DxO DeepPrime versus Prime, results and controls

CAcreeks · Nov 5, 2020

MaynardGKrebbs said:
For my part, I'm struggling to see how the new software is worth $80 an I'm not sure that I'll upgrade. Your mileage will, of course, vary depending what kind of camera and what you do with the photo (e.g., how big you print, crop, or view on screen).

To a large extent, I agree with you, although $80 is not much. The difference is mostly in detail rendered. In two cases the Deep PRIME images are larger, indicating that noise is similar but JPEG shows more detail. This is especially visible in text (e.g. CAUTION) in the third set.

MaynardGKrebbs said:
Nikon D750, ISO 51K, Underexposed by 4.66 stops, OldPrime, Default Settings

Nikon D750, ISO 51K, Underexposed by 4.66 stops, DeepPrime, Default Settings

IanYorke · Nov 5, 2020

MaynardGKrebbs said:
It may be my lack of discernment, but a lot of this discussion seems to overstate the difference between DeepPrime and Old Prime. Except for shots taken with small sensor compacts, I don't see much, if any difference, certainly not something that would show up in an 8x10 print. I can give some examples from my own photos. I had to reach a bit because most of my photos are of family and I didn't want to post people shots. So here goes:

I saw the least difference with normally exposed shots from my D750. Below is a shot taken at maximum ISO of 51K with OldPrime (first) and DeepPrime (second). I can definitely see some difference in the details, but, as I stated, I don't think it would be noticeable in an 8x10 print.

Even in the shot taken with my D7000 at its max ISO of 25K, the difference isn't any more dramatic. The only time that I saw a dramatic difference was with a shot that I took to test ISO independence and effect of bit depth and compression on my D750. This shot (below) was taken at ISO 51K and underexposed by 4.66 stops. (By the by, up underexposed by 3.66 stops, I saw virtually no difference between 12 bit and 14 bit, between lossy compression and lossless compression. The shot below is at 14 bit, lossy.)

For my part, I'm struggling to see how the new software is worth $80 an I'm not sure that I'll upgrade. Your mileage will, of course, vary depending what kind of camera and what you do with the photo (e.g., how big you print, crop, or view on screen).

Nikon D750, ISO51k, DxO Prime, Default settings

Nikon D750, ISO51k, DxO DeepPrime, Default settings

Nikon D7000, ISO 25K, OldPrime, Default Settigs

Nikon D7000, ISO 25K, DeepPrime, Default Settings

Nikon D750, ISO 51K, Underexposed by 4.66 stops, OldPrime, Default Settings

Nikon D750, ISO 51K, Underexposed by 4.66 stops, DeepPrime, Default Settings

In the DXO forums Beta testers said that they worked with DXO to fine tune the results from DeepPrime so that the default values gave similar results. It looks like you have confirmed the results of their work

Ian

Austinian · Nov 5, 2020

CAcreeks said:
MaynardGKrebbs said:

For my part, I'm struggling to see how the new software is worth $80 an I'm not sure that I'll upgrade. Your mileage will, of course, vary depending what kind of camera and what you do with the photo (e.g., how big you print, crop, or view on screen).

Click to expand...

To a large extent, I agree with you, although $80 is not much. The difference is mostly in detail rendered. In two cases the Deep PRIME images are larger, indicating that noise is similar but JPEG shows more detail.

The increased detail is also what I observed processing TIFs from an old 24Mp -3EV APS-C image from a HDR set.

NAwlins Contrarian · Nov 5, 2020

MaynardGKrebbs said:
It may be my lack of discernment, but a lot of this discussion seems to overstate the difference between DeepPrime and Old Prime. Except for shots taken with small sensor compacts, I don't see much, if any difference, certainly not something that would show up in an 8x10 print. I can give some examples from my own photos. I had to reach a bit because most of my photos are of family and I didn't want to post people shots. So here goes:

I saw the least difference with normally exposed shots from my D750. Below is a shot taken at maximum ISO of 51K with OldPrime (first) and DeepPrime (second). I can definitely see some difference in the details, but, as I stated, I don't think it would be noticeable in an 8x10 print.

I think that is more-or-less true, but the issue is much more general: the larger the sensor and the smaller the output size, the less noise is an issue, period. Across some range of the sensor size-output size relationship, even no noise reduction is just fine. Obviously where that range is depends on the gain.

So any noise reduction, more so with Prime, and even more so with DeepPrime, is more of a useful feature for the more problematic shots. And what is problematic is a function of sensor size, output size, gain, and sensor technology.

As previously raised, in my limited testing, where DeepPrime excels is not necessarily so much noise suppression as it is detail retention for a given amount of noise suppression.

Your example is using a relatively large (full-frame) sensor, and one of the better ones at that,* and then for printing downsampling about three sensor pixels to each print pixel.

This has long been true. With film, I found that 35mm Ilford Delta 3200 produced perfectly usable 4x6 inch prints, but by 8x10 inches they were often decidedly grain.

*E.g., DxO measurements (as interpolated) show the D750 falling below their criterion of a 30 dB signal-to-noise ratio at ISO 2956. The very highest such figure for any full-frame camera is ISO 3730 in the Sony A7 III. So the lowest-noise FF sensor, by DxO criteria / measurements, is only about 1/3 stop better than your D750.

MaynardGKrebbs · Nov 6, 2020

CAcreeks said:
MaynardGKrebbs said:

For my part, I'm struggling to see how the new software is worth $80 an I'm not sure that I'll upgrade. Your mileage will, of course, vary depending what kind of camera and what you do with the photo (e.g., how big you print, crop, or view on screen).

Click to expand...

To a large extent, I agree with you, although $80 is not much. The difference is mostly in detail rendered. In two cases the Deep PRIME images are larger, indicating that noise is similar but JPEG shows more detail. This is especially visible in text (e.g. CAUTION) in the third set.

I'm kinda old fashioned and still print and hang photos, so my threshold for significant difference is whether it would be noticeable in an 8x10 or 11x14 print viewed at normal viewing distance. For me (and I repeat that this is my own personal take and not meant to imply how anyone else should think) smaller differences are academic and that's how I judge the difference between the two NR options.

I agree with you that on the third set, there is a significant difference. However, that set was essentially taken at ISO 163K on a camera with max ISO of 51K. While I'm amazed at how good the shots are, and Deep PRIME does make a noticeable difference, I don't think that I would print it even if the subject matter was interesting. For the other two sets taken at max ISO on DX and FX cameras, to me, the difference falls below my threshold.

I also agree that $80 isn't much and I may upgrade to get the ability to select which adjustments get cut and pasted. A few years ago, I suggested DxO add the ability to cut and past just white balance. I also suggested that they add adjust white balance to match a single color between shots, something that would be really useful when photographing an event where the light varies across the space or changes.

Jacques Cornell · Nov 6, 2020

MaynardGKrebbs said:
It may be my lack of discernment, but a lot of this discussion seems to overstate the difference between DeepPrime and Old Prime. Except for shots taken with small sensor compacts, I don't see much, if any difference, certainly not something that would show up in an 8x10 print. I can give some examples from my own photos. I had to reach a bit because most of my photos are of family and I didn't want to post people shots.

I see clear differences with my 16MP, 20MP and 24MP MFT & 35mm-format images shot at high ISO. PRIME was already excellent at noise removal. I mean, my MFT ISO 3200 images already had almost no noise with PRIME. The improvements in DeepPRIME are subtle but significant. It does a much better job of retaining fine detail and eliminating artifacts along high-contrast edges. At small print/display sizes, these differences may not be obvious, but DeepPRIME will enable much larger display/printing of images with lots of fine detail.

MaynardGKrebbs said:
So here goes:

I saw the least difference with normally exposed shots from my D750. Below is a shot taken at maximum ISO of 51K with OldPrime (first) and DeepPrime (second). I can definitely see some difference in the details, but, as I stated, I don't think it would be noticeable in an 8x10 print.

Probably not. But, larger prints are a very different story. If you're only making 8x10 prints, the D750 is already ridiculous overkill.

MaynardGKrebbs said:
Even in the shot taken with my D7000 at its max ISO of 25K, the difference isn't any more dramatic. The only time that I saw a dramatic difference was with a shot that I took to test ISO independence and effect of bit depth and compression on my D750. This shot (below) was taken at ISO 51K and underexposed by 4.66 stops. (By the by, up underexposed by 3.66 stops, I saw virtually no difference between 12 bit and 14 bit, between lossy compression and lossless compression. The shot below is at 14 bit, lossy.)

For my part, I'm struggling to see how the new software is worth $80 an I'm not sure that I'll upgrade. Your mileage will, of course, vary depending what kind of camera and what you do with the photo (e.g., how big you print, crop, or view on screen).

I shoot corporate events with MFT up to ISO 3200 and 35mm format up to ISO 25,600. DeepPRIME won't make any difference in the images my clients post on social media, but it'll make their large-format printed marketing materials (e.g. posters) look better.

MaynardGKrebbs said:
Nikon D750, ISO51k, DxO Prime, Default settings

Nikon D750, ISO51k, DxO DeepPrime, Default settings

Nikon D7000, ISO 25K, OldPrime, Default Settigs

Nikon D7000, ISO 25K, DeepPrime, Default Settings

Nikon D750, ISO 51K, Underexposed by 4.66 stops, OldPrime, Default Settings

Nikon D750, ISO 51K, Underexposed by 4.66 stops, DeepPrime, Default Settings

For event shooters like me, DeepPRIME is almost as big an advance as the original PRIME was, and that was huge. I am over the moon about this upgrade. Now I just gotta get me a better GPU...

--
"I learned long ago, never to wrestle with a pig. You get dirty, and besides, the pig likes it." - George Bernard Shaw
http://jacquescornell.photography
http://happening.photos

CAcreeks · Nov 6, 2020

MaynardGKrebbs said:
I'm kinda old fashioned and still print and hang photos, so my threshold for significant difference is whether it would be noticeable in an 8x10 or 11x14 print viewed at normal viewing distance. For me ... smaller differences are academic and that's how I judge the difference between the two NR options.

Good for you, making and hanging photos. I wouldn't know where to hang them! At home our walls are full of children's art, which is not only good to look at, but makes the kids feel appreciated when visiting.

There is a huge difference between looking at a > 20Mp photo on a 27" 4K monitor, and looking at an 8x11 print of it. It was significant before but with 4K it's ridiculous.

MaynardGKrebbs said:
A few years ago, I suggested DxO add the ability to cut and past just white balance. I also suggested that they add adjust white balance to match a single color between shots, something that would be really useful when photographing an event where the light varies across the space or changes.

That is a great idea! Did they do it? When taking a series of shots, especially where you want blue sky to match hue between frames, because that's the most visible WB issue, you have two choices:

use camera white balance, and hope
set specific white balance and try to maintain angle and exposure

Neither works. :-(

Nice, I just noticed that DxO Photolab is available for Mac OS X.

mujana · Nov 6, 2020

Jacques Cornell said:
MaynardGKrebbs said:

It may be my lack of discernment, but a lot of this discussion seems to overstate the difference between DeepPrime and Old Prime. Except for shots taken with small sensor compacts, I don't see much, if any difference, certainly not something that would show up in an 8x10 print. I can give some examples from my own photos. I had to reach a bit because most of my photos are of family and I didn't want to post people shots.

Click to expand...

I see clear differences with my 16MP, 20MP and 24MP MFT & 35mm-format images shot at high ISO. PRIME was already excellent at noise removal. I mean, my MFT ISO 3200 images already had almost no noise with PRIME. The improvements in DeepPRIME are subtle but significant. It does a much better job of retaining fine detail and eliminating artifacts along high-contrast edges. At small print/display sizes, these differences may not be obvious, but DeepPRIME will enable much larger display/printing of images with lots of fine detail.

Jacques Cornell said:

So here goes:

I saw the least difference with normally exposed shots from my D750. Below is a shot taken at maximum ISO of 51K with OldPrime (first) and DeepPrime (second). I can definitely see some difference in the details, but, as I stated, I don't think it would be noticeable in an 8x10 print.

Click to expand...

Probably not. But, larger prints are a very different story. If you're only making 8x10 prints, the D750 is already ridiculous overkill.

Jacques Cornell said:

Even in the shot taken with my D7000 at its max ISO of 25K, the difference isn't any more dramatic. The only time that I saw a dramatic difference was with a shot that I took to test ISO independence and effect of bit depth and compression on my D750. This shot (below) was taken at ISO 51K and underexposed by 4.66 stops. (By the by, up underexposed by 3.66 stops, I saw virtually no difference between 12 bit and 14 bit, between lossy compression and lossless compression. The shot below is at 14 bit, lossy.)

For my part, I'm struggling to see how the new software is worth $80 an I'm not sure that I'll upgrade. Your mileage will, of course, vary depending what kind of camera and what you do with the photo (e.g., how big you print, crop, or view on screen).

Click to expand...

I shoot corporate events with MFT up to ISO 3200 and 35mm format up to ISO 25,600. DeepPRIME won't make any difference in the images my clients post on social media, but it'll make their large-format printed marketing materials (e.g. posters) look better.

Jacques Cornell said:

Nikon D750, ISO51k, DxO Prime, Default settings

Nikon D750, ISO51k, DxO DeepPrime, Default settings

Nikon D7000, ISO 25K, OldPrime, Default Settigs

Nikon D7000, ISO 25K, DeepPrime, Default Settings

Nikon D750, ISO 51K, Underexposed by 4.66 stops, OldPrime, Default Settings

Nikon D750, ISO 51K, Underexposed by 4.66 stops, DeepPrime, Default Settings

Click to expand...

For event shooters like me, DeepPRIME is almost as big an advance as the original PRIME was, and that was huge. I am over the moon about this upgrade. Now I just gotta get me a better GPU...

--
"I learned long ago, never to wrestle with a pig. You get dirty, and besides, the pig likes it." - George Bernard Shaw
http://jacquescornell.photography
http://happening.photos

+1. Agree 100%

NAwlins Contrarian · Nov 6, 2020

CAcreeks said:
CAcreeks said:

A few years ago, I suggested DxO add the ability to cut and past just white balance. I also suggested that they add adjust white balance to match a single color between shots, something that would be really useful when photographing an event where the light varies across the space or changes.

Click to expand...

That is a great idea! Did they do it? When taking a series of shots, especially where you want blue sky to match hue between frames, because that's the most visible WB issue, you have two choices:

use camera white balance, and hope

set specific white balance and try to maintain angle and exposure

Neither works. :-(

We're getting a bit OT, and maybe this deserves its own thread, but I too would really like a feature that let you select a small to moderate-size area in one image and a corresponding area in another, and have software automatically adjust white balance (or even curves) to match the second to the first. This would be great for skin tones. I've done some experimenting there, with no technique I've developed being wholly satisfactory.

Smitty1 · Nov 6, 2020

Austinian said:
Breakfastographer 2 said:

Austinian said:

Breakfastographer 2 said:

Thanks for adding the SOOC jpeg. I'm still curious about the untampered output from DxO (no PRIME, DeepPRIME or HQ denoising) compared to PRIME and DeepPRIME, but seeing the SOOC jpeg gives a first idea.

I would be doing my own tests if I currently had a machine that could run it, so I appreciate your efforts.

Click to expand...

I don't have a PRIME, but here's a no-NR image vs. DeepPrime. Heavily edited and shadows brightened.

No NR

DeepPrime

Bad as the half-size JPEG is, the full-size No NR TIFF looked worse.

Click to expand...

Thank you for posting this. What I take from this is that DeepPRIME is rather rigorous about ditching shadow detail in its quest to return a nice-looking image. So the positive is that it doesn't seem to want to hallucinate detail where there is none. On the other hand, it's rather decidedly cutting down on dynamic range.

I can't help but wonder if that's a big part of how it works - just ditch lots of shadow detail... as a result, the well-lit parts look super crisp while the shadows look a bit like an oil painting.

I have to confess this one section has me in awe, and a little suspicious (sample reproduced in the name of expediency - hope this is acceptable):

Click to expand...

Sure, glad to see some skepticism; I find some of what DeepPrime does suspiciously good.

The image was from Carlsbad in 2015 with a APS-C camera; one exposure from what was originally a 5-shot HDR series, -2 EV from what the AE wanted.

BTW, that colored lighting no longer exists.

RANT ON.

the last time I was there the park had replaced it with boring, near-tungsten-like LEDs! They told me this was 'more natural'. WTH is "natural" in a cave hundreds of feet below the surface, except total darkness! Grrr! The old lighting was beautiful.

RANT OFF.

Austinian said:

[ATTACH alt=""Before" on the left, probably obvious!"]2596385[/ATTACH]
"Before" on the left, probably obvious!

Were they really able to get that detailed texture back without "help"? Hmm...

Click to expand...

All editing was done on the entire image, with and without NR as the last step.

I did clone out a couple of blown areas near the middle large, bright stalagmite, and also a number plaque and information sign where the right-side railing makes a sharp left turn.

I hope this is at least plausible. :-D

Oh, and DeepPrime was at the default setting. Luminance 40. Plus heavy adjustment to light and shadows, of course.

It does a fine job on color noise and removing the purple cast seen in most high ISO shots (Across camera brands). Seems to retain a lot of the detail too.

I think this is a best case for the feature.

I was going to get Capture One but this feature alone has me reconsidering for PhotoLab. It isn't going to make an ISO 25600 shot look like ISO 100 but, in my tests and yours, it does look to clean up a stop or two.

Thanks for the before/after comparison shots

Austinian · Nov 6, 2020

Smitty1 said:
It does a fine job on color noise and removing the purple cast seen in most high ISO shots (Across camera brands). Seems to retain a lot of the detail too.

That's what makes DeepPRIME superior to the older versions of PRIME, IMO.

Member said:
I think this is a best case for the feature.

I was going to get Capture One but this feature alone has me reconsidering for PhotoLab. It isn't going to make an ISO 25600 shot look like ISO 100 but, in my tests and yours, it does look to clean up a stop or two.

Thanks for the before/after comparison shots

You're welcome. Capture One is also a fine raw processor (I've used the free Express version at times), but I prefer PhotoLab over every other raw processor I've tried.

Breakfastographer 2 · Nov 10, 2020

MaynardGKrebbs said:
Nikon D750, ISO 51K, Underexposed by 4.66 stops, OldPrime, Default Settings

Nikon D750, ISO 51K, Underexposed by 4.66 stops, DeepPrime, Default Settings

Thank you so much for this. Finally, we have a pretty clear example of an artefact. In this example, DeepPRIME presents with a great amount of confidence the print on the back of the cooking book ("ZAHAV").

Here's an actual picture of what that book looks like:

https://static01.nyt.com/images/2015/10/07/dining/07COOKBOOK3/07COOKBOOK3-articleLarge-v3.jpg

So DeepPRIME is making a bad inference of what the book looks like, and then presenting it with a lot of confidence by removing any noise or blur from it.

I also suspect that wall doesn't have quite that strange anisotropic appearance DeepPRIME is assigning to it.

--
https://breakfastographer.wordpress.com

Raist3d · Nov 10, 2020

Seems to me it is prime the one with the artifacts on the book Can you point out kore specifically what am I missing? Thanks

NAwlins Contrarian · Nov 10, 2020

MaynardGKrebbs said:
It may be my lack of discernment, but a lot of this discussion seems to overstate the difference between DeepPrime and Old Prime.... I can give some examples from my own photos. I had to reach a bit because most of my photos are of family and I didn't want to post people shots. So here goes:

Nikon D750, ISO 51K, Underexposed by 4.66 stops, OldPrime, Default Settings

Nikon D750, ISO 51K, Underexposed by 4.66 stops, DeepPrime, Default Settings

As Breakfastographer 2 pointed out (at https://www.dpreview.com/forums/post/64543924), your third example shows artifacts caused by DeepPrime. I've taken your shots, cropped to the book pointed out, and marked a couple of them:

Be sure to view at 100% to see this clearly.

So although I'm generally a big DxO fan and I think DeepPrime is quite promising, apparently it can induce artifacts.

Raist3d · Nov 10, 2020

yes, an AI driven ML algorithm for this can indeed introduce artifacts. So far I am finding more artifact situations with PRIME than deepPRIME.

Marco Cinnirella · Nov 10, 2020

Raist3d said:
yes, an AI driven ML algorithm for this can indeed introduce artifacts. So far I am finding more artifact situations with PRIME than deepPRIME.

Raist3d · Nov 10, 2020

Marco Cinnirella said:
Raist3d said:

yes, an AI driven ML algorithm for this can indeed introduce artifacts. So far I am finding more artifact situations with PRIME than deepPRIME.

Click to expand...

I do have to wonder what folks expect- I’m yet to find any noise reduction solution that is perfect- Deep PRIME seems the most promising solution to come along in a while but using NR has always been a game of gains and losses- it’s not a magic bullet.

Oh I agree. I think it's good to know what things do. For PRIME itself I liked it but started to bump tiny artifacts that were somewhat common like someone's shirt - the buttons. Or door knobs. So far deepPRIME seems to do significantly better for me here. I'll accept whatever new artifacts are there instead of the PRIME artifacts.

Marco Cinnirella said:
What DXO have done is move things on substantially and what’s an added bonus for me is the massive speed increase even though my PC is not state of the art

Oh, I agree with that, not disputing that at all.

Breakfastographer 2 · Nov 11, 2020

Raist3d said:
Seems to me it is prime the one with the artifacts on the book Can you point out kore specifically what am I missing? Thanks

There are several problems with the DeepPrime image, but I'll give you the one that's probably easiest to find.

One letter "A" has two holes and one has none. I provided you with the clean image that shows the same book so you know what the back should look like.

Make sure you're looking at the right image - the one labelled "OldPrime" is just a lot more blurry, which is an honest way to deal with the situation rather than an artefact.

Breakfastographer 2 · Nov 11, 2020

Marco Cinnirella said:
Raist3d said:

yes, an AI driven ML algorithm for this can indeed introduce artifacts. So far I am finding more artifact situations with PRIME than deepPRIME.

Click to expand...

I do have to wonder what folks expect- I’m yet to find any noise reduction solution that is perfect- Deep PRIME seems the most promising solution to come along in a while but using NR has always been a game of gains and losses- it’s not a magic bullet. What DXO have done is move things on substantially and what’s an added bonus for me is the massive speed increase even though my PC is not state of the art

Hi Marco,

As you may be aware aware, image processing is currently at an important cross-road with respect to the future of imaging. If you've followed the news, you may be aware of what "Hallucination" is in the context of super-resolution and noise reduction.

Hallucination can be thought of as the little brother of techniques used in FakeApp and similar developments that allow seamlessly replacing content in videos and still images (i.e. falsifying the original image content). FakeApp-like techniques are now banned on many video platforms, for very good reasons.

Hallucination has not yet hit the news in a big way because the consequences so far have been accidental, like the example of it inserting an actual person's image into a stained glass window. But aside from being an actual recognised algorithm, machine learning techniques that use known image content to create or match new image content are at risk from developing hallucination-like effects.

As a somewhat abstract and imagined example, imagine an image of a tennis court that shows an unassignable shadow in the shape of an ellipse. Machine learning might infer that there should be an additional ball in the image, because that's statistically probable. And just like that, you create a tennis game with two balls in play - obviously not Federation rules!

So that's easy to spot, but we should be concerned about images where the mistake is plausible, but strongly suggestive of facts that aren't true, like an image of soldiers in an armistice, where hallucination suggests a gun visible in the background should be fired because it's statistically probable based on images it has seen before - or some spec of noise or dust in front of the gun that the AI can't, statistically probably, assign to facts in any other way.

If the image with the fake fired gun were published, it might result in unwanted political consequences.

In my opinion, we need to all be sceptical and observant of technological developments in the AI arena, and especially the temptation to create, and use, an "auto" button that triggers or includes AI modifications to an image. I can honestly say that when I've processed a sharp 50-ish megapixel image, I can no longer reasonably monitor for the possible presence of pixel patterns somewhere in the image that might suggest things that aren't true. There simply isn't time. And I do think that's worrying.

--
https://breakfastographer.wordpress.com

Raist3d · Nov 11, 2020

Breakfastographer 2 said:
Raist3d said:

Seems to me it is prime the one with the artifacts on the book Can you point out kore specifically what am I missing? Thanks

Click to expand...

There are several problems with the DeepPrime image, but I'll give you the one that's probably easiest to find.

One letter "A" has two holes and one has none. I provided you with the clean image that shows the same book so you know what the back should look like.

Make sure you're looking at the right image - the one labelled "OldPrime" is just a lot more blurry, which is an honest way to deal with the situation rather than an artefact.

Ok. Well, all I can say is that deepPRIME is working better than PRIME for me.

DxO DeepPrime versus Prime, results and controls

Forum Pro

Veteran Member

Dances With Werewolves

Forum Pro

New member

Forum Pro

Forum Pro

Forum Pro

Forum Pro

Smitty1

Guest

Dances With Werewolves

Leading Member

Raist3d

Guest

Forum Pro

Raist3d

Guest

Veteran Member

Raist3d

Guest

Leading Member

Leading Member

Raist3d

Guest

Similar threads

Keyboard shortcuts