Just for New Users.. A700 High ISO NR Why Adobe Tests are not valid.

zxaar said:

zxaar said:

In the beginning there was CCD and CMOS. CMOS was more noisy so most
cameras used CCD. Then Canon added on sensor noise handeling and
that removed the noise and grain of higher ISO and it was good. Then
on the next Day.. Sony added more Noise control and it was good
except for those crying in the dessert for more grain...

Click to expand...

you are correct about this part, and all the camera makers do this.
The only issue is when it comes to SONY Bary thinks it is noise
reduction (as in noise Ninja type) vs the reduction of noise (as what
CANON applies to their sensors).
As a person who has to write such algorithms, I could think that
implementing an algorithm is not bad thing, and if it is applied at
sensor level then it would be more effective given that there is live
information about signal present on sensor. (the whole philosophy of
high performance sensors).

Click to expand...

Sony is in betweem.. they have added in the sensor design some features that might have been reserved for software.. but that is how things go.

I am saying that some RAW processors (cough Adobe) make look more heavy handed that it is, and that this is the RAW for this sensor design.

but I did advocate that value of giving a user more control of the NR levels even from the sensor.
--
------------
Ken - Happy A700 Owner
http://www.cascadephotoworks.com

Chadd Thompson said:

You are grasping here Ken :O)

I still don't agree, looks to me that the bible image was sharpened
and contains some atrifacts. But if you look at your samples without
any color noise the LR images appears to me to have more tonal range
to it. Whereas the Bible image is more one color of grey for the same
area. But like I said I can't really tell much from what you have
posted.

Click to expand...

We are splitting hairs.. for sure.. bibble, Raw therapee and from what Saw ACDsee all before they begin PP start with finer grain and list blotches than Adobe.

you and I both like the output.. the question has been did Sony go to far? with the right converter there is much less difference between sony output and others..that is all I am saying.

Chadd Thompson said:

Ken_5D said:

Chadd Thompson said:

Ken_5D said:

Except your wrong you CS2 image has shapening artifacts along the
color blob edges..

Click to expand...

As I stated in another post Bible is sharpening the image in its
conversion. RAW files in general are not sharp unless you apply some
sharpening. Also the image I posted of yours has been jpg compressed
at least 3 times.

Chadd Thompson said:

Chadd Thompson said:

My point still stands that "if the RAW files are cooked then it is
the A700 that is doing it not your raw converter".

Now do you guys see why I wish this whole thing would just be
dropped? We are never going to come to a conclusion on it.

Click to expand...

Problem is others bring it up.. and you in this case create an image
that a casual glance says Adobe is OK..

Click to expand...

The images I posted are not a casual glance, they are a 100% crop of
the images that you posted. Again the image I edited has been jpg
compressed at least three times.

Click to expand...

So lets look closer.. at the patterns, I can see this different at
100% maybe you cant;

Not the same.. more evenly distribuited pixel level noise in Bibble
than LR

Chadd Thompson said:

Chadd Thompson said:

------------
Ken - Happy A700 Owner
http://www.cascadephotoworks.com

Click to expand...

--
Thanks,
Chadd
http://www.pbase.com/chadd

Click to expand...

--
------------
Ken - Happy A700 Owner
http://www.cascadephotoworks.com

Click to expand...

--
Thanks,
Chadd
http://www.pbase.com/chadd

Click to expand...

--
------------
Ken - Happy A700 Owner
http://www.cascadephotoworks.com

that I'm in awe of all the technical knowledge on this forum. I can't even imagine being able to argue the technical issues that you all are trying to make... I sincerely am amazed. The more this digital world develops the more it leaves me "technically" behind...

To me and what I focus on are:

1. Either the image looks one way or another (equiating to what I like/want), and

2. It either "can or can't" be adjusted in or out of the camera to creat an image I like. Caveat - being the simpler the adjustments needed - the better.

All this allows me to focus (no pun intended) on composition, focus points (or focus type), ambient or added lighting and metering.

I'm so impressed with what Sony has done in their user interface, feel, features and especailly in-body stabilization and connection w/Zeiss. (Used Minolta for years and loved it.) I have at present only two considerations on whether to go A700 or D300:

1) IQ - at all ISO's up to and including 1600 (and "to my eyes" how they handle and render noise and detail; and

2) Flash System - specifically, consistency of exposure in both direct and bounce (indoors) and exposure consistency outdoors for "fill".

"How" the system makes all this happen is well beyond my level of technical understanding... what matters is simply whether it does or doesn't do it. My personal ability to facilitate a Nikon or Sony decision to change the way the design cameras is obviously a moot point (it doen't exist).

Bob

--
Bob

I agree,

I do think Sony is doing some strange stuff with the NR but like I said it has not caused me to miss any shots and I have been able to process them to acceptable levels for me. The a700 is by no means a broken camera and can produce some very nice images.

I say we both get some nice high ISO shots to have ready for the next bout of noise issues, we post them and say "hey if it is so bad explain this".

Sound like a deal?

Ken_5D said:

you and I both like the output.. the question has been did Sony go to
far? with the right converter there is much less difference between
sony output and others..that is all I am saying.

Ken_5D said:

Ken_5D wrote:

Click to expand...

--
------------
Ken - Happy A700 Owner
http://www.cascadephotoworks.com

Click to expand...

--
Thanks,
Chadd
http://www.pbase.com/chadd

Lord, as much as it pains me to do this....

http://www.sonolta.com/sony-photos/v/Sony+A700+review/sony-a700-iso-6400/

Just point them to these 6400iso shots.
(man, I gotta get a drink)
--

'Dodging and burning are steps to take care of mistakes God made in establishing tonal relationships.'

Chadd Thompson said:

I agree,

I do think Sony is doing some strange stuff with the NR but like I
said it has not caused me to miss any shots and I have been able to
process them to acceptable levels for me. The a700 is by no means a
broken camera and can produce some very nice images.

I say we both get some nice high ISO shots to have ready for the next
bout of noise issues, we post them and say "hey if it is so bad
explain this".

Sound like a deal?

Click to expand...

Can we get a LOUD AMEN!!! Chadd, Ken, thank you for making this clear for ALL of us! Ken especially for you for fighting the good fight..

Bob,

I'm surprised you STILL haven't purchased either system yet!!! I remember a few months ago you were still pondering...? LOL! That's ok, I think that you be happy either the A700 or D300, the A700 is really nice but with the D300, you'll be getting more ooohs and ahhhs from people who recognize the NIKON label and see that BIG camera... My brother-in-law w/ his D200 was very impressed with the A700 as much as I was of his D200 (when I had my A100)...

You can't go wrong with either system and as you can see from PMA 2008 SONY is in the dSLR market for the LONG RUN so don't worry about being a "closed end" system. Sony is here to stay and they're going to EXPAND the A-MOUNT system with new lenses and new bodies... (I remember this was one of your original concerns with the SONY...)

--
-Alex

From the minds of Minolta to the imagination of Sony, Alpha, like no other.

http://www.pbase.com/lonewolf69

To me, that diagram implies NR processing on RAW.

I cannot see anywhere in that diagram specific mention of JPG which people are assuming. Is it there, have I missed it?

What I read with that diagram is a signal flow from sensor to memory card and the word conversion being used to convert the signal that comes off the sensor into a format that can be written to memory card, ie the RAW file format.

The data that comes off the sensor has to be converted whether it is going into a raw file or a jpg file simply because it has to be molded into a series of numbers that can be stored in the file and in the format defined for that file type. Conversion does not necessarily imply jpg only.

However, I do understand that the diagram can be taken to mean jpg only but that is far from clear and with the wording as it is on the diagram, both conclusions (raw and jpg or jpg only) are as valid as each other.

Ed

It's interesting to some, Bob; and to others they really couldn't give a flip. If you had someone sit down and explain it to you, you may be interested to hear how it is done, but in the end what you are saying is what really matters. To Ken, and Chadd, I understand both of your arguments, which in the end is not really that different. But while you two were conferring, and also in other posts, there are those of us who are really looking for the best software to settle on to handle our A-700 files. While I now use Raw Therapee, and CS-2 with DNG converter, I am hoping to find a one shot solution to this method. I have recently acquired CS-3 and was looking at Lightroom. But if Bibble Pro or ACDSee will handle these in one workflow, that is what some of us are looking at! This is why there is still interest in threads such as this(for me) anyway! Glad you have agreed to match 3200 raw files to qualm the naysayers, and with a good software solution, maybe we as listeners and learners can help! Thanks to you all for working on this issue! No Barry, I'm not going to leave you out, because it is like continual research for you, so we follow your threads on this issue also! And a big "up" to David K. who always intervenes with his knowlege to help out! To all of you I say Thanks!
--
Glenn

I'm kinda partial to video, but I'm hangin!

robminchin said:

So the BIONZ chip works on the RAW to make a better JPEG, but does
not modify the data saved to the RAW file, which is the input data to
the BIONZ chip.

Click to expand...

The Bionz chip does work on the RAW data before writing it. First of all, it performs the entire task of compressing the cRAW file into an 8-bit version which retains all salient 12-bit information; with either the RAW or cRAW, it handles the I/O operations.

The BIONZ chip is responsible for creating the JPEG preview, which is like the standard separate JPEG passed through the IRIDIX processor when switched into the pipeline for DRO. This is embedded in the RAW file.

As explained to me by Paul Genge at the product launch, the raw file is subject to NR. This was a major part of his explanation about the additional NR steps in the Sony workflow, compared to other cameras - that there was 'raw noise reduction' before 'JPEG high ISO noise reduction' and after 'on sensor' NR. Paul is a technical sales manager - a former training expert, turned product manager, then returned to a modified role with much technical involvement. He probably has a better technical understanding of the workings of his products than 90 per cent of Sony staff, and he has been responsible (in Europe) for English language training and sales presentations.

Frankly, if there is no Raw NR, then Sony hardly need to have mentioned it. All cameras normally put NR as part of the JPEG writing process, after the RAW is arrayed in the buffer ready to write to card. Sony seem to me to have made a specific point that NR has an additional stage of NR, before the raw file is converted to JPEG (which is the stage where High ISO NR takes place).

I think the flowcharts are ambiguous but Barry's intepretation is in line with the Sony press conferences at the time of the camera launch, and Ken's intepretation from PMA contradicts the original carefully explained three-stage NR process.

David

--
Publishing & Editing Photoworld (photoclubalpha.com) and Master Photo Digital
Currently writing for f2 and the BJP

David,

Given this interpretation is right, is there something like a final verdict? Sony certainly raised a controversy, but was it a mistake, a good idea, or something that does not make much of a difference in the long run (as long as we do appropriate PP)? I remember that you seemed to like Nikon's approach to NR better but that you preferred the A700's color accuracy. What's your current view of the comparative IQ of the various top ASP-C cameras?
--Jan

David Kilpatrick said:

David Kilpatrick said:

So the BIONZ chip works on the RAW to make a better JPEG, but does
not modify the data saved to the RAW file, which is the input data to
the BIONZ chip.

Click to expand...

The Bionz chip does work on the RAW data before writing it. First of
all, it performs the entire task of compressing the cRAW file into an
8-bit version which retains all salient 12-bit information; with
either the RAW or cRAW, it handles the I/O operations.

The BIONZ chip is responsible for creating the JPEG preview, which is
like the standard separate JPEG passed through the IRIDIX processor
when switched into the pipeline for DRO. This is embedded in the RAW
file.

As explained to me by Paul Genge at the product launch, the raw file
is subject to NR. This was a major part of his explanation about the
additional NR steps in the Sony workflow, compared to other cameras -
that there was 'raw noise reduction' before 'JPEG high ISO noise
reduction' and after 'on sensor' NR. Paul is a technical sales
manager - a former training expert, turned product manager, then
returned to a modified role with much technical involvement. He
probably has a better technical understanding of the workings of his
products than 90 per cent of Sony staff, and he has been responsible
(in Europe) for English language training and sales presentations.

Frankly, if there is no Raw NR, then Sony hardly need to have
mentioned it. All cameras normally put NR as part of the JPEG writing
process, after the RAW is arrayed in the buffer ready to write to
card. Sony seem to me to have made a specific point that NR has an
additional stage of NR, before the raw file is converted to JPEG
(which is the stage where High ISO NR takes place).

I think the flowcharts are ambiguous but Barry's intepretation is in
line with the Sony press conferences at the time of the camera
launch, and Ken's intepretation from PMA contradicts the original
carefully explained three-stage NR process.

David

--
Publishing & Editing Photoworld (photoclubalpha.com) and Master Photo
Digital
Currently writing for f2 and the BJP

Click to expand...

--
http://www.pbase.com/koster10/

This is why I suggested that whomever created/approved that diagram be shot.

ed209 said:

To me, that diagram implies NR processing on RAW.

I cannot see anywhere in that diagram specific mention of JPG which
people are assuming. Is it there, have I missed it?

What I read with that diagram is a signal flow from sensor to memory
card and the word conversion being used to convert the signal that
comes off the sensor into a format that can be written to memory
card, ie the RAW file format.

The data that comes off the sensor has to be converted whether it is
going into a raw file or a jpg file simply because it has to be
molded into a series of numbers that can be stored in the file and in
the format defined for that file type. Conversion does not
necessarily imply jpg only.

However, I do understand that the diagram can be taken to mean jpg
only but that is far from clear and with the wording as it is on the
diagram, both conclusions (raw and jpg or jpg only) are as valid as
each other.

Ed

Click to expand...

--
Direct your eye right inward, and you'll find a thousand
regions in your mind Yet undiscovered. Travel them, and
be Expert in home-cosmography.
-H.D. Thoreau

that Sony went for color accuracy and Nikon went for noise. Additionally, their tests show that the grain of the noise is smaller on the D300 than on the A700.

Hopefully Sony will give us something to turn off Bionz.

Cheers!

jan-k said:

David Kilpatrick said:

jan-k said:

So the BIONZ chip works on the RAW to make a better JPEG, but does
not modify the data saved to the RAW file, which is the input data to
the BIONZ chip.

Click to expand...

The Bionz chip does work on the RAW data before writing it. First of
all, it performs the entire task of compressing the cRAW file into an
8-bit version which retains all salient 12-bit information; with
either the RAW or cRAW, it handles the I/O operations.

The BIONZ chip is responsible for creating the JPEG preview, which is
like the standard separate JPEG passed through the IRIDIX processor
when switched into the pipeline for DRO. This is embedded in the RAW
file.

As explained to me by Paul Genge at the product launch, the raw file
is subject to NR. This was a major part of his explanation about the
additional NR steps in the Sony workflow, compared to other cameras -
that there was 'raw noise reduction' before 'JPEG high ISO noise
reduction' and after 'on sensor' NR. Paul is a technical sales
manager - a former training expert, turned product manager, then
returned to a modified role with much technical involvement. He
probably has a better technical understanding of the workings of his
products than 90 per cent of Sony staff, and he has been responsible
(in Europe) for English language training and sales presentations.

Frankly, if there is no Raw NR, then Sony hardly need to have
mentioned it. All cameras normally put NR as part of the JPEG writing
process, after the RAW is arrayed in the buffer ready to write to
card. Sony seem to me to have made a specific point that NR has an
additional stage of NR, before the raw file is converted to JPEG
(which is the stage where High ISO NR takes place).

I think the flowcharts are ambiguous but Barry's intepretation is in
line with the Sony press conferences at the time of the camera
launch, and Ken's intepretation from PMA contradicts the original
carefully explained three-stage NR process.

David

--
Publishing & Editing Photoworld (photoclubalpha.com) and Master Photo
Digital
Currently writing for f2 and the BJP

Click to expand...

--
http://www.pbase.com/koster10/

Click to expand...

--
Direct your eye right inward, and you'll find a thousand
regions in your mind Yet undiscovered. Travel them, and
be Expert in home-cosmography.
-H.D. Thoreau

jan-k said:

David,
Given this interpretation is right, is there something like a final
verdict? Sony certainly raised a controversy, but was it a mistake, a
good idea, or something that does not make much of a difference in
the long run (as long as we do appropriate PP)? I remember that you
seemed to like Nikon's approach to NR better but that you preferred
the A700's color accuracy. What's your current view of the
comparative IQ of the various top ASP-C cameras?

Click to expand...

For my purposes, the Nikon D300 works slightly better than the A700 because my files are checked for 'artefacts' by library editors, and the A700 file generally looks to have more artificial edges and transitions present.

However, since the Nikon requires twice the exposure of the A700 in real terms to give the same amount of shadow detail, the comparison is flawed. The Sony at 3200 is generally going to give the same dynamic range (with whatever raw adjustments are needed) as the Nikon at 6400, and far superior colour.

Here is an example at IS0 3200, set to -2 compensation after testing on the A700, processed using the dreaded ACR 4.3.1 - which retains very good colour. The shot is taken on the 11-18mm at full aperture, 1/10th of a second hand held. The processing does not clip the blacks (Black Level 5, default) but the night sky has just a trace of noise speckle - it is extremely clean. The sharpness is ridiculous for a handheld shot with SSS. The roof tiles, the details inside the windows of the towers (handrails and stuff), the fences at the extreme left, etc - they are as sharp as you could ever possibly expect from a 16mm equivalent wide open zoom.



The colour is vibrant, there are no reflections or ghosts or even any hard colour fringers (regular -30 CA applied) despite extremely tasking light conditions. I have put the full size image on pBase:

http://www.pbase.com/davidkilpatrick/image/92554900

I have not had anything equal from the D300 when using it, and the A700 has consisently got some great low light images with lots of colour and plenty of detail, using ACR not Bibble (I can't use Bibble, the sharpening artefacts are too strong for library use - the conversion is not soft enough, may sound contradictory to this thread, but I don't need emphasised edge sharpness).

This image has been accepted on Alamy rezzed up to 5120 pixels, and was submitted with a complete set of night images for quality control, not slipped in in the middle of some big batch of shots. It's not an ultimate criterion but you would be amazed at the Canon 5D and even 1Ds MkXX shooters who complain of getting their files rejected for artefacts or interpolation or sofness (etc etc).

David

--
Publishing & Editing Photoworld (photoclubalpha.com) and Master Photo Digital
Currently writing for f2 and the BJP

This makes a ton of sense and if the Sony techincal guys says they do NR on RAW then I consider that better confirmation than a SR Product Manager. Since that is my training, I know technical stuff can get lost in the over all product management which is a very big job.

I hope to talk to Mark in the future soon. and am bookmarking your post. so I can make a case for a Firmware off to the Bionze NR (assuming LOW isn't a miss named Off to Bionz) Not sure it will happen, since I am sure they are well into the A750 update planning now.. for Next fall or PMA 2009.

I officially stand properly corrected on where the final NR on RAW is..

It still it is not as bad as some RAW convertrs make it look when they enforce a another level of NR.. but that is a different story.

Barry, you were right and I was wrong. about the Bionz.. I am still right about everything else he he he

------------
Ken - Happy A700 Owner
http://www.cascadephotoworks.com

JusLookN said:

It's interesting to some, Bob; and to others they really couldn't
give a flip. If you had someone sit down and explain it to you, you
may be interested to hear how it is done, but in the end what you are
saying is what really matters. To Ken, and Chadd, I understand both
of your arguments, which in the end is not really that different.
But while you two were conferring, and also in other posts, there are
those of us who are really looking for the best software to settle on
to handle our A-700 files. While I now use Raw Therapee, and CS-2
with DNG converter, I am hoping to find a one shot solution to this
method. I have recently acquired CS-3 and was looking at Lightroom.
But if Bibble Pro or ACDSee will handle these in one workflow, that
is what some of us are looking at! This is why there is still
interest in threads such as this(for me) anyway! Glad you have
agreed to match 3200 raw files to qualm the naysayers, and with a
good software solution, maybe we as listeners and learners can help!
Thanks to you all for working on this issue! No Barry, I'm not going
to leave you out, because it is like continual research for you, so
we follow your threads on this issue also! And a big "up" to David
K. who always intervenes with his knowlege to help out! To all of
you I say Thanks!
--
Glenn

I'm kinda partial to video, but I'm hangin!

Click to expand...

I am working on some ACDSee tests from ISO 200-6400 to help with that need.
------------
Ken - Happy A700 Owner
http://www.cascadephotoworks.com

David Kilpatrick said:

As explained to me by Paul Genge at the product launch, the raw file
is subject to NR. This was a major part of his explanation about the
additional NR steps in the Sony workflow, compared to other cameras -
that there was 'raw noise reduction' before 'JPEG high ISO noise
reduction' and after 'on sensor' NR. Paul is a technical sales
manager - a former training expert, turned product manager, then
returned to a modified role with much technical involvement. He
probably has a better technical understanding of the workings of his
products than 90 per cent of Sony staff, and he has been responsible
(in Europe) for English language training and sales presentations.

Frankly, if there is no Raw NR, then Sony hardly need to have
mentioned it. All cameras normally put NR as part of the JPEG writing
process, after the RAW is arrayed in the buffer ready to write to
card. Sony seem to me to have made a specific point that NR has an
additional stage of NR, before the raw file is converted to JPEG
(which is the stage where High ISO NR takes place).

I think the flowcharts are ambiguous but Barry's intepretation is in
line with the Sony press conferences at the time of the camera
launch, and Ken's intepretation from PMA contradicts the original
carefully explained three-stage NR process.

David

Click to expand...

Guys,

If you ever get to talk to the tech folks at Sony again, just ask them this question...

Does the Bionz do a NR pass on the RAW data before writing it as RAW on the card. That's a real simple yes or no answer, and won't give away any trade secrets. That's the answer we're looking for. Yes or No.
--

'Dodging and burning are steps to take care of mistakes God made in establishing tonal relationships.'

Ken_5D said:

I hope to talk to Mark in the future soon. and am bookmarking your
post. so I can make a case for a Firmware off to the Bionze NR
(assuming LOW isn't a miss named Off to Bionz) Not sure it will
happen, since I am sure they are well into the A750 update planning
now.. for Next fall or PMA 2009.

Click to expand...

This is really what counts, to see what sony do, if anything.

NR off = Buy..and you can bookmark this too ;-)

--



Clint is on holiday! Soon to return! ;-)

Ken_5D said:

I am working on some ACDSee tests from ISO 200-6400 to help with that
need.

Click to expand...

Since it does not seem to have a mac version it's of no use to me.

Since you and a bunch of others have just called Mark a flat out liar, are you willing to say that to his face? How about the rest? Let's see actual posts from each.

And for those who wondered why I should note problems with a a700 and ask me for what, you still don't have the answer and won't get it from me.

Walt

--
Direct your eye right inward, and you'll find a thousand
regions in your mind Yet undiscovered. Travel them, and
be Expert in home-cosmography.
-H.D. Thoreau

--Thank You Ken for the Explanation....But by Gum! You lot leave me baffled with all this techno babble..LOL..

All I really want to know is this...If I take an iso200/400/800 RAW..Put it through Bibble or Adobe RAW 4.2 then Save as a TIFF file, will the Finished Image Look EXCELLENT as an A3/15x11..If NOT I won't buy..If so I'll Definatly Buy when price drops to £700 or less.
THANKS.

MrScary (DennisR)
Swansea, Wales. UK

http://www.pbase.com/dennisr
http://community.webshots.com/user/mrscarecrow
http://www.russ4to.KMD5DImages.photoshare.co.nz