, or it is possible something is not quite right with that particular unit. It almost appears as if there is a calibration error with the sensor calibration file itself. It looks similar to what I've seen with other DB's when there have been corrupted calibration files. Hard to say...


Steve hendrix
Phase One

well, i studied various bayer artefact patterns in the past  and for me they look like. ofcourse they jump into the eye more for the already applied
sharpening. no idea how a calibration file could affect this, and for sure its not blur or movement.
one word about "default" sharpening. i understand that this is a nice marketing thing cause your files look sharper
in direct comparision.  further there is no problem with this way of konverting as long you dont uprez the images by much.
but if you do it can be very difficult to start with an already sharpened file, more so because the applied sharpening in C1
at 0setting is not so little. i do this quite often ( uprezzing ) and starting with sharpened files i cant avoid to run into stairstepping
issues and other artefacts, which are nearly impossibler to hide at big sizes.  also blurring the (bayer?)artefacts becomes very visible
in this case, they would be much softer if starting with unsharpened files.
there is a reason why many people suggest to sharpen files at the end of the chain.
« Last Edit: November 07, 2008, 07:32:57 AM by rainer_v » Logged

rainer viertlböck
architecture photographer
munich / germany

www.tangential.de
BernardLanguillier
Sr. Member
****
Offline Offline

Posts: 8070



WWW
« Reply #53 on: November 07, 2008, 08:24:29 AM »
ReplyReply

Thanks for the link.

Interesting results indeed, the A900 can now be had at 2500 US$ in Tokyo...

If Nikon doens't announce something interesting before the end of November, I will probably become a Sony user soon... that would be as an addition to my Nikon set up.

Cheers,
Bernard
« Last Edit: November 07, 2008, 04:33:09 PM by BernardLanguillier » Logged

A few images online here!
douglasf13
Sr. Member
****
Offline Offline

Posts: 547


« Reply #54 on: November 08, 2008, 12:34:40 PM »
ReplyReply

Quote from: Dustbak
Furthermore. I consider the 150 the weakest lens of the HC lenses right after the 80 (the 80 is the weakest). At least of all the HC lenses I own these 2 have the least performance. The 150 has excellent AF speed though (for MF standards that is ).

Having said that I think the Sony has great performance but we already knew the 24MP DSLR's are great tools.

  True, but keep in mind that this test is done with a 20+ year old Minolta lens that goes for <$250 on ebay.  The Sony ZA 135mm 1.8 or 70-200mm 2.8 would probably blow that old Minolta "beercan" lens out of the water.
Logged
douglasboyd
Newbie
*
Offline Offline

Posts: 31


WWW
« Reply #55 on: November 08, 2008, 12:47:11 PM »
ReplyReply

Bernard,

I was in the same position as you since I owned a bunch of Nikon lenses.  But even if Nikon anounces at the end of November, it may still be months before you can actually buy it.  In the meantime, the a900 is already in the price reduction cycle, and I'm sure Sony is hard at work on the 48 mpixel sensor for the next model.   Right now I am testing my Nikon lenses on the a900 using a Nikon to A-mount adapter.  You loose AE and AF, but the resolution doesn't look bad so far.   In any case it is not expensive to load up on a collection of Sony A-mount lenses if you buy the older Minolta lenses.  These seem excellent so far, and are very affordable on eBay. I did buy two Zeiss lenses, but I'm having a hard time finding any advantage to them.  In fact, my CZ 24-70mm f2.8 lens is both overweight and overpriced compared to Minolta 24-105mm f3.5-4.5.  I can't see (or hear) any difference in focus performance.

==Doug




Quote from: BernardLanguillier
Thanks for the link.

Interesting results indeed, the A900 can now be had at 2500 US$ in Tokyo...

If Nikon doens't announce something interesting before the end of November, I will probably become a Sony user soon... that would be as an addition to my Nikon set up.

Cheers,
Bernard
Logged
tho_mas
Sr. Member
****
Offline Offline

Posts: 1696


« Reply #56 on: November 16, 2008, 01:41:14 PM »
ReplyReply

Quote from: rainer_v
there is no problem with this way of konverting as long you dont uprez the images by much.
but if you do it can be very difficult to start with an already sharpened file, more so because the applied sharpening in C1
at 0setting is not so little. i do this quite often ( uprezzing ) and starting with sharpened files i cant avoid to run into stairstepping
issues and other artefacts, which are nearly impossibler to hide at big sizes.
Hi Rainer,

very true!

But...
If there is some default sharpening in Capture One for Phase files (and there might be some) it's quite well balanced. At least for sensors without microlenses.
Attached a comparision of a P45 file crop in Capture One and ACR.
With ACR I just don't get the same details - even after sharpening (in Photoshop here with the values you proposed above and without noise reduction).
The C1 file is slightly sharper but at the same time has more details so looks more filigreed.
And on the other hand I can't see C1 pushes artifacts of my P45 files; sole thing is moiree but I repair only the affected parts on layers in Photoshop so this works quite well (though takes time).
I do a lot of upscaling, too (about 300% or so depending on the image). But files out of C1 are still "smooth" enough for upscaling without significant artifacts.
Starting with the same file out of ACR... I end up with squashy images and sharpening afterwards (even with sophisticated procedures) can't bring back the details C1 produces from the very beginning.
So... might be different with files from microlenses sensors but for those without I can't note problems with C1 even when upscaling.

Best Regards,
Thomas

[attachment=9734:c1_acr.jpg]
« Last Edit: November 16, 2008, 01:44:57 PM by tho_mas » Logged
BernardLanguillier
Sr. Member
****
Offline Offline

Posts: 8070



WWW
« Reply #57 on: November 16, 2008, 04:43:29 PM »
ReplyReply

Quote from: douglasboyd
Bernard,

I was in the same position as you since I owned a bunch of Nikon lenses.  But even if Nikon anounces at the end of November, it may still be months before you can actually buy it.

Which would be OK in fact. I am getting excellent results from my D3 and stitching has become a second nature, I could wait until mid February.

I am not sure that it would take months until availability though. Nikon has been much better with this recently and both the D700 and D90 were available for purchase within 3 weeks of announcement. Most Japanese employees get a significant bonus on Dec 25th, and my bet would be that Nikon would be targetting availability around mid December... that is of course if something significant is announced on Nov 20th.

Cheers,
Bernard
Logged

A few images online here!
krystian maj
Newbie
*
Offline Offline

Posts: 14


WWW
« Reply #58 on: November 18, 2008, 08:25:15 AM »
ReplyReply

besides all responds: I've been there durning this test. all the results you have to split by yourself. The mamiya back is an old demo, overheated (you can it clearly see on ISO400) and the raws were developed by photoshop, not a mamiya digital studio. hassy? forget it - all the time problems with software, monits "pull out the cable", "remove the lens", "remove the battery". damn, one of the worst cameras i've ever have. and one more thing: people who made and published this article, are sony alpha's fans. what can i add? if you have any other questions about this test: please feel free to ask.
Logged

ixpressraf
Guest
« Reply #59 on: November 18, 2008, 09:15:19 AM »
ReplyReply

That was clear from day one i saw this test ( the part that they were sony enthusiasts). i never got the part how it would be possible that a mediocre sensor builder got the best chip ever made, supposedly better than the canon cmos.... and even better than anything available on the MF marked. Until now I haven't seen anything from canon or nikon that delivers images of a quality that surpasses my old leaf Cantare or other 6Mp from last millenium. Yes they are faster, portable, higher iso etc but they never deliver a cleaner image like file as from even the oldest MF backs. It is like people are in a smoke of some good hashes or other overwhelming narcotic.
Logged
Pages: « 1 2 [3] 4 5 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad