Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
04-23-2018, 02:21 PM   #211
Pentaxian




Join Date: May 2015
Posts: 1,267
It is surprising that the accelerator kicks in as early as iso 800 considering the quality of the sensor. Anyone looking should be able to see evidence of noise reduction in the samples provided. Cooked raws are increasingly common it seems. There must be something driving this development. Review sites or users? Who knows?

I guess the manufacturer have better data on sensor noise and are thus able to produce better denoise defaults. However this is only the case because the manufacturers fail to share the data. And the downside is that you can't turn it off when you prefer noise to avoid loosing detail. It's an unfortunate way to proceed but most camera brands seem to have adopted it.

04-23-2018, 02:22 PM - 2 Likes   #212
Veteran Member




Join Date: Feb 2015
Photos: Gallery | Albums
Posts: 8,004
is it just me or ? i find the photographs from my K1 , well exposed, much more compelling than 100% crops of underexposed test shots in bad light. Hopefully K1 II is at least as good as K1 mk1

Last edited by biz-engineer; 04-25-2018 at 11:34 AM.
04-23-2018, 02:24 PM - 1 Like   #213
Pentaxian




Join Date: Apr 2007
Location: Romania
Posts: 13,517
QuoteOriginally posted by house Quote
Anyone looking should be able to see evidence of noise reduction in the samples provided.
I looked, hard. Then I looked again. There is no evidence of compulsory noise reduction in the samples provided.

Get the RAWs, and make sure your RAW developing software doesn't apply any NR by itself. Defaults varies with ISO.
04-23-2018, 02:35 PM   #214
Junior Member
aikaarska's Avatar

Join Date: Sep 2016
Location: Finland
Posts: 44
QuoteOriginally posted by house Quote
It is surprising that the accelerator kicks in as early as iso 800 considering the quality of the sensor. Anyone looking should be able to see evidence of noise reduction in the samples provided. Cooked raws are increasingly common it seems. There must be something driving this development. Review sites or users? Who knows?

I guess the manufacturer have better data on sensor noise and are thus able to produce better denoise defaults. However this is only the case because the manufacturers fail to share the data. And the downside is that you can't turn it off when you prefer noise to avoid loosing detail. It's an unfortunate way to proceed but most camera brands seem to have adopted it.
+1

Indeed. It's an annoying trend. I understand that if inexpensive APS-C amateur camera (<1000$), but not with flagship models.

04-23-2018, 02:38 PM   #215
Loyal Site Supporter
Loyal Site Supporter
BigMackCam's Avatar

Join Date: Mar 2010
Location: North-East of England
Posts: 15,936
QuoteOriginally posted by house Quote
It is surprising that the accelerator kicks in as early as iso 800 considering the quality of the sensor. Anyone looking should be able to see evidence of noise reduction in the samples provided.
Absolutely, there's noise reduction. That's no surprise - we know the sensor is the same, hence the improved noise performance has to be a result of noise reduction, albeit applied at an earlier point in the architecture.

What we should be interested in is whether the noise reduction carried out by the image accelerator results in better images overall (which involves some give and take) than we can produce ourselves with software. The OP suggested that it might be possible to get better results ourselves... I'm not yet convinced of that (but I remain open-minded).

QuoteOriginally posted by house Quote
Cooked raws are increasingly common it seems. There must be something driving this development. Review sites or users? Who knows?
Cooked RAW files have been with us for years. My K-5, K-3 and K-3II had them... I assume other models too. Very much the same for the other brands. And then, of course, there's the cooking by RAW development software at the time of import. What you see is most definitely not what was imported... there's a lot of default processing going on, and much of it (demosacing algorithms, for example) is different between the various applications. So the concept of genuinely "raw" RAW hasn't been around for some time, if ever, IMHO.
04-23-2018, 02:41 PM   #216
Junior Member
aikaarska's Avatar

Join Date: Sep 2016
Location: Finland
Posts: 44
QuoteOriginally posted by Kunzite Quote
No more theories, please. I just debunked the "smoothing" one, I don't need another.
That in-camera algorithm seems apply NR, but also a bit sharpening - when ISO800 or higher. ...Anyway, I would like desperately to see the corresponding K-1 vs. K-1M2 comparison!
04-23-2018, 03:04 PM   #217
Pentaxian




Join Date: May 2015
Posts: 1,267
QuoteOriginally posted by BigMackCam Quote
What you see is most definitely not what was imported... there's a lot of default processing going on, and much of it (demosacing algorithms, for example) is different between the various applications. So the concept of genuinely "raw" RAW hasn't been around for some time, if ever, IMHO.
This all depends on what software you use. With Rawtherapee for instance all those steps are transparent including demosaic algo. Of course raw data is of no use as an image. 0100100 etc isn't all that communicative but being able to choose your pipeline is.
04-23-2018, 03:10 PM - 1 Like   #218
Loyal Site Supporter
Loyal Site Supporter
BigMackCam's Avatar

Join Date: Mar 2010
Location: North-East of England
Posts: 15,936
QuoteOriginally posted by house Quote
This all depends on what software you use. With Rawtherapee for instance all those steps are transparent including demosaic algo. Of course raw data is of no use as an image. 0100100 etc isn't all that communicative but being able to choose your pipeline is.
Agreed

As well as Lightroom, I use Darktable 2.4.2. It's very transparent and selectable, very easy to switch everything off. But here's a for-instance - even in Darktable and RawTherapee, you have to select a demosacing algorithm. There is no such thing in any RAW development software (RawTherapee included - and I have used it, occasionally still do for testing) as an import that doesn't involve a demosaicing algorithm (it has to use something, right?). And you've obviously seen the significant difference that such algorithms can have on an imported image. Then there's the base tone curve, or - if no base curve is applied - the software's interpretation and rendering of the data to whatever the developers believe is a neutral tone curve.

I could go on... but I don't think that's necessary

My point is, no-one's probably seen "RAW" for real... because every rendering of a RAW file, by any piece of software, involves the developer's interpretations...

04-23-2018, 03:13 PM   #219
Pentaxian




Join Date: Apr 2007
Location: Romania
Posts: 13,517
QuoteOriginally posted by BigMackCam Quote
Absolutely, there's noise reduction. That's no surprise - we know the sensor is the same, hence the improved noise performance has to be a result of noise reduction, albeit applied at an earlier point in the architecture.
Sorry, I was imprecise in my previous posts - of course there is a lot of effort put into noise suppression, from attempting to avoid generating noise to attempting to detect and remove noise. As I said, we probably wouldn't like a "raw RAW" at all.
What there isn't - according to the samples, and contrary to the OP's claim - is a compulsory, detail-destroying RAW NR.

---------- Post added 24-04-18 at 01:15 AM ----------

QuoteOriginally posted by aikaarska Quote
That in-camera algorithm seems apply NR, but also a bit sharpening - when ISO800 or higher.
No, it doesn't. Not the kind of NR you're talking about.
I won't go into sharpening - those images have sharpening applied, by the way. Some sharpening is even applied at demosaicing stage.

L.E. Obviously, I forgot to set demosaicing sharpness to the same value in samples provided... yep, slightly higher for ISO 100. But I'm done with this comparison.

Last edited by Kunzite; 04-23-2018 at 03:22 PM.
04-23-2018, 03:25 PM   #220
Junior Member
aikaarska's Avatar

Join Date: Sep 2016
Location: Finland
Posts: 44
QuoteOriginally posted by Kunzite Quote
. But I'm done with this comparison.
...Good for you, I guess. But it doesn't solve the original claim, unfortunately.
04-23-2018, 03:29 PM - 1 Like   #221
Loyal Site Supporter
Loyal Site Supporter
BigMackCam's Avatar

Join Date: Mar 2010
Location: North-East of England
Posts: 15,936
QuoteOriginally posted by aikaarska Quote
...Good for you, I guess. But it doesn't solve the original claim, unfortunately.
For the sake of balance though, nothing supports that claim yet, either. More evidence, in the forum of well-controlled K-1 vs K-1II examples, is needed before any of us can assess and analyse what the differences are, and what they mean to a wide group of potential users...
04-23-2018, 03:33 PM   #222
Moderator
Not a Number's Avatar

Join Date: Mar 2012
Location: Venice, CA
Posts: 7,671
Is it sharpening or more detail? If the accelerator is able to better discriminate signal from noise you can get more detail. Sharpening enhances or highlights edges or fine detail to make it easier seen by the eye.
04-23-2018, 03:37 PM   #223
Pentaxian




Join Date: Apr 2007
Location: Romania
Posts: 13,517
QuoteOriginally posted by aikaarska Quote
...Good for you, I guess. But it doesn't solve the original claim, unfortunately.
It's solved, done and now should be buried - it was a false alarm.
04-23-2018, 03:40 PM   #224
Junior Member
aikaarska's Avatar

Join Date: Sep 2016
Location: Finland
Posts: 44
QuoteOriginally posted by Kunzite Quote
It's solved, done and now should be buried - it was a false alarm.
Well - It remains to be seen...
04-23-2018, 04:42 PM   #225
Pentaxian




Join Date: Jun 2009
Posts: 11,913
QuoteOriginally posted by house Quote
Cooked raws are increasingly common it seems. There must be something driving this development.
Competition - eg Sony is doing it too, as are others: from the DPR's recent A7 III review:

QuoteQuote:
The updated sensor works in tandem with a new 'front-end LSI' (essentially an additional processor) to capture up to 10 frames per second with full autofocus and autoexposure, as well as allowing for oversampled (read: very detailed) 4K/24p capture

Despite Ricoh claiming the Accelerator chip is their own invention, I wouldn't be surprised if it was actually sourced from someone else like Socionext or even Sony.

QuoteOriginally posted by Kunzite Quote
There is no evidence of compulsory noise reduction in the samples provided.
I had a look and share this view.

But more testing is needed to map all of the contributions the Accelerator chip makes to camera RAWs.

I wish Bill Claff had been able to complete his testing on the K-P, but that effort seems to have died.

Last edited by rawr; 04-23-2018 at 04:48 PM.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
accelerator, camera, claims, crop, dslr, exposure, full frame, full-frame, image, information, iso800, iv, k-1, k-1 mk2, k1, mess, mk2, op, pentax k-1, sensor, settings, software, sony, unit
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Macro Best moderate priced macro lens for newbie HGMerrill Photographic Technique 18 10-20-2014 06:08 PM
Old Moderate Mitt is Back! boriscleto General Talk 3 10-10-2012 02:12 PM
"Moderate" Mitt jeffkrol General Talk 2 10-08-2012 01:59 PM
Excesive noise in moderate light, please help Al_s14 Pentax K-r 4 08-03-2011 03:28 PM
Pentax lens all-in-one with moderate zoom fevbusch Pentax DSLR Discussion 17 01-27-2007 05:13 PM



All times are GMT -7. The time now is 09:00 PM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top