What happened to all D100 softness posts??

Gabriele Sartori said:

That is not the truth, Nikon is softer since they don't have a very good
jpeg scheme otherwise it would be impossible to explain why jpeg
out of PS7 is sharper even with the same file size.

Click to expand...

I don't know exactly what you mean by a "JPEG scheme", but the
JPEG encoding engine (if this is what you are talking about) can't
be responsible for the softening of the images when you compare them
to RAW. This encoding is well standardised, and there is only one
correct implementation. Varying the amount of compression
introduces artifacts that look totally different.
The blurring must be due to some step in the chain prior to JPEG
encoding. Several things happen before, e.g. the Bayer interpolation.

I agree with you that, since Nikon offers three levels of sharpening,
at least the highest sharpening should be similar to what you get from
e.g. a Fuji S2. So, in my opinion, the real problem of the D100 is not
that the images look soft with the default settings, but that the
sharpening filter that is used in the higher sharpening settings is lousy
and is inferior to what is found in other camera's.

If Nikon should improve something on the sharpening in a firmware
upgrade, it should be a better sharpening filter in the high sharpening
mode. That would at least make the people happy that refuse to do
post-processing.

Personally, I prefer to save JPEGS without sharpening, and use an
advanced filter afterwards. I believe this allows one to rival RAW,
at least in terms of sharpness.

Vtie

PS Did anybody compare the high sharpening mode between
firmware 1.01 and 2.00 ?

Pete Perry said:

You're right about one thing, with Canon you don't have to do that
so you should probably go and buy a D60 so you won't have to
complain anymore. Then when you realize the back focus issues
cause you to completely miss the pictures you were trying to get

Click to expand...

Member said:

all. Oh and another thing, if you go to the Canon forum, there are
complaints of softness and underexposure there also.

Click to expand...

I know that and it is one of the reasons (there are others as well) why I keep the D100, but I keep also my right to complain to a company that in the year 2002 doesn't know how to make a decent jpeg engine.

--
Regards
Gabriele Sartori

Vtie said:

Gabriele Sartori said:

That is not the truth, Nikon is softer since they don't have a very good
jpeg scheme otherwise it would be impossible to explain why jpeg
out of PS7 is sharper even with the same file size.

Click to expand...

I don't know exactly what you mean by a "JPEG scheme", but the
JPEG encoding engine (if this is what you are talking about) can't
be responsible for the softening of the images when you compare them
to RAW. This encoding is well standardised, and there is only one
correct implementation.

Click to expand...

With all the due respect but you are clearly telling to the world that you don't know nothing about jpeg encoding engines. I'm involved with my profession in this type of thngs and it is not as you say. You can have two perfectly functional jpeg encoder one is soft, one is not, one show blocks, one doesn't. When you have a good jpeg encoder you should not see any difference from a raw if the compression is 1/5 or so.

Vtie said:

I agree with you that, since Nikon offers three levels of sharpening,
at least the highest sharpening should be similar to what you get from
e.g. a Fuji S2. So, in my opinion, the real problem of the D100 is not
that the images look soft with the default settings, but that the
sharpening filter that is used in the higher sharpening settings is
lousy
and is inferior to what is found in other camera's.

Click to expand...

OK, this is your opinion and may be it is the truth. THe result don't change though.

Vtie said:

PS Did anybody compare the high sharpening mode between
firmware 1.01 and 2.00 ?

Click to expand...

It seems that nothing has been changed.

--
Regards
Gabriele Sartori

You sound as if you designed D100, not just like you own it. I know it can hurt spending $3000 on a camera only to find out that it is so inconsistant that it can best be described and used as a toy.

I don't know much (or at least I think I don't know much, but then again I have masters in pure math and think I don't know much about math either) photography, but if what I can say for sure is that Nikon's software is 10 years behind the rest of the world. It looks like their cameras might be too.

Or could it be that they don't know how to write good software so their in camera embedded software sucks at getting the most out of those pixels coming from the sensor.

You can all say whatever you want, but PowerShot G2 will in 99% of circumstances give you better or just as good photo under any light conditions! If you are after quality and don't have "small *****" complex, you might as well save lots of money and buy yourself a camera that works!! (or at least its embedded image processing algorithms work).

If you need to have the biggest lens around,.. well then continue using your D100 and "look like a pro", lol...

It's pathetic and humorous to see how many complex and elaborate reasons people have come up with to explain to themselves ( and then try to persuade the others too), of why D100 sucks. I don't know if it's the amount of money they blew or if its purely pride. I don't care. I'm not afraid to call a spade a spade. I do have extremely high standards that I apply equally to myself and others, and according to those, D100 is a TOY.

Ken,

I understand that u can get sharp image out of NEF, but NEF transfer to JPEG or TIFF also have higher noise than straight JPEG or remove sharpening in NC3 and the save as JPEG or TIFF. What is your workflow to reduce noise?

jack

Ken Mai said:

Ken Mai said:

1. Chromatic abberations (CA) and

Click to expand...

With what lenses are you seeing the CA? This may be a function of
the lens more than the CCD. Athough, the CCDs can introduce some CA
problems as well (see Rob Galbraith's 1Ds article).

Ken Mai said:

2. The soft focus issue (thus the post here)

Click to expand...

Uh, I don't think it's a soft focus issue. I haven't seen anyone
complain about the autofocus module (those complaints are in the
Fuji S2 forum ). The problem is that the in-camera JPEG engine
generates somewhat soft images when set to normal sharpening and
too sharp images when set to high sharpening. So the solution is to
shoot in NEF and deal with the sharpening in post processing. I'm
coming around to the conclusion that shooting in NEF all the time
is basically the right way to

Click to expand...

Hi

You gave me a good smile and you are right in some points.

Mario Grgic said:

You can all say whatever you want, but PowerShot G2 will in 99% of
circumstances give you better or just as good photo under any light
conditions!

Click to expand...

This is why I didn't sell my G2 (-;

Mario Grgic said:

If you need to have the biggest lens around,.. well then continue
using your D100 and "look like a pro", lol...

Click to expand...

In my case I was aware that I trade things in against each other.

I myself experienced how sophisticated G2's internal processing is in high contrast situations or how "pleasing" colours are.

But I disliked the G2 "speed" and AF performance. This was my personal reason to look for a D-SLR.

The D100 reacts much quicker than my G2 an this is important for certain kind of actions shots. I still find myself in situations were I feel a very little lag but D100 is worlds quicker than G2

Beond this I agree that I am still in a situation that out of the D100 camera images do not fully satisfy me. Neutral colours (and I think D100 colours are relatively neutral) are not pleasing in every circumstance. Exposure is more difficult and the system in itself has serious flaws.

Mario Grgic said:

It's pathetic and humorous to see how many complex and elaborate
reasons people have come up with to explain to themselves ( and
then try to persuade the others too), of why D100 sucks.

Click to expand...

Well, let's share the pain and it get's half (-;

I don't mind people like you who criticize without owning a D100. This camera is no real dream, in some aspects a nightmare in others ahead of P&S cameras. It is for sure no replacement for film based cameras, at least not for MF.

Mario Grgic said:

I don't
know if it's the amount of money they blew or if its purely pride.

Click to expand...

I hate the money I spent for it, is like monetarian masochism. Tell me a cheaper camera that is as quick and as comfortable in handling, convince me quality wise and possibly I trade the D100 in...

Mario Grgic said:

I don't care. I'm not afraid to call a spade a spade. I do have
extremely high standards that I apply equally to myself and others,
and according to those, D100 is a TOY.

Click to expand...

I pushed this "toy" expression myself because I think in a way like you do. The D100 does not satisfy my imaginations of accuracy I like to work with too... It seems to be that D1/D1x/D1h are not any better here...

But I see no alternative when you want speed and a better AF and you still own some N* lenses.

Regards, A. Schiele

Unfortunately I don't have PowerShot G2 anymore, I traded it for D100. I wonder sometimes if I made a BIG mistake.

Gabriele Sartori said:

With all the due respect but you are clearly telling to the world
that you don't know nothing about jpeg encoding engines. I'm
involved with my profession in this type of thngs and it is not as
you say.

Click to expand...

Have you ever implemented the JPEG compression/decompression
algorithm yourself? Have you ever studied the theory of
Fourier transforms, FFT and cosine transforms? I have, and believe me, I
know what I am talking about. I have more than 12 years of
experience in high end digital image analysis software.
I simply makes me laugh if I hear someone telling me that I don't
know anything about JPEG encoding.

A correct implementation of a JPEG compression will never introduce
the softness observed with the D100 images. It is simply not possible
to have a overall smooth blur as a JPEG artifact. Everything that degrades
from a perfect image will immediately show blocks, due to the very
nature of the JPEG algorithm itself. It works on 8x8 blocks.
Having a smooth blur that runs over these blocks would be real magic,
because the encoding of one 8x8 block is completely independent
of the neighbors. Take any book that describes the JPEG compression
algorithm if you want to know further details.

Vtie

Steve,

When I got the E-10 there was a learning curve. I learned to shoot at ISO 80, sharpening set to none (to keep the noise down), low contrast and jpeg compression set to 1/2.7. I would then post process every image to sharpen it up. Raw was not an option because of the slow processing speed.

I am now a happy D100 owner. There is a learning curve. I shoot either JPEG Fine or Raw. I set sharpening to none with jpeg and normal in RAW mode. I use low contrast, Adobe RGB colorspace. The D100 is a joy to use. Write times are fast compared to the E-XX (even raw mode is fast). Review times are super-fast (which was always an issue for me with the E-10). Jpeg or Raw images should be sharpened in post processing to get the best results.

As for D100 "underexposure" I do see it sometimes. I think the D100 is trying to be conservative and not blow the highlights.If you shot Raw it is easily corrected.

So... If you don't mind post processing and you are willing to shoot raw mode (9.5mb files = 51 images on my 512mb card) then you will be happy with the D100 after a little practice

Check out my gallery to see some of my full size soft JPEGS : )
http://www.pbase.com/seanu

the test of Phil indicated that the images in RAW were sharper but had more noise. This could let think of a in camera sharpening with threshold value to 0, which unfortunately seem to practise much of people, what increases the noise. A test of an image RAW with "normal" sharpening, then the same image sharpening "off" + fuzzy mask in NC3 (to find same sharpness) but with a threshold valus to 3 , confirmed my intuition.
images test :
onthe left: RAW "normal" sharpening

on the right:the same but sharpenind deactivate in NC3 and usm withi threshold: 3

Jack310 said:

I understand that u can get sharp image out of NEF, but NEF
transfer to JPEG or TIFF also have higher noise than straight JPEG
or remove sharpening in NC3 and the save as JPEG or TIFF. What is
your workflow to reduce noise?

jack

Ken Mai said:

Jack310 said:

1. Chromatic abberations (CA) and

Click to expand...

With what lenses are you seeing the CA? This may be a function of
the lens more than the CCD. Athough, the CCDs can introduce some CA
problems as well (see Rob Galbraith's 1Ds article).

Jack310 said:

2. The soft focus issue (thus the post here)

Click to expand...

Uh, I don't think it's a soft focus issue. I haven't seen anyone
complain about the autofocus module (those complaints are in the
Fuji S2 forum ). The problem is that the in-camera JPEG engine
generates somewhat soft images when set to normal sharpening and
too sharp images when set to high sharpening. So the solution is to
shoot in NEF and deal with the sharpening in post processing. I'm
coming around to the conclusion that shooting in NEF all the time
is basically the right way to

Click to expand...

Click to expand...

I have missed than it is an 200% image, and Colour Noise reducer on, in NC3 (medium)

rome said:

the test of Phil indicated that the images in RAW were sharper but
had more noise. This could let think of a in camera sharpening with
threshold value to 0, which unfortunately seem to practise much of
people, what increases the noise. A test of an image RAW with
"normal" sharpening, then the same image sharpening "off" + fuzzy
mask in NC3 (to find same sharpness) but with a threshold valus to
3 , confirmed my intuition.
images test :
onthe left: RAW "normal" sharpening
on the right:the same but sharpenind deactivate in NC3 and usm
withi threshold: 3

Jack310 said:

I understand that u can get sharp image out of NEF, but NEF
transfer to JPEG or TIFF also have higher noise than straight JPEG
or remove sharpening in NC3 and the save as JPEG or TIFF. What is
your workflow to reduce noise?

jack

Ken Mai said:

rome said:

1. Chromatic abberations (CA) and

Click to expand...

With what lenses are you seeing the CA? This may be a function of
the lens more than the CCD. Athough, the CCDs can introduce some CA
problems as well (see Rob Galbraith's 1Ds article).

rome said:

2. The soft focus issue (thus the post here)

Click to expand...

Uh, I don't think it's a soft focus issue. I haven't seen anyone
complain about the autofocus module (those complaints are in the
Fuji S2 forum ). The problem is that the in-camera JPEG engine
generates somewhat soft images when set to normal sharpening and
too sharp images when set to high sharpening. So the solution is to
shoot in NEF and deal with the sharpening in post processing. I'm
coming around to the conclusion that shooting in NEF all the time
is basically the right way to

Click to expand...

Click to expand...

Click to expand...

Vtie said:

Vtie said:

With all the due respect but you are clearly telling to the world
that you don't know nothing about jpeg encoding engines. I'm
involved with my profession in this type of thngs and it is not as
you say.

Click to expand...

Have you ever implemented the JPEG compression/decompression
algorithm yourself? Have you ever studied the theory of
Fourier transforms, FFT and cosine transforms?

Click to expand...

Yes, Yes and yes

Vtie said:

I simply makes me laugh if I hear someone telling me that I don't
know anything about JPEG encoding.

Click to expand...

Is what you proved saying that all the jpeg compressors are the same, this is clearly not true. Don't mislead people than. If you were trying to say something else, rephrase what you said.

Vtie said:

A correct implementation of a JPEG compression will never introduce
the softness observed with the D100 images. It is simply not possible
to have a overall smooth blur as a JPEG artifact. Everything that
degrades
from a perfect image will immediately show blocks, due to the very
nature of the JPEG algorithm itself. It works on 8x8 blocks.
Having a smooth blur that runs over these blocks would be real magic,
because the encoding of one 8x8 block is completely independent
of the neighbors. Take any book that describes the JPEG compression
algorithm if you want to know further details.

Click to expand...

You perfectly know that is up to the encoder designer to put some amount of integration or low pass filtering in order to minimize blocking. There you can lose resolution. Now, to save the day you may hide yourself behind technicalities and say that this is not part of the jpeg engine but you know what I meant in the first place. You also perfectly know that you can make a DCT engine in many possible ways, some are better than others. Also read entirely my post please, I don't exclude that it may be just a big low pass at the beginning of everything.

--
Regards
Gabriele Sartori

Gabriele Sartori said:

You perfectly know that is up to the encoder designer to put some
amount of integration or low pass filtering in order to minimize
blocking. There you can lose resolution. Now, to save the day you
may hide yourself behind technicalities and say that this is not
part of the jpeg engine but you know what I meant in the first
place. You also perfectly know that you can make a DCT engine in
many possible ways, some are better than others. Also read entirely
my post please, I don't exclude that it may be just a big low pass
at the beginning of everything.

Click to expand...

Yes, a low pass filter prior to JPEG encoding may explain the D100
softness of JPEG images. Personally, I would never call this filtering
a part of the JPEG encoding engine. I would call it, well, a low
pass filter prior to JPEG encoding. If you call this a part of the
encoding we are getting into terminology and I am not going to argue
over that. Then I guess we totally agree, apart from what you call
JPEG encoding.

However, it sounds like a very exotic design to me if they would apply
a sharpening filter to the data (like most of the modes are doing, at
least according to their names), and then again a low pass filter prior to
encoding! Do you think Nikon would do that kind of strange things?
It's a waste of quality anyway!

Vtie

explaining the obvious and easily researchable.
those who really know how jpeg encoding works are behind you.

Regards,

--

Vtie said:

Vtie said:

You perfectly know that is up to the encoder designer to put some
amount of integration or low pass filtering in order to minimize
blocking. There you can lose resolution. Now, to save the day you
may hide yourself behind technicalities and say that this is not
part of the jpeg engine but you know what I meant in the first
place. You also perfectly know that you can make a DCT engine in
many possible ways, some are better than others. Also read entirely
my post please, I don't exclude that it may be just a big low pass
at the beginning of everything.

Click to expand...

Yes, a low pass filter prior to JPEG encoding may explain the D100
softness of JPEG images. Personally, I would never call this filtering
a part of the JPEG encoding engine. I would call it, well, a low
pass filter prior to JPEG encoding. If you call this a part of the
encoding we are getting into terminology and I am not going to argue
over that. Then I guess we totally agree, apart from what you call
JPEG encoding.

However, it sounds like a very exotic design to me if they would apply
a sharpening filter to the data (like most of the modes are doing, at
least according to their names), and then again a low pass filter
prior to
encoding! Do you think Nikon would do that kind of strange things?
It's a waste of quality anyway!

Vtie

Click to expand...

--

Pixelman said:

Why didn't Nikon ask you to write their D100 Guide like Fuji did?

Click to expand...

Fuji didn't ask me to write an S2 Guide. I wrote one, Fuji has licensed it for a specific purpose (registration premium).

As for Nikon, sometimes I wonder if they know I exist.

--
Thom Hogan
author, Nikon Field Guide
author, Nikon Flash Guide
author, Complete Guide to the Nikon D100
author, Complete Guide to the Nikon D1, D1h, & D1x
http://www.bythom.com

Casey,

With posts like this, maybe you should consider posting here even less often. Honestly, this has been talked about to DEATH. If you really want to reminisce, then use the search tool or go back a couple months. It really is a shame to see so many people wasting their time saying the same things over and over again.

Just a thought,

--
Dan

http://www.med.umich.edu/pibs
http://www.pbase.com/efatapo
http://www.hillsdale.edu

i don't know as i have a fuji S1 but have heard that this problem was fixed by the last firmware update. it was an inofficial fix, so not mentioned by nikon.

read this some weeks ago.
thomas

Casey Cheung said:

A few months ago, the Nikon SLR forum was flooded with posts about
the D100 producing "soft" images. It seemed that every other post
was the same thing over and over again.

I was actually turned off by these posts, and sort of took a break
from this forum for a little while. There have been so many new
members with all kinds of esoteric code names all saying the same
things for awhile. I realize that the advent of the $2,000 digital
SLR has brought in many new members, which is a good thing, I just
didn't expect so many to keep asking the same ol' same ol'.

However, as I recently look at the latest postings and headlines, I
have noticed that the D100 softness issue has died down a bit.
Things now appear to be "normal" again, in terms of having a good
amount of diverse topics to discuss about. I think this is a
positive change.

So whatever happened to all the softness posts? Did the D100 users
simply get smarter and learned to deal with it? Did they convert to
(gasp!) Canon users instead? Did they give up and went back to
their Coolpix P/S cams?

Okay, I realize I have nothing of significance to contribute
regarding this post, I was just making a casual observation. Don't
take anything I wrote too seriously--no harm meant.

Regards,
Casey

D1x & D100 owner

Click to expand...

Casey Cheung said:

Okay, I realize I have nothing of significance to contribute

Click to expand...

That's the part I understood..

,Jason

--
Gallery - http://www.digitalphotos.pwp.blueyonder.co.uk

You're not wasting your time. You have an audience larger than any one reader, and I find your explanations highly informative.

A coherent explanation of JPEG mode "softness" is still lacking, although the conventional wisdom of late, I think, you're debunking. Interesting.

Hogan ruled out hardware low-pass filtering, which he originally judged unusually "aggressive", in favor of blaming the JPEG encoding scheme. It is possible that he (and some others) are lumping together software low-pass filtering and the JPEG compression scheme.

However, you've said that this is a bizarre and unlikely answer, too. Do you have any other theory that might make sense?

Although, in the end, whatever it might be could theoretically be addressed in a firmware upgrade, unfortunately the upgrade method and the fact they've already sort-of "recalled" the camera twice makes another upgrade exceedingly unlikely, IMHO.

Best wishes to all.

Vtie said:

Vtie said:

You perfectly know that is up to the encoder designer to put some
amount of integration or low pass filtering in order to minimize
blocking. There you can lose resolution. Now, to save the day you
may hide yourself behind technicalities and say that this is not
part of the jpeg engine but you know what I meant in the first
place. You also perfectly know that you can make a DCT engine in
many possible ways, some are better than others. Also read entirely
my post please, I don't exclude that it may be just a big low pass
at the beginning of everything.

Click to expand...

Yes, a low pass filter prior to JPEG encoding may explain the D100
softness of JPEG images. Personally, I would never call this filtering
a part of the JPEG encoding engine. I would call it, well, a low
pass filter prior to JPEG encoding. If you call this a part of the
encoding we are getting into terminology and I am not going to argue
over that. Then I guess we totally agree, apart from what you call
JPEG encoding.

However, it sounds like a very exotic design to me if they would apply
a sharpening filter to the data (like most of the modes are doing, at
least according to their names), and then again a low pass filter
prior to
encoding! Do you think Nikon would do that kind of strange things?
It's a waste of quality anyway!

Vtie

Click to expand...