Imaging resource test shots are up

[Erik Magnuson]

The IR review says it's noisy, White balance is off, Autofocus is
bad. Few of the shots lack dynamic range, other sets have
"excellent dynamic range".

Phil tested Mr. Hyde, while Dave tested Dr. Jeckle ;-)

About the JPEG issue: it's certainly an issue if the most popular option has a significant problem. Is it fair to treat this camera differently than other cameras that also support RAW? Phil solved this dilemma with re-testing with D60 RAW, but choosing the alternate path is just as valid.

Besides, I think that people are ascribing too many things to the JPEG bugs. 2x2 subsampling will mainly affect some color boundaries and the (IHMO separate) solid color posterization issue was not one of the problems mentioned.

--
Erik

 

[Ladehaug]

About the JPEG issue: it's certainly an issue if the most popular
option has a significant problem. Is it fair to treat this camera
differently than other cameras that also support RAW? Phil solved
this dilemma with re-testing with D60 RAW, but choosing the
alternate path is just as valid.

Besides, I think that people are ascribing too many things to the
JPEG bugs. 2x2 subsampling will mainly affect some color boundaries
and the (IHMO separate) solid color posterization issue was not one
of the problems mentioned.

Regarding the JPG artifacts issue, it will certainly depend on what properties you want to consider regarding a given image. I have earlier proven that solid color posterization can occur because of this JPG bug / limitation in the Sigma Photo Pro software:



The posterziation effect becomes clearly visible when you magnify this sample 200%.

If you want more comparison samples, I can make a lot of them.. (thanks to Samuel, Tony and Jeffrey - I now have 79 RAW files to pick my samples from!)

Cheers,
Geir Rune

 

[Kevin Blair]

Erik,

I'm not that suprised that two different reviewers could reach different conclusions about the SD9 for several of reasons.

First, if the reviewer focuses on the potential of the camera/concept, they may be more likely to de-emphasize some of the first release issues. Conversely, if the reviewer focuses on the camera 'as-is' for a present day purchaser, they are going to look for the warts.

Second, there may be inconsistent performance between individual cameras. Based on the problems people have experienced with dust, and general comments about the level of QC with respect to Sigma's lenses, its possible that Phil may have received a better review unit than did Dave.

Third, each reviewer may have different approaches to conducting a review and place greater importance on different areas of a camera's operation.

People need to evaluate reviews in total and make a decision based upon their intended use of the camera and their experience level. There isn't a DSLR out there that meets everyone's needs.

Kevin

The IR review says it's noisy, White balance is off, Autofocus is
bad. Few of the shots lack dynamic range, other sets have
"excellent dynamic range".

Phil tested Mr. Hyde, while Dave tested Dr. Jeckle ;-)

About the JPEG issue: it's certainly an issue if the most popular
option has a significant problem. Is it fair to treat this camera
differently than other cameras that also support RAW? Phil solved
this dilemma with re-testing with D60 RAW, but choosing the
alternate path is just as valid.

Besides, I think that people are ascribing too many things to the
JPEG bugs. 2x2 subsampling will mainly affect some color boundaries
and the (IHMO separate) solid color posterization issue was not one
of the problems mentioned.

--
Erik

 

[Erik Magnuson]

Regarding the JPG artifacts issue, it will certainly depend on what
properties you want to consider regarding a given image. I have
earlier proven that solid color posterization can occur because of
this JPG bug / limitation in the Sigma Photo Pro software:

It's clear that JPEG problems do exist. However, it does not seem that these problems had a significant effect on Dave's tests or conclusions.

--
Erik

 

[Frank C.]

You like it so you praise it. And then what? This is the general
attitude in this forum - including you... Everybody can read what
he likes and be happy.
My opinion is that IR review is worthless unless he counts in the
Jpeg compression artifects at least. His review is loaded with
subjective comments about AF and noise also.

yep, I agree with Zoltan. It appears that IR did that "I'm smarter
than everyone else and will prove it by showing all of you
'REAL' problems with the SD-9 " thingy. Gallant try on their part
and they may have a couple of valid points but it's still funny how
their review came out AFTER Phil's...mmm... I thought IR
had their camera way before Phil, or am I mistaken.....8o ?

Also, from the way I read their review I get the impression the SD-9's
inherent sharpness is helped in a marked way from the fact it doesn't
have an AA filter. Isn't it the way the X3 sensor is built that permits
such a thing to happen and wouldn't this be considered a good
thing? IR seems to be knocking it. bohhh... I'm really dazed and
confused now.....are we progressing forward with the X3 or not?

It is clear Dave has a much better understanding of Alaising than
most here, if he had posted it here, people would beat up on him as
a basher.

Read his complete explanation of the topic, you just might learn
something.

http://www.imaging-resource.com/PRODS/SD9/SD9PICS.HTM

--
Best wishes,
Zoli

--
Best wishes,
Zoli

 

[Peter G]

yep, I agree with Zoltan. It appears that IR did that "I'm smarter
than everyone else and will prove it by showing all of you
'REAL' problems with the SD-9 " thingy. Gallant try on their part

The majority is not always right, several of us here who have knowledge of signal theory mentioned the same issue with respect to Alaising. Dave tried to correct a misconception based on his greater experience in the area. I don't see this as an ego trip.

Also, from the way I read their review I get the impression the SD-9's
inherent sharpness is helped in a marked way from the fact it doesn't
have an AA filter. Isn't it the way the X3 sensor is built that
permits
such a thing to happen and wouldn't this be considered a good
thing? IR seems to be knocking it. bohhh... I'm really dazed and
confused now.....are we progressing forward with the X3 or not?

It is less subject to artifacts than a bayer camera would be without an AA filter, but there will still be some. And they have already showed up.

http://www.dpreview.com/forums/read.asp?forum=1027&message=3766107

This is from the clothesline shot.

You tend to get luminance moire, it is still objectionable. As Dave said, Alaising is not a good thing in general. It will crop up at odd times to bite you. When it is not going screwy, it will result in a sharper image appearance. It is a tradeoff. Here is a Quote of Dave:

"Just to state it clearly, aliasing is always a bad thing, because it introduces spurious information into an image that doesn't reflect the subject detail (and in most cases, actually obscures it)."

This is true and anyone who knows signal theory will agree with this.

So yes the sigma will be sharper because of the lack of the filter, but be aware that if you are shooting very fine patterns near the limit, that you may be unexpected moire showing up when you least want or expect it.

I think fabrics will exhibit this most, just as fabrics do strange things on TV.

If you are a landscape photographer you would probably never see it.

But no matter how this tradeoff affects you, the idea of detail beyond nyquist is nonsence.

Peter

 

[Frank C.]

yep, I agree with Zoltan. It appears that IR did that "I'm smarter
than everyone else and will prove it by showing all of you
'REAL' problems with the SD-9 " thingy. Gallant try on their part

The majority is not always right, several of us here who have
knowledge of signal theory mentioned the same issue with respect to
Alaising. Dave tried to correct a misconception based on his
greater experience in the area. I don't see this as an ego trip.

Also, from the way I read their review I get the impression the SD-9's
inherent sharpness is helped in a marked way from the fact it doesn't
have an AA filter. Isn't it the way the X3 sensor is built that
permits
such a thing to happen and wouldn't this be considered a good
thing? IR seems to be knocking it. bohhh... I'm really dazed and
confused now.....are we progressing forward with the X3 or not?

It is less subject to artifacts than a bayer camera would be
without an AA filter, but there will still be some. And they have
already showed up.

http://www.dpreview.com/forums/read.asp?forum=1027&message=3766107

This is from the clothesline shot.

You tend to get luminance moire, it is still objectionable. As Dave
said, Alaising is not a good thing in general. It will crop up at
odd times to bite you. When it is not going screwy, it will result
in a sharper image appearance. It is a tradeoff. Here is a Quote
of Dave:

"Just to state it clearly, aliasing is always a bad thing, because
it introduces spurious information into an image that doesn't
reflect the subject detail (and in most cases, actually obscures
it)."

This is true and anyone who knows signal theory will agree with this.

So yes the sigma will be sharper because of the lack of the filter,
but be aware that if you are shooting very fine patterns near the
limit, that you may be unexpected moire showing up when you least
want or expect it.

I think fabrics will exhibit this most, just as fabrics do strange
things on TV.

for some reason it was fabrics that most impressed me with the
SD-9.... the bayer cameras would smudge everything ( because
of the AA filter + interplolation combo ) whereas as the X3
delivered crisp details and patterns. Maybe this was that
famous "detail beyond nyquist" bit, I dunno, but whatever it was
made the SD-9 images incredibly lifelike. Maybe it goes down to
personal choice. I prefer these 'pseudo' or 'imaginary' X3 details
over the smudges and washed out details of a bayer camera...

If you are a landscape photographer you would probably never see it.

But no matter how this tradeoff affects you, the idea of detail
beyond nyquist is nonsence.

Peter

 

[Peter G]

pg > I think fabrics will exhibit this most, just as fabrics do strange
pg > things on TV.

for some reason it was fabrics that most impressed me with the
SD-9.... the bayer cameras would smudge everything ( because
of the AA filter + interplolation combo ) whereas as the X3
delivered crisp details and patterns. Maybe this was that
famous "detail beyond nyquist" bit, I dunno, but whatever it was
made the SD-9 images incredibly lifelike. Maybe it goes down to
personal choice. I prefer these 'pseudo' or 'imaginary' X3 details
over the smudges and washed out details of a bayer camera...

Well I can guarantee the location pointed to by the arrow looks more like the d60 photo than the Sigma photo. You take a photo and all of a sudden you get this wierd waveform pattern, you may like it, but I don't and if you are taking photos for money, "My camera just does that" may not cut it as an explanation.

Note this pattern is not fine grain detail, its low frequency noise, the pattern is very big and would easily be resolved by the D60, if it were actually present. Sometimes the artifacts are fine grained and give the look of texture and that can be visually pleasing. Other times you can get big ugly low frequency patterns. It will be very hard to predict unless you can calculate the beat frequency just by looking at the fabric pattern.

 

[Frank C.]

pg > I think fabrics will exhibit this most, just as fabrics do strange
pg > things on TV.

for some reason it was fabrics that most impressed me with the
SD-9.... the bayer cameras would smudge everything ( because
of the AA filter + interplolation combo ) whereas as the X3
delivered crisp details and patterns. Maybe this was that
famous "detail beyond nyquist" bit, I dunno, but whatever it was
made the SD-9 images incredibly lifelike. Maybe it goes down to
personal choice. I prefer these 'pseudo' or 'imaginary' X3 details
over the smudges and washed out details of a bayer camera...

Well I can guarantee the location pointed to by the arrow looks
more like the d60 photo than the Sigma photo. You take a photo and
all of a sudden you get this wierd waveform pattern, you may like
it, but I don't and if you are taking photos for money, "My camera
just does that" may not cut it as an explanation.

Note this pattern is not fine grain detail, its low frequency
noise, the pattern is very big and would easily be resolved by the
D60, if it were actually present. Sometimes the artifacts are fine
grained and give the look of texture and that can be visually
pleasing. Other times you can get big ugly low frequency patterns.
It will be very hard to predict unless you can calculate the beat
frequency just by looking at the fabric pattern.

SD-9 moire looks pretty bad but then again all the detail
in the shoulder section of the apparel in the D60 shot
is lost or looks washed out. I've seen D60/ S2 shots with
moire as bad as this SD-9 picture exhibits. Visible interference
like this is not as frequent as one might think
tho, it only pops up when certain frequencies manifest
themselves. On the whole the X3 is certainly well-bahaved
and the bonus is that every pixel contributes to picture
definition/resolution ( a BIG plus )... or as Phil
atates it- SD9 is first camera with single pixel resolution.

 

[Peter G]

SD-9 moire looks pretty bad but then again all the detail
in the shoulder section of the apparel in the D60 shot
is lost or looks washed out. I've seen D60/ S2 shots with
moire as bad as this SD-9 picture exhibits. Visible interference
like this is not as frequent as one might think
tho, it only pops up when certain frequencies manifest
themselves. On the whole the X3 is certainly well-bahaved
and the bonus is that every pixel contributes to picture
definition/resolution ( a BIG plus )... or as Phil
atates it- SD9 is first camera with single pixel resolution.

Just pointing out that the point Dave makes is a valid one and it occurs in real life. There are about 6 or 7 more serious issues I have with this camera. There could probably be a weak AA filter that would prevent the gross moire in the hood while still pulling more detail in the fabric. It is primarily a money decision to leave it out.

In the same shootout I also see objectionable moire in the S2 as well (not quite as obvious as this), I think it has a weaker AA filter.

I have read all 3 reviews (here, IR and Steves) and I don't really see much in the way of contradictions in their results, just the spin on the final conclusions. This is 1.0 technology which has promise.

I stay away from the bleeding edge of technology these days. Have fun beta testing.

Pter

 

[Roger]

I'm not that suprised that two different reviewers could reach
different conclusions about the SD9 for several of reasons.

First, if the reviewer focuses on the potential of the
camera/concept, they may be more likely to de-emphasize some of the
first release issues. Conversely, if the reviewer focuses on the
camera 'as-is' for a present day purchaser, they are going to look
for the warts.

Second, there may be inconsistent performance between individual
cameras. Based on the problems people have experienced with dust,
and general comments about the level of QC with respect to Sigma's
lenses, its possible that Phil may have received a better review
unit than did Dave.

Third, each reviewer may have different approaches to conducting a
review and place greater importance on different areas of a
camera's operation.

People need to evaluate reviews in total and make a decision based
upon their intended use of the camera and their experience level.
There isn't a DSLR out there that meets everyone's needs.

Kevin

The IR review says it's noisy, White balance is off, Autofocus is
bad. Few of the shots lack dynamic range, other sets have
"excellent dynamic range".

Phil tested Mr. Hyde, while Dave tested Dr. Jeckle ;-)

About the JPEG issue: it's certainly an issue if the most popular
option has a significant problem. Is it fair to treat this camera
differently than other cameras that also support RAW? Phil solved
this dilemma with re-testing with D60 RAW, but choosing the
alternate path is just as valid.

Besides, I think that people are ascribing too many things to the
JPEG bugs. 2x2 subsampling will mainly affect some color boundaries
and the (IHMO separate) solid color posterization issue was not one
of the problems mentioned.

--
Erik

As with any new product there will problems to work out. Like the Minolta Dimage 7 a year ago. It needed some changes but now we have the Dimage 7HI the best prosumer non DSLR camera for the money. The SD9 looks like it will do a great job if you know how to use it. I'm sure the price will decrease making it the best buy for the money like the Sigma SA9. These reviews are just that, reviews, their all saying some different, good and bad. I'll bet you can find something good and bad about the rest of the digital cameras as well. I know that the SD9 really jerks the Nikon and Canon people, why I don't know??? Go figure. They might as well get use to it, because Sigma is here to stay. Who pays for these reviews? I thought Phil did a great job on his both pros and cons.
Enjoy
Roger J