Detail Man: Richard Butler and Rishi Sanyal:
These two articles about sources of image-noise are well presented and explained in your statements of clarification made in comments sections.
I like the way that DPReview writers appear to have the adopted the terminology "ISO-invariant" / "ISO-variant" in lieu of the (IMO, less descriptive) terms "ISO-less" / "ISO-full", and recall suggesting the alternate use of those particular terms in this post replying to "gollywop":
... when he was in the course of editing his to be published DPReview article here:
... and which he continued to use in his subsequent DPReview published article here:
The more information regarding these subjects that is made accessible to readers within articles published on the DPReview site (in addition to appearing in forum posts), the better !
Upstream/downstream demarcation based upon what we (outside of "black boxes") are able to measure makes sense in cases of separable system-blocks. The existence of multiple on-sensor parallel-configured system sub-blocks does not preclude the comparison of various conceptual models to measurement results.
These noise-sources are most relevant in cases where "photon shot noise" is not a dominant (input-referred) noise source.
If, "downstream read noise includes all sources from and after" ... " ISO amplification", the "demarcation" exists at photosite source-follower outputs - with multiplexed switched-capacitor variable-gain devices and S/Hs preceeding ADC device inputs "downstream".
It appears that DxO SNR measurements include periodic (as well as random) noise components. DxO "DR" and derived "Read Noise" estimations are extrapolations from those measurements.
(Unless periodic can be separated from random),"classic" noise-models may tend towards becoming less tenable
While the authors of this article have done a good job introducing concepts of "upstream" (Footnote *1) and "downstream" (Footnote *2) for the purposes of introducing some "working conceptual models" within this particular article, ...
... it seems to me that many recent image-sensor design architectures accomplish (non digital scaling implemented) "amplification" of photosite source-follower stage outputs using switched-capacitor variable feedback impedance amplification stages, and utilize multiple column-parallel analog/digital conversion devices - all of which reside on the image-sensor LSI Circuitry itself ...
... thus (IMO) tending towards a "blurring of the determinable lines" demarcating where within the system architecture "upstream" ends and "downstream" begins.
A "Low/High Downstream Read Noise" dichotomy provokes the question, "relative to what?". These days such sub-system differentiations are becoming harder to determine with certainty (thus becoming more "murky")
Richard Butler and Rishi Sanyal:
(35mm equivalent) FL=24mm, F=5.6, Ts=1/40, ISO=100;
DxO-7.23, Mpixels=5.626, (16-bit) Lanczos-3 downsized followed by mild USM.
Dear Lady of the Creek
In secret places of sanctity, through time and tides, aeons pass within moments, like droplets of sweet rainslowly nurturing perennial swaths of emerald green mosses gently wrapping around Nature's primeval, delicate frame,eventually finding their way back to the ancient Salish Sea.While thoughtfully perched upon primeval, stoic stonesembraced and softly caressed by the lady of the creek,dark nights of the soul immersed within tempestuous seas wane with dawn's graceful glow; remembrances of eternity amidst moments timeless, within sanctuaries sublime.Through such places of quiet sanctity I often wanderamidst a sliver of ancient evergreen forest abounding with lush mosses, ferns, plethoric and ubiquitous floragently embracing perennial springs and sacred streamsslowly finding their way to ponds, lakes, and the Sea.
Those who came before us here for some 12,000 years held similar metaphors for river, mouth, and language.Stoic stones amidst the creeks still make water speak of aeons before, of Nature's gravity unheard amidsttoday's concrete jungle; beyond ontological vanity.Oft while lightly perched upon mossy, stoic stones amidst the natural wonder of yet unspoiled streams, a little bird protean wistfully daydreams of visitingsecret lush and green places of sanctity,glimpses of Nature's primeval and delicate frame,timeless beauty in perennial swaths of emerald moss,ancient flora flourishing amidst lush and muddy bogs,rocks gently caressed by the dear lady of the creekhaving, merely a few blinks before in the eye of time, nurtured what may have been the earth's tallest trees.
The cool, damp soil abounds with reverberations of time,signs of those who came before for some 12,000 years;human hearts, hopes, struggles, standings, and fears, trails smoothed by many feet, many unrecognized tears fallen into eternal springs joining the lady of the creek.Following this conscious stream as she winds by gravity, one part above the ground, with three parts as well below, she finds her riverine essence, gathering her timeless stones of ancient origin so long ago in the taciturn folds of eternity,where only the stillness confides that which was once to be.
The subject-matter outside of the cropped image-frame seen was prohibitive ...
John Kay: "Neither does DxO Optics Pro 8 allow you to highlight a slider's corresponding value box to adjust the numbers via your keyboard."
I think you'll find that you can! I get a blue outline on the value box when I click on it, and can then type a value in. BTW, that's on a Mac.
Thanks for the article. What it shows is that the software developers have now narrowed their differences substantially in the image editing results, but still have different approaches to management, etc. Wonderful for the consumer!
That has always worked just fine on all Windows versions of DxO Optics Pro (6.1 - 7.23) when installed on WinXP Pro. It seems unlikely that DxO 8.x would drop that functionality ?
Also - when the user single-left-clicks on (any of the) control-sliders themselves (at the location of where the control's "cursor" is set), the keyboard's Up/Down Arrow keys will advance or decrement that particular control setting-value, incrementing by one gradation-unit, as opposed to the super-irritating individual steps of 5 gradation-units that Lightroom 3.x insists upon stepping-by (in all cases except where it comes to the "Black Level" control-slider).
Detail Man: While it is true that it is necessary to purchase the DxO Optics Pro Elite 8.x version (normally priced at $299 USD) in order to ensure compatibility with some relatively higher-priced dSLRs, DxO Optics Pro 8.x Standard version (normally priced at $169 USD) can suffice in all other cases - where it is, as a result, much more competitively priced in comparison with Lightroom 4.x's $149 USD costs.
<< << Hello Peter,
at DxO we also offer special upgrade price to existing customer. ...
The above statement relating to the availability of upgrade licenses for DxO Optics Pro could only be interpreted to be accurate if the word "special" is construed to have been intended to mean:
"possibly for certain time periods if and when DxO Labs may so choose" (with reference to the existence of an upgrade license price that is reduced in cost relative to the full license price in effect at any given time).
... as opposed to being construed to have been intended to mean:
"special" (with reference to the monetary cost of an upgrade license that is reduced in cost relative to the full license price in effect at any given time).
Thus, it appears that customers who have purchased licenses for previous release-versions of DxO Optics Pro are "special" to DxO Labs only if and when DxO Labs may at times so choose.
<< Hello Peter,
at DxO we also offer special upgrade price to existing customer.
Concerning the price of the upgrade on DxO Optics Pro 8 :From version "x" (any version) to v8 the upgrade price was as low as 49USD (Standard edition) and 69 USD (Elite edition) during the launch.Those prices were only available for those who have purchased a previous version of course. (only online via the DxO customer account)
This is to inform DxO Labs that the statement in the first paragraph (above) is only partially accurate, and past-tense references of the second paragraph (above) are not accurate.
Since the Release of Version 8.0 on Oct 30, 2012, DxO Labs emails sent to members of the DxO Labs mailing-list show that there was no notification of any upgrade license availability whatsoever, until ...
... four emails (dated 22 and 27 Dec, 2102, and 4 and 16 Jan, 2013) advertise $49 USD (Standard) and $69 USD (Elite) upgrade licenses still being offered through 31 Jan, 2013.
Detail Man: Regarding DxO Optics Pro:
In the case of Panasonic, quality of rectilinear distortion correction appears to be notably higher than Panasonic correction meta-data utilized by applications such as Adobe LR/CR. Presumably characterized at more Focal Lengths.
"Lens Softness" correction default settings were changed (in V 6.x) to "V2" (Global=-0.5, Detail=50). Have found that these settings can overdo the sharpening effects. I typically use the original "V1" (Global=0.0, Detail=0).
I find the auto NR controls' settings can be (uniformly) reduced by factors of between 2 and 5.
Adobe LR/CR Color NR is more effective than DxO Chrominance NR. DxO Luminance NR (particularly at settings <= 10) appears to result in less detail-smearing than LR/CR Luminance NR. This article speculates as to why that may be so.
See Section (1) here:
Almost all of my applications are nature/landscape - and I am the type who (with any RAW processor) prefers to process each image individually using manual controls (Versions 6.x followed by Version 7.23 using the "DxO Lighting" interface).
It is not clear to me whether I would prefer the new ("just like Lightroom") tone-control-sliders. I don't tends to let much in the way of detailed subject-matter "clip" when shooting RAW.
Lightroom 4.x's "highlight recovery" is less than impressive for things other than wispy clouds - and it sounds like DxO's offering is not better.
Then there is the question of how well the "Lens Softness" corrections function in conjunction with "highlight recovery"?
"Auto-brain" controls are not for me. I like the use of the Gamma control (in "DxO Lighting" interface) combined with Exposure Compensation - though that (pre-versions 8.x) interface can tend to overdo the shadow-tones. That "DxO Lighting" interface is available as an option in versions 8.x.
My experience is primarily with the Panasonic DMC-LX3, and LGV 14-45mm and LGV 7-14mm lenses on a DMC-GH2 body. Sorry, no Nikon camera/lens experiences to recount.
It is possible that other manufacturers provide more rectilinear distortion correction image-file meta-data pertaining to more individual ranges of Focal Length. Don't know.
I have a feeling the the "Lens Softness" setting characteristics (may) well apply in general relative to different manufacturers.
The business about (in the case of Panasonic cameras) being able to reduce the NR-control settings below what is automatically selected (may) apply (to some extent) in general - as DxO Labs states that they characterize the image-noise spectrums for each camera model, and adjust the (internal) weighting of their NR controls accordingly.
OK. Thanks for that specific information. I guess that DxO Labs figures that "full-frame > full price".
Now that DxO Labs writes the essential core of the DxO Optical Correction Modules into the software application itself, rather than into the downloadable Modules themselves (ever since the release of Version 7.50) ...
... that means that roughly once per year (as in the case of Versions 8.x replacing Versions 7.x), if you want support for that newly added camera body or lens, the customer will be forced to purchase (Version 9.x, etc.) all over again.
$299 USD per year could get rather expensive, indeed ...
Jolly Oly: Fantastic read, thank you very much.
Having both Lr4 and DxO 7.5, and sometimes also using Olympus Viewer2 (for the OMD files) one thing is not clear and I would appreciate if someone could explain it to me.
Why the same WB color temperature value will result in more or less different colors in all three on them ?
There is a lot more than the de-mosacing algorithm used that influences color rendering.
While image-file meta-data contains color matrix coefficients (not absolute color temperature and tint data), different RAW processors use differing color matrices applied to the de-mosaiced RGB image-data. The latitude taken is significant - and does not lend itself to consistency in absolute color temperature comparisons.
A given indicated color temperature in Lightroom 3.x yields an image (viewed on a sRGB-calibrated monitor) that is appears to the eyes as being approximately 500 Degrees Kelvin "cooler" than DxO Optics Pro Versions 6.x and 7.x.
In example, 5000 Degrees Kelvin in DxO (6.x and 7.x) looks about the same as 5500 Degrees Kelvin in Lightroom 3.x (for sunlit scenes).
In addition to the numerical value of WB Color Temperature, the amount of applied WB Tint alters chromaticity as well.
Most "As Shot" WB co-ordinates include both Temperature and Tint; these are not simple matters.
While it is true that it is necessary to purchase the DxO Optics Pro Elite 8.x version (normally priced at $299 USD) in order to ensure compatibility with some relatively higher-priced dSLRs, DxO Optics Pro 8.x Standard version (normally priced at $169 USD) can suffice in all other cases - where it is, as a result, much more competitively priced in comparison with Lightroom 4.x's $149 USD costs.
Neodp: Shoot Raw
Pull the embedded JPEG out, with dcraw ("draw -e *", and setup in your chosen file manger; as a right click option.) This is not developing, and is instant, and camera dependent(size).
Use the JPEG(s), where sufficient to your own goals. AKA, you got the settings right, in the camera, and controlled the light where you could.
For time purposes: For those fewer Raw shoots; that would benefit, from 16 bit light adjustments (including color), then just click the Raw file, that you preset, to bring up "ufraw".
Learn what the ufraw (free, and cross-platform) settings do, from it's web page. You do not have to adjust all of them; but only the few, that matter, and to effect the change you are after. With ufraw, you just download, or make your camera's profile, first, and then ufraw produces a default photo; matching your cameras JPEG, if that's what you want. It mostly finished developing; upon loading. You can match any style, this way. Do the rest in gimp, and it's plug-ins.
UFRaw 0.18 is positively "ancient" at this point, and uses older DCRaw binaries which do not support any camera released in what is now (nearly) 2 years time.
20/02/2011 - UFRaw-0.18 released, based on DCRaw v 9.06:
Further, it will not function with versions of GIMP later than 2.611 on WinXP OS. I uninstalled UFRaw 0.18 (and the 8-bit only GIMP 2.80) completely as a result of this unfortunate situation.
Regarding DxO Optics Pro:
Very nice ! Great capture, GW !
DxOReview ... featuring controlled-exposure, pixel-resolution-normalized, and non-Adobe-massaged Studio Comparison Tool RAW image-samples for the purpose of readers actually being able to make reliable comparisons between RAW-recording cameras !