DPreview K-01 RAW high ISO samples botched!

JohnBee · 05-03-2012, 04:43 AM

Sorry to be the barer of bad news but the DPreview RAW Studio scene comparison samples for the K-01 have NR applied to them

I just downloaded the ISO6400 RAW image file from the Studio scene and found that all of the noise was present and accounted for. Whereras the scene(on the other hand) was virtually noise free.

This seemingly surreal noise performance has stirred quite a few threads where people have been pointing-out that the K-01 was somehow surpassing most FF camera's. Which sort of made me look

Anyways, when all is said and done, I'm afraid there won't be any low level NR miracles this round from Pentax. But... hey... you gotta admit. This even had the RAW purist guys biting their tongues.

Pentaxor · 05-03-2012, 06:01 AM

Originally posted by JohnBee

Sorry to be the barer of bad news but the DPreview RAW Studio scene comparison samples for the K-01 have NR applied to them

I just downloaded the ISO6400 RAW image file from the Studio scene and found that all of the noise was present and accounted for. Whereras the scene(on the other hand) was virtually noise free.

This seemingly surreal noise performance has stirred quite a few threads where people have been pointing-out that the K-01 was somehow surpassing most FF camera's. Which sort of made me look

Anyways, when all is said and done, I'm afraid there won't be any low level NR miracles this round from Pentax. But... hey... you gotta admit. This even had the RAW purist guys biting their tongues.

not really surprised JB. as I mentioned the possibility of RAW NR application. the sharpening pattern is somewhat a gimme that something was done with the RAW. I didn't say that it has superior RAW but better RAW NR.

ScooterMaxi Jim · 05-03-2012, 12:23 PM

They still look very good after the corrected posting. Very low chroma noise at high ISO, and good detail retained. On the down side, the saturation and contrast are lower than the K-5. The red channel suppression looks better than the K-5, but that could be mostly a matter of the difference in vibrance.

To my eye, it is pretty clear cut that the IQ is higher than the NEX-5n and NEX-7 RAW samples, but then I'm a big fan of realistic (subdued) rendering.

JNR

regor · 05-03-2012, 03:28 PM

Originally posted by JohnBee

...the ISO6400 RAW image file from the Studio scene and found that all of the noise was present and accounted for. Whereras the scene(on the other hand) was virtually noise free.

The "scene" cannot be "seen" unless the RAW is PP'd. Your processing may differ from DPR processing, what matters for comparative purposes is that DPR Processing is consistent from one camera to another, and this is what they've corrected I believe. I just looked at the "scene" and to me (on a completely different computer, so I am not looking at the cached image here), it is still outstanding, and easily measure up to FF Nikon D800, which in itself is rather surprising. What I also see is that the nature of noise (mostly luminance) is easily removed without affecting details, whereas color noise (which is extremely well controlled) cannot be removed without loosing details; hence the RAW K-01 images can be PP'd to virtually remove all noise with little or no loss of details; THAT is what I think is most significant.

Last edited by regor; 05-03-2012 at 03:42 PM.

JohnBee · 05-03-2012, 03:44 PM

Originally posted by regor

The "scene" cannot be "seen" unless the RAW is PP'd.

Actually what happened(I'm almost sure of it) is that they(DPReview) put-up RAW images with NR applied to them whereas they wouldn't do this normally since part of the assessment is to compare noise performances between units(respectively).

Anyways, I didn't look very deeply into this, but it looks like they used ACR to create the images at which point, it's likely that they allowed ACR to run with default NR enabled, which I think is evaluative(based on noise readings) or not... I'm not entirely sure about that.

Whatever the case they've since fixed the issue and the K-01 is showing RAW noise as it should be.
Unfortunately, I didn't save any of the original screens, though I will say they were pretty outstanding against the competition. Enough to say that even the D800 people were on the defensive

**monochrome** · 05-03-2012, 04:40 PM

Originally posted by JohnBee

Enough to say that even the D800 people were on the defensive

Suppose Pentax works this magic (they seem to be very good at extracting clean RAW files from the Sony CMOS sensor(s) with whatever the next cameras are? Imagine if the K-01 eventually overcomes the negative initial response by showing itself to be an IQ beast - what will DPR do then? What if the reason they dropped PEF is because Adobe DNG is just better - will we shout huzzah?

Pentaxor · 05-03-2012, 05:06 PM

Originally posted by JohnBee

Actually what happened(I'm almost sure of it) is that they(DPReview) put-up RAW images with NR applied to them whereas they wouldn't do this normally since part of the assessment is to compare noise performances between units(respectively).

Anyways, I didn't look very deeply into this, but it looks like they used ACR to create the images at which point, it's likely that they allowed ACR to run with default NR enabled, which I think is evaluative(based on noise readings) or not... I'm not entirely sure about that.

Whatever the case they've since fixed the issue and the K-01 is showing RAW noise as it should be.
Unfortunately, I didn't save any of the original screens, though I will say they were pretty outstanding against the competition. Enough to say that even the D800 people were on the defensive

35MP sensor versus 16MP, they should stop complaining and be happy with the D800.

anyway, it became obvious that there is some sort K-01 noise pattern. this maybe distracting especially once NR is applied. I don't worry much about detail and color lost since it seems that color is not that compromised as well.

gazonk · 05-04-2012, 04:26 AM

Originally posted by ScooterMaxi Jim

They still look very good after the corrected posting. Very low chroma noise at high ISO, and good detail retained. On the down side, the saturation and contrast are lower than the K-5. The red channel suppression looks better than the K-5, but that could be mostly a matter of the difference in vibrance.

To my eye, it is pretty clear cut that the IQ is higher than the NEX-5n and NEX-7 RAW samples, but then I'm a big fan of realistic (subdued) rendering.

JNR

Looks better than the K-5 to me! Apparently, the K-01 has a weaker anti alias filter, since moire is seen in the striped paper to the left of the Bailey bottle. So K-01 images look slightly sharper. And even after the correction the high iso images are less noisy than the K-5 images. Wow! Can't wait for the K-5n or whatever its name is...

Regarding differences in PP of RAW images: When comparing with the K-01, IMHO the NEX-5n RAW images look over-sharpened to me. They may look slightly "crisper" when observing letters etc., but I see several cases of what looks like sharping artifacts to me.

JohnBee · 05-04-2012, 05:08 AM

Originally posted by gazonk

Looks better than the K-5 to me! Apparently, the K-01 has a weaker anti alias filter, since moire is seen in the striped paper to the left of the Bailey bottle. So K-01 images look slightly sharper. And even after the correction the high iso images are less noisy than the K-5 images. Wow! Can't wait for the K-5n or whatever its name is...

Regarding differences in PP of RAW images: When comparing with the K-01, IMHO the NEX-5n RAW images look over-sharpened to me. They may look slightly "crisper" when observing letters etc., but I see several cases of what looks like sharping artifacts to me.

I think you're right. I scoured the scene at 200% and found the K-01 to resolve detail that the K-5 just didn't have. However, I noticed the K-5 was shot at f/9 whereas the K-01 was at f/8. And after looking at MFT ratings, it looks like the D FA 50/2.8 falls-off past 5.6 and so there might be a little of that coming through in the scene. Having said that, I'm really hoping Pentax is leaning toward a weaker(perhaps improved) AA filter this round as that could spell better per pixel sharpness of the upcoming K-3 also.

Now if only they can give us an NR: ON/OFF feature in RAW, that would really put the K-5 successor in a class on its own imo.

PS. No offense Pentax, I think your low level NR Is splendid, but I prefer to have the choice with NR in RAW rather than you choosing for me. :ugh:

05-03-2012, 04:43 AM	#1
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667	DPreview K-01 RAW high ISO samples botched! Sorry to be the barer of bad news but the DPreview RAW Studio scene comparison samples for the K-01 have NR applied to them I just downloaded the ISO6400 RAW image file from the Studio scene and found that all of the noise was present and accounted for. Whereras the scene(on the other hand) was virtually noise free. This seemingly surreal noise performance has stirred quite a few threads where people have been pointing-out that the K-01 was somehow surpassing most FF camera's. Which sort of made me look Anyways, when all is said and done, I'm afraid there won't be any low level NR miracles this round from Pentax. But... hey... you gotta admit. This even had the RAW purist guys biting their tongues.

05-03-2012, 12:23 PM	#3
ScooterMaxi Jim Veteran Member Join Date: Jul 2009 Location: Minnesota Posts: 1,520	DPreview K-01 RAW high ISO samples fixed now They still look very good after the corrected posting. Very low chroma noise at high ISO, and good detail retained. On the down side, the saturation and contrast are lower than the K-5. The red channel suppression looks better than the K-5, but that could be mostly a matter of the difference in vibrance. To my eye, it is pretty clear cut that the IQ is higher than the NEX-5n and NEX-7 RAW samples, but then I'm a big fan of realistic (subdued) rendering. JNR

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
K-5 Low light/ High ISO dpreview vs dxo mark conflict	vodanh1982	Pentax K-5 & K-5 II	5	10-07-2011 02:52 AM
K-5 low light/High ISO samples	johnmflores	Pentax K-5 & K-5 II	21	04-19-2011 05:03 PM
Nikon D7000 High ISO Samples	vancmann	Canon, Nikon, Sony, and Other Camera Brands	48	11-05-2010 06:38 PM
Dedicated Pentax K5 High ISO samples Thread	vancmann	Pentax K-5 & K-5 II	2	11-03-2010 09:25 AM
Some K-r high ISO JPEG samples	Asahiflex	Pentax News and Rumors	34	10-11-2010 04:29 PM

05-03-2012, 06:01 AM	#2
Pentaxor Veteran Member Join Date: May 2009 Location: Vancouver, B.C. Photos: Gallery Posts: 6,513	Originally posted by JohnBee Sorry to be the barer of bad news but the DPreview RAW Studio scene comparison samples for the K-01 have NR applied to them I just downloaded the ISO6400 RAW image file from the Studio scene and found that all of the noise was present and accounted for. Whereras the scene(on the other hand) was virtually noise free. This seemingly surreal noise performance has stirred quite a few threads where people have been pointing-out that the K-01 was somehow surpassing most FF camera's. Which sort of made me look Anyways, when all is said and done, I'm afraid there won't be any low level NR miracles this round from Pentax. But... hey... you gotta admit. This even had the RAW purist guys biting their tongues. not really surprised JB. as I mentioned the possibility of RAW NR application. the sharpening pattern is somewhat a gimme that something was done with the RAW. I didn't say that it has superior RAW but better RAW NR.

05-03-2012, 03:28 PM	#4
regor Veteran Member Join Date: Jan 2009 Location: Victoria, BC Photos: Gallery \| Albums Posts: 356	Originally posted by JohnBee ...the ISO6400 RAW image file from the Studio scene and found that all of the noise was present and accounted for. Whereras the scene(on the other hand) was virtually noise free. The "scene" cannot be "seen" unless the RAW is PP'd. Your processing may differ from DPR processing, what matters for comparative purposes is that DPR Processing is consistent from one camera to another, and this is what they've corrected I believe. I just looked at the "scene" and to me (on a completely different computer, so I am not looking at the cached image here), it is still outstanding, and easily measure up to FF Nikon D800, which in itself is rather surprising. What I also see is that the nature of noise (mostly luminance) is easily removed without affecting details, whereas color noise (which is extremely well controlled) cannot be removed without loosing details; hence the RAW K-01 images can be PP'd to virtually remove all noise with little or no loss of details; THAT is what I think is most significant. Last edited by regor; 05-03-2012 at 03:42 PM.

05-03-2012, 03:44 PM	#5
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	Originally posted by regor The "scene" cannot be "seen" unless the RAW is PP'd. Actually what happened(I'm almost sure of it) is that they(DPReview) put-up RAW images with NR applied to them whereas they wouldn't do this normally since part of the assessment is to compare noise performances between units(respectively). Anyways, I didn't look very deeply into this, but it looks like they used ACR to create the images at which point, it's likely that they allowed ACR to run with default NR enabled, which I think is evaluative(based on noise readings) or not... I'm not entirely sure about that. Whatever the case they've since fixed the issue and the K-01 is showing RAW noise as it should be. Unfortunately, I didn't save any of the original screens, though I will say they were pretty outstanding against the competition. Enough to say that even the D800 people were on the defensive

05-03-2012, 04:40 PM	#6
monochrome Loyal Site Supporter Join Date: Mar 2008 Location: Working From Home Photos: Gallery \| Albums Posts: 26,276	Originally posted by JohnBee Enough to say that even the D800 people were on the defensive Suppose Pentax works this magic (they seem to be very good at extracting clean RAW files from the Sony CMOS sensor(s) with whatever the next cameras are? Imagine if the K-01 eventually overcomes the negative initial response by showing itself to be an IQ beast - what will DPR do then? What if the reason they dropped PEF is because Adobe DNG is just better - will we shout huzzah?

05-03-2012, 05:06 PM	#7
Pentaxor Veteran Member Join Date: May 2009 Location: Vancouver, B.C. Photos: Gallery Posts: 6,513	Originally posted by JohnBee Actually what happened(I'm almost sure of it) is that they(DPReview) put-up RAW images with NR applied to them whereas they wouldn't do this normally since part of the assessment is to compare noise performances between units(respectively). Anyways, I didn't look very deeply into this, but it looks like they used ACR to create the images at which point, it's likely that they allowed ACR to run with default NR enabled, which I think is evaluative(based on noise readings) or not... I'm not entirely sure about that. Whatever the case they've since fixed the issue and the K-01 is showing RAW noise as it should be. Unfortunately, I didn't save any of the original screens, though I will say they were pretty outstanding against the competition. Enough to say that even the D800 people were on the defensive 35MP sensor versus 16MP, they should stop complaining and be happy with the D800. anyway, it became obvious that there is some sort K-01 noise pattern. this maybe distracting especially once NR is applied. I don't worry much about detail and color lost since it seems that color is not that compromised as well.

05-04-2012, 04:26 AM	#8
gazonk Pentaxian Join Date: Feb 2009 Location: Oslo area, Norway Photos: Gallery \| Albums Posts: 3,746	Originally posted by ScooterMaxi Jim They still look very good after the corrected posting. Very low chroma noise at high ISO, and good detail retained. On the down side, the saturation and contrast are lower than the K-5. The red channel suppression looks better than the K-5, but that could be mostly a matter of the difference in vibrance. To my eye, it is pretty clear cut that the IQ is higher than the NEX-5n and NEX-7 RAW samples, but then I'm a big fan of realistic (subdued) rendering. JNR Looks better than the K-5 to me! Apparently, the K-01 has a weaker anti alias filter, since moire is seen in the striped paper to the left of the Bailey bottle. So K-01 images look slightly sharper. And even after the correction the high iso images are less noisy than the K-5 images. Wow! Can't wait for the K-5n or whatever its name is... Regarding differences in PP of RAW images: When comparing with the K-01, IMHO the NEX-5n RAW images look over-sharpened to me. They may look slightly "crisper" when observing letters etc., but I see several cases of what looks like sharping artifacts to me.

05-04-2012, 05:08 AM	#9
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	Originally posted by gazonk Looks better than the K-5 to me! Apparently, the K-01 has a weaker anti alias filter, since moire is seen in the striped paper to the left of the Bailey bottle. So K-01 images look slightly sharper. And even after the correction the high iso images are less noisy than the K-5 images. Wow! Can't wait for the K-5n or whatever its name is... Regarding differences in PP of RAW images: When comparing with the K-01, IMHO the NEX-5n RAW images look over-sharpened to me. They may look slightly "crisper" when observing letters etc., but I see several cases of what looks like sharping artifacts to me. I think you're right. I scoured the scene at 200% and found the K-01 to resolve detail that the K-5 just didn't have. However, I noticed the K-5 was shot at f/9 whereas the K-01 was at f/8. And after looking at MFT ratings, it looks like the D FA 50/2.8 falls-off past 5.6 and so there might be a little of that coming through in the scene. Having said that, I'm really hoping Pentax is leaning toward a weaker(perhaps improved) AA filter this round as that could spell better per pixel sharpness of the upcoming K-3 also. Now if only they can give us an NR: ON/OFF feature in RAW, that would really put the K-5 successor in a class on its own imo. PS. No offense Pentax, I think your low level NR Is splendid, but I prefer to have the choice with NR in RAW rather than you choosing for me. :ugh: