Pages: [1] 2   Go Down

Author Topic: H2/P21 put to test.  (Read 11445 times)

samuel_js

  • Guest
H2/P21 put to test.
« on: December 14, 2007, 06:36:37 pm »

Hi all,
I've been testing my P21 under the last days. I thought maybe some of you could be interested in the results.
I'll probably be doing a big project that will last a few years (beside assignments), a research project the will be planned around music and photography put together. This could imply very big prints. There comes the question of my 18 mpx being capable of producing that king of prints. We are talking maybe 1 meter high.
So I put a simple scenario were I'll put my H2 and P21 to hard test. I want  to see sharpness and  DR in the first place, and I want to know how much the files could be upsized without loosing (too much) quality.
Can I get comparable quality from a resized P21 file beside a P30 or P45 file?
The motive of this thread is to find out how resizable medium format files are. I'm not very familiar with resizing but I know a 18 mpx DB can stretch a lot. But of course, there must be a limit. I'd like to get, if possible, real information here, not too much charts or thesis if you know what I mean. Anyway, any medium format user will recognize the sharpness and detail rendition of these captures. For 35mm new users here I suppose this is a kind of eye-training about what a DB can do.

So I'll start posting the full picture version with 100% crops of the file. And then, after your advice,  I'll resize the picture to see where I can go.

Details:
- Camera Hasselblad h2 and PhaseOne P21
- Lens: Hasselblad 210mm HC
- No flash used.
- ISO 100 1/2s at f7.1 on tripod and mirror LU
- Processed 16 bit tiff file in C1 with everything to 0 (sharpening or noise reduction)
- No adjustments of any kind, the file is totally flat.
- Then imported in PS and made a print. Added sharpening amount 35 ratio 1 ( just what I thought the print needed).
So there's minimal processing, almost flat.


My initial observation is that the 210mm is scary sharp actually. Very good and natural color rendition. The picture got a little warm after the web conversion but color isn't my main concern at this point. Also the combination is capable of resolving a lot of detail. Some of the crops are amazing, like the texture in the card or the small clock's knob.
The the jpg versions also loose quality againt the 16 bit tiffs but...

Full image cropped 1x1 ratio:
 

/Samuel
« Last Edit: December 14, 2007, 06:56:05 pm by samuel_js »
Logged

AndreNapier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 422
    • Andre Napier Photography
H2/P21 put to test.
« Reply #1 on: December 14, 2007, 08:38:08 pm »

I just finished printing some exhibition prints 44x60 from A75. They are printed at 2880DPI on Museo Silver Rag. Files were up to the size at 300 resolution. The prints are undisputable photo quality from any distance.
I am not a guru on resizing but the following approach was learned from a print master.

1. Never resize in one step if you are going up more than 30%.
2. Make small increasements of 10% at a time.
3. Adjust the sharpness in small steps as well.

As it was explain to me it helps the software to make a better guess on a new pixel since 10 original pixels generate one new pixel. If you would up size 100% at one step each pixel would just clone itself. It works for me.
Http://AndreNapier.com
Logged

Jonathan Wienke

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5829
    • http://visual-vacations.com/
H2/P21 put to test.
« Reply #2 on: December 14, 2007, 08:50:09 pm »

Quote
I am not a guru on resizing but the following approach was learned from a print master.

1. Never resize in one step if you are going up more than 30%.
2. Make small increasements of 10% at a time.
3. Adjust the sharpness in small steps as well.

Resizing in multiple steps hasn't been necessary since Photoshop added 16-bit processing and Bicubic Smoother to the feature list several versions ago. Sizing in one step or several in 16-bit mode makes little if any difference.
Logged

jing q

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 596
    • we are super
H2/P21 put to test.
« Reply #3 on: December 14, 2007, 10:28:49 pm »

Quote
Resizing in multiple steps hasn't been necessary since Photoshop added 16-bit processing and Bicubic Smoother to the feature list several versions ago. Sizing in one step or several in 16-bit mode makes little if any difference.
[a href=\"index.php?act=findpost&pid=160768\"][{POST_SNAPBACK}][/a]

not true.
I can tell you that resizing in small increments helps the quality.
DON'T use bicubic smoother.
I compared various methods of upsizing and upsizing with bicubic smoother caused a loss of detail compared to upsizing by bicubic.

my resizing workflow went as such
10 percent increments, bicubic
one pass of sharpening for the edges (low radius high percentage)
one pass of local contrast sharpening (high radius low percentage)
and one more pass of verrrry fine sharpening of the edges (very low radius high percentage)
Logged

Dustbak

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2442
    • Pepperanddust
H2/P21 put to test.
« Reply #4 on: December 15, 2007, 03:25:46 am »

I have made prints from my A17 (comparable to your P21) that were 1mx1m (I also lost on cropping!).

I have tried:

OnOne Genuine fractals
PS bicubic
Alien skin blow-up

I am still wondering what really gave the best results. I am leaning towards either blow-up or genuine fractals where I would give the edge to blow-up. PS was not bad either.

I also thought it was no longer necessary to upsize in increments since PS6.0. Maybe there is a difference between using bicubic and bicubic smoother. I have never really looked into that.

Another thing you could do with static scenes like this is stitching.
Logged

rueyloon

  • Full Member
  • ***
  • Offline Offline
  • Posts: 221
    • http://www.36frames.com
H2/P21 put to test.
« Reply #5 on: December 15, 2007, 03:35:10 am »

in my own test I found photozoom to be superior.

check it out.

cheers
rgs
rueyloon
Logged

samuel_js

  • Guest
H2/P21 put to test.
« Reply #6 on: December 15, 2007, 09:00:21 pm »

Quote
in my own test I found photozoom to be superior.

check it out.

cheers
rgs
rueyloon
[a href=\"index.php?act=findpost&pid=160814\"][{POST_SNAPBACK}][/a]
I've been trying photozoom today. Amazing! I'll post a few samples tomorrow.  
I'll try Genuine fractals as well.
« Last Edit: December 15, 2007, 09:06:26 pm by samuel_js »
Logged

AndrewDyer

  • Full Member
  • ***
  • Offline Offline
  • Posts: 125
    • http://www.andrewdyer.com
H2/P21 put to test.
« Reply #7 on: December 16, 2007, 07:51:44 am »

Quote
I've been trying photozoom today. Amazing! I'll post a few samples tomorrow. 
I'll try Genuine fractals as well.
[a href=\"index.php?act=findpost&pid=160946\"][{POST_SNAPBACK}][/a]
I had a look at Photozoom just now on their website, and although their examples of blowing up such a small file 400% is just idiocy, and enlarging a better quality file would probably look better, I find the way it enlarges the images makes them look like they are digital paintings. Genuine Fractals gave me much the same awful look.
I would rather a softer look to the image that you get from Photoshop's enlargement than what you get from these 2 programs. I would prefer to enlarge in Photoshop and try some various sharpening methods (High Pass filter method, USM layers blended in etc.)

Samuel. It is hard to believe those images you posted were not sharpened... actually, the softening you might get from enlarging may help the hard-jagged edged highlights I see there.
Personally I don't like ever enlarging more than 200%, although you could get away with more if you knew that no-one would ever be able to view it up close.
Logged
Andrew
 ht

samuel_js

  • Guest
H2/P21 put to test.
« Reply #8 on: December 16, 2007, 07:58:51 am »

Quote
Samuel. It is hard to believe those images you posted were not sharpened... actually, the softening you might get from enlarging may help the hard-jagged edged highlights I see there.
Personally I don't like ever enlarging more than 200%, although you could get away with more if you knew that no-one would ever be able to view it up close.
[a href=\"index.php?act=findpost&pid=161002\"][{POST_SNAPBACK}][/a]

Andrew, as I wrote in my original post, they are sharpened. Amount 35 ratio 1. I did sharpen a little when printing from PS but not when exported from C1.

I will try to compare these programs with photoshop. I suppose the best method is exporting the RAW to a tiff without any sharpening at all? To see the whole effect? And then sharpen both individually?
Logged

hubell

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1135
H2/P21 put to test.
« Reply #9 on: December 16, 2007, 10:19:13 am »

There is a comparative review of  PS Bicubic, Genuine Fractals, Blowup and Photozoom Pro in the Nov/Dec issue of Photo Techniques Magazine. Bottom line: PS Bicubic and Blowup were pretty comparable(and quite good), and Genuine Fractals and Photozoom Pro were  comparable  to each other and a bit better than the other two. The testing was done with images from a Leica M8. I wonder if the differences between these programs would be even less pronounced with files from higher resolution MFDBs where the amount of enlargement to reach say a 30"x40" print size is much lower.

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
H2/P21 put to test.
« Reply #10 on: December 16, 2007, 04:21:13 pm »

I did a review of Genuine Fractals some years ago. Some of the readers of my reviews at the time said they liked GF for product shots, and indeed on testing I could see that GF was spectacular on clean curves like the type on product boxes;  if I recall rightly, Michael Reichmann, the owner of this site, looked at GF and found it to be nothing special for his purposes. We corresponded, and as I recall Michael tersely but cogently indicated that he was  sticking by his opinion because he didn't do box shots, which I considered to be a perfectly straight and reasonable answer, given that he is a landscape and art photographer. I don't know whether our respective opinions need updating now there are new versions and new software out there, but the lesson seems to be that the software you choose for blowups is determined by your subject matter.

Edmund

Quote
There is a comparative review of  PS Bicubic, Genuine Fractals, Blowup and Photozoom Pro in the Nov/Dec issue of Photo Techniques Magazine. Bottom line: PS Bicubic and Blowup were pretty comparable(and quite good), and Genuine Fractals and Photozoom Pro were  comparable  to each other and a bit better than the other two. The testing was done with images from a Leica M8. I wonder if the differences between these programs would be even less pronounced with files from higher resolution MFDBs where the amount of enlargement to reach say a 30"x40" print size is much lower.
[a href=\"index.php?act=findpost&pid=161012\"][{POST_SNAPBACK}][/a]
« Last Edit: December 16, 2007, 04:24:59 pm by eronald »
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

RobertJ

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 706
H2/P21 put to test.
« Reply #11 on: December 16, 2007, 04:26:31 pm »

Quote
I find the way it enlarges the images makes them look like they are digital paintings.

In PhotoZoom Pro, you should tone down the "Sensitivity" slider, and it won't look painterly anymore.  That slider increases the contrast of edges, or however you want to describe it.  You have to create your own settings, not use the defaults.

I find PZ Pro to be the best out there.
Logged

vjbelle

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 635
H2/P21 put to test.
« Reply #12 on: December 18, 2007, 06:32:53 pm »

FWIW..... I have used all of the above upsizing software and have found all of them unacceptable.  They have a plastic look that generally distorts the image.  I suggest that you try to use C1 4 or RD and upsample at that stage to whatever size you will need for your print (the bigger the better).  Then, compare that to an upsized version with those plugins that claim to be better than the holy grail.  You will see a VERY big difference.  If you prefer Fractals or those others then that is your choice.  As for me I will only upsize at the Raw converter stage or use PS for upsampling.  BTW, ACR does not function as well as C1 or RD and also exhibits that plastic look that is evident with the plugins.  

Victor
Logged

samuel_js

  • Guest
H2/P21 put to test.
« Reply #13 on: December 18, 2007, 07:27:51 pm »

Quote
FWIW..... I have used all of the above upsizing software and have found all of them unacceptable.  They have a plastic look that generally distorts the image.  I suggest that you try to use C1 4 or RD and upsample at that stage to whatever size you will need for your print (the bigger the better).  Then, compare that to an upsized version with those plugins that claim to be better than the holy grail.  You will see a VERY big difference.  If you prefer Fractals or those others then that is your choice.  As for me I will only upsize at the Raw converter stage or use PS for upsampling.  BTW, ACR does not function as well as C1 or RD and also exhibits that plastic look that is evident with the plugins. 

Victor
[a href=\"index.php?act=findpost&pid=161565\"][{POST_SNAPBACK}][/a]
I've never tried C1 for upsizing. I'll give it a try too.
Thank's
Logged

samuel_js

  • Guest
H2/P21 put to test.
« Reply #14 on: December 18, 2007, 08:23:12 pm »

Quote
FWIW..... I have used all of the above upsizing software and have found all of them unacceptable.  They have a plastic look that generally distorts the image.  I suggest that you try to use C1 4 or RD and upsample at that stage to whatever size you will need for your print (the bigger the better).  Then, compare that to an upsized version with those plugins that claim to be better than the holy grail.  You will see a VERY big difference.  If you prefer Fractals or those others then that is your choice.  As for me I will only upsize at the Raw converter stage or use PS for upsampling.  BTW, ACR does not function as well as C1 or RD and also exhibits that plastic look that is evident with the plugins. 

Victor
[a href=\"index.php?act=findpost&pid=161565\"][{POST_SNAPBACK}][/a]
This was unexpected, C1 wins easily against Photoshop.  
I resized both files up to 39 mpx. This time all settings were 0, inclusive sharpening, the files are completely flat. Above PS, under C1. In my opinion, all lines and details are better in C1. It holds color and shadows better too...











Logged

hubell

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1135
H2/P21 put to test.
« Reply #15 on: December 18, 2007, 10:15:42 pm »

Quote
This was unexpected, C1 wins easily against Photoshop.   
I resized both files up to 39 mpx. This time all settings were 0, inclusive sharpening, the files are completely flat. Above PS, under C1. In my opinion, all lines and details are better in C1. It holds color and shadows better too...
Interesting how different perceptions of relative  quality can be. I find the PS upsizing much more natural and attractive. The C1 examples look brittle and overly sharpened for my taste.

jing q

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 596
    • we are super
H2/P21 put to test.
« Reply #16 on: December 18, 2007, 11:50:22 pm »

once you go past a certain size you need to do test strips to see the effect of your enlarging.
I've done big prints where I've had to sharpen to the point of looking overly so, but when printing as a strip on an epson 4800 it looks natural.
Logged

samuel_js

  • Guest
H2/P21 put to test.
« Reply #17 on: December 19, 2007, 03:07:15 am »

Quote
Interesting how different perceptions of relative  quality can be. I find the PS upsizing much more natural and attractive. The C1 examples look brittle and overly sharpened for my taste.
[a href=\"index.php?act=findpost&pid=161623\"][{POST_SNAPBACK}][/a]
If you look closer almost every edge looks pixelated in the PS version. It's more visible in the card samples.
Logged

Dustbak

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2442
    • Pepperanddust
H2/P21 put to test.
« Reply #18 on: December 19, 2007, 03:18:38 am »

From these samples I find it very hard to determine which I prefer. Looking at teh cards, the PS version appears to be slightly more contrasty and saturated, the C1 version slightly more sharpened. The shadows of the C1 samples also appear to have more detail which is probably also why I am experiencing the PS files to have more contrast.

I agree you have to see test prints on the desired material to be really sure what provides the best solution here.
« Last Edit: December 19, 2007, 03:20:21 am by Dustbak »
Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
H2/P21 put to test.
« Reply #19 on: December 19, 2007, 04:16:52 am »

Indeed, C1 appears to be a surprise contender. My experience is that special software only shows a real difference with PS around 4x - 10x, maybe that's where you should test.

Edmund

Quote
From these samples I find it very hard to determine which I prefer. Looking at teh cards, the PS version appears to be slightly more contrasty and saturated, the C1 version slightly more sharpened. The shadows of the C1 samples also appear to have more detail which is probably also why I am experiencing the PS files to have more contrast.

I agree you have to see test prints on the desired material to be really sure what provides the best solution here.
[a href=\"index.php?act=findpost&pid=161682\"][{POST_SNAPBACK}][/a]
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald
Pages: [1] 2   Go Up