Low image quality on newer GPUs

[MarkFMB]

Made a video to show one of the games where i can perceive the problem the most MGSV. Notice all the sharp lines, fences, railings etc.etc. in the distance. I know this game has poor AA in its engine, so that's prolly why it's more pronounced than in others. Still you can see in the vid how bad the isse gets when disabling DSR and reverting to native. That's seriously not how the game plays normally at native, i've seen it in other rigs, i've seen it in other videos, i've played it before on my old pc. Something is at work there. You can see it's even pretty shimmery with 4k dsr as well. Also on my old card the game ran smoothly without any of these artifacts (or if any, they were so limited to the point i never noticed any in 40 hours of gameplay, while i noticed these ones in less than 10 second on first startup and you can very well guess why seeing how pronounced they are without DSR). Now that i tried that old card on my rig and possibly "infected" it, it shows these same symptoms.

(At a certain point i revert back to 4k DSR and the screen capture program seems to be unable to pick it up again for some reason, but there's plenty to go on even before that point so don't mind that).

[tsrbino]

I am from the US. Several from the NVidia forum with this problem were from Russia, and I believe MarkFMB said he was from Spain, so it doesn't seem to be geographically exclusive. My electrical system is fairly old, it uses a fuse box and is not grounded.

[tsrbino]

I thought of a way we should be able to prove or disprove the "bad power" theory.  Take a computer that is experiencing this problem and instead run it on a high quality UPS.  Does the problem remain?  I want to know

I do not have access to a UPS.  I remember reading somewhere that someone supposedly fixed the issue by using one.  However, I have read several posts of other people who have tried it and said that it did not fix it.

[LokiFire]

Made a video to show one of the games where i can perceive the problem the most MGSV. Notice all the sharp lines, fences, railings etc.etc. in the distance. I know this game has poor AA in its engine, so that's prolly why it's more pronounced than in others. Still you can see in the vid how bad the isse gets when disabling DSR and reverting to native. That's seriously not how the game plays normally at native, i've seen it in other rigs, i've seen it in other videos, i've played it before on my old pc. Something is at work there. You can see it's even pretty shimmery with 4k dsr as well. Also on my old card the game ran smoothly without any of these artifacts (or if any, they were so limited to the point i never noticed any in 40 hours of gameplay, while i noticed these ones in less than 10 second on first startup and you can very well guess why seeing how pronounced they are without DSR). Now that i tried that old card on my rig and possibly "infected" it, it shows these same symptoms.

(At a certain point i revert back to 4k DSR and the screen capture program seems to be unable to pick it up again for some reason, but there's plenty to go on even before that point so don't mind that).

The railings are all square shaped, this is why aliasing is such an issue in this game for you, it's purely aliasing.

 

Are people really just passing normal things off as something only they have witnessed?

 

 

OMFG.

[LokiFire]

From Wikipedia, the free encyclopedia

 

 

This article is about aliasing in signal processing, including computer graphics. For aliasing in computer programming, see aliasing (computing).

 

Properly sampled image of a brick wall

 

Spatial aliasing in the form of amoiré pattern

In signal processing and related disciplines, aliasing is an effect that causes different signals to become indistinguishable (or aliases of one another) when sampled. It also refers to the distortion or artifact that results when the signal reconstructed from samples is different from the original continuous signal.

Aliasing can occur in signals sampled in time, for instance digital audio, and is referred to as temporal aliasing. Aliasing can also occur in spatially sampled signals, for instance digital images. Aliasing in spatially sampled signals is called spatial aliasing.

 

Contents   [hide] 

• 1Description
• 2Bandlimited functions
• 3Bandpass signals
• 4Sampling sinusoidal functions
  • 4.1Folding
  • 4.2Complex sinusoids
  • 4.3Sample frequency
• 5Historical usage
• 6Angular aliasing
• 7More examples
  • 7.1Online audio example
  • 7.2Direction finding
• 8See also
• 9Notes
• 10Citations
• 11Further reading

 

Description[edit]

 

Left: An aliased image of the letter A in Times New Roman. Right: An anti-aliased image. (See also: Font rasterization)

When a digital image is viewed, a reconstruction is performed by a display or printer device, and by the eyes and the brain. If the image data is not properly processed during sampling or reconstruction, the reconstructed image will differ from the original image, and an alias is seen.

An example of spatial aliasing is the moiré pattern one can observe in a poorly pixelized image of a brick wall. Spatial anti-aliasing techniques avoid such poor pixelizations. Aliasing can be caused either by the sampling stage or the reconstruction stage; these may be distinguished by calling sampling aliasing prealiasing and reconstruction aliasing postaliasing.^[1]

Temporal aliasing is a major concern in the sampling of video and audio signals. Music, for instance, may contain high-frequency components that are inaudible to humans. If a piece of music is sampled at 32000 samples per second (Hz), any frequency components above 16000 Hz (the Nyquist frequency for this sampling rate) will cause aliasing when the music is reproduced by a digital to analog converter (DAC). To prevent this an anti-aliasing filter is used to remove components above the Nyquist frequency prior to sampling.

In video or cinematography, temporal aliasing results from the limited frame rate, and causes the wagon-wheel effect, whereby a spoked wheel appears to rotate too slowly or even backwards. Aliasing has changed its apparent frequency of rotation. A reversal of direction can be described as a negative frequency. Temporal aliasing frequencies in video and cinematography are determined by the frame rate of the camera, but the relative intensity of the aliased frequencies is determined by the shutter timing (exposure time) or the use of a temporal aliasing reduction filter during filming.^[2]

Like the video camera, most sampling schemes are periodic; that is, they have a characteristic sampling frequency in time or in space. Digital cameras provide a certain number of samples (pixels) per degree or per radian, or samples per mm in the focal plane of the camera. Audio signals are sampled (digitized) with an analog-to-digital converter, which produces a constant number of samples per second. Some of the most dramatic and subtle examples of aliasing occur when the signal being sampled also has periodic content.

Bandlimited functions[edit]

Main article: Nyquist–Shannon sampling theorem

Actual signals have finite duration and their frequency content, as defined by the Fourier transform, has no upper bound. Some amount of aliasing always occurs when such functions are sampled. Functions whose frequency content is bounded (bandlimited) have infinite duration. If sampled at a high enough rate, determined by the bandwidth, the original function can in theory be perfectly reconstructed from the infinite set of samples.

Bandpass signals[edit]

Main article: Undersampling

Sometimes aliasing is used intentionally on signals with no low-frequency content, called bandpass signals. Undersampling, which creates low-frequency aliases, can produce the same result, with less effort, as frequency-shifting the signal to lower frequencies before sampling at the lower rate. Some digital channelizers^[3] exploit aliasing in this way for computational efficiency. See Sampling (signal processing), Nyquist rate (relative to sampling), and Filter bank.

Sampling sinusoidal functions[edit]

Sinusoids are an important type of periodic function, because realistic signals are often modeled as the summation of many sinusoids of different frequencies and different amplitudes (as, for example, with a Fourier series or transform). Understanding what aliasing does to the individual sinusoids is useful in understanding what happens to their sum.

 

Two different sinusoids that fit the same set of samples.

Here, a plot depicts a set of samples whose sample-interval is 1, and two (of many) different sinusoids that could have produced the samples. The sample-rate in this case is . For instance, if the interval is 1 second, the rate is 1 sample per second. Nine cycles of the red sinusoid and 1 cycle of the blue sinusoid span an interval of 10 samples. The corresponding number of cycles per sample are   and .  If these samples were produced by sampling functions  cos(2π(0.9)x−θ)  and  cos(2π(0.1)x−φ),  they could also have been produced by the trigonometrically identical functions,  cos(2π(−0.9)x +θ)  and  cos(2π(−0.1)x +φ),  which introduces the useful concept of Negative frequency.

In general, when a sinusoid of frequency  is sampled with frequency  the resulting number of cycles per sample is  (known asnormalized frequency), and the samples are indistinguishable from those of another sinusoid (called an alias) whose normalized frequency differs from  by any integer (positive or negative).^{[note 1]}  Replacing negative frequency sinusoids by their equivalent positive frequency representations, we can express all the aliases of frequency  as    for any integer N,  with    being the true value, and N has units of cycles per sample. Then the N = 1 alias of   is    (and vice versa).

Aliasing matters when one attempts to reconstruct the original waveform from its samples. The most common reconstruction technique produces the smallest of the   frequencies. So it is usually important that  be the unique minimum. A necessary and sufficient condition for that is  where  is commonly called the Nyquist frequency of a system that samples at rate   In our example, the Nyquist condition is satisfied if the original signal is the blue sinusoid ().  But if   the usual reconstruction method will produce the blue sinusoid instead of the red one.

 

The black dots are aliases of each other. The solid red line is an example of amplitude varying with frequency. The dashed red lines are the corresponding paths of the aliases.

Folding[edit]

In the example above,   and  are symmetrical around the frequency   And in general, as  increases from 0 to   decreases from   to   Similarly, as  increases from   to    continues decreasing from  to 0.

A graph of amplitude vs frequency for a single sinusoid at frequency  and some of its aliases at   and  would look like the 4 black dots in the adjacent figure. The red lines depict the paths (loci) of the 4 dots if we were to adjust the frequency and amplitude of the sinusoid along the solid red segment (between  and ). No matter what function we choose to change the amplitude vs frequency, the graph will exhibit symmetry between 0 and  This symmetry is commonly referred to as folding, and another name for  (the Nyquist frequency) is folding frequency. Folding is most often observed in practice when viewing the frequency spectrum of real-valued samples using a discrete Fourier transform.

 

Two complex sinusoids, colored gold and cyan, that fit the same sets of real and imaginary sample points when sampled at the rate (f_s) indicated by the grid lines. The case shown here is: 

Complex sinusoids[edit]

Complex sinusoids are waveforms whose samples are complex numbers, and the concept of negative frequency is necessary to distinguish them. In that case, the frequencies of the aliases are given by just:   Therefore, as  increases from   to    goes from   up to 0.  Consequently, complex sinusoids do not exhibit folding. Complex samples of real-valued sinusoids have zero-valued imaginary parts and do exhibit folding.

Sample frequency[edit]

 

Illustration of 4 waveforms reconstructed from samples taken at six different rates. Two of the waveforms are sufficiently sampled to avoid aliasing at all six rates. The other two illustrate increasing distortion (aliasing) at the lower rates.

When the condition  is met for the highest frequency component of the original signal, then it is met for all the frequency components, a condition known as the Nyquist criterion. That is typically approximated by filtering the original signal to attenuate high frequency components before it is sampled. These attenuated high frequency components still generate low-frequency aliases, but typically at low enough amplitudes that they do not cause problems. A filter chosen in anticipation of a certain sample frequency is called an anti-aliasing filter.

The filtered signal can subsequently be reconstructed, by interpolation algorithms, without significant additional distortion. Most sampled signals are not simply stored and reconstructed. But the fidelity of a theoretical reconstruction (via the Whittaker–Shannon interpolation formula) is a customary measure of the effectiveness of sampling.

Historical usage[edit]

Historically the term aliasing evolved from radio engineering because of the action of superheterodyne receivers. When the receiver shifts multiple signals down to lower frequencies, from RF to IF by heterodyning, an unwanted signal, from an RF frequency equally far from the local oscillator (LO) frequency as the desired signal, but on the wrong side of the LO, can end up at the same IF frequency as the wanted one. If it is strong enough it can interfere with reception of the desired signal. This unwanted signal is known as an image oralias of the desired signal.

Angular aliasing[edit]

Aliasing occurs whenever the use of discrete elements to capture or produce a continuous signal causes frequency ambiguity.

Spatial aliasing, particular of angular frequency, can occur when reproducing a light field^[4] or sound field with discrete elements, as in 3D displays or wave field synthesis of sound.

This aliasing is visible in images such as posters with lenticular printing: if they have low angular resolution, then as one moves past them, say from left-to-right, the 2D image does not initially change (so it appears to move left), then as one moves to the next angular image, the image suddenly changes (so it jumps right) – and the frequency and amplitude of this side-to-side movement corresponds to the angular resolution of the image (and, for frequency, the speed of the viewer's lateral movement), which is the angular aliasing of the 4D light field.

The lack of parallax on viewer movement in 2D images and in 3-D film produced by stereoscopic glasses (in 3D films the effect is called "yawing", as the image appears to rotate on its axis) can similarly be seen as loss of angular resolution, all angular frequencies being aliased to 0 (constant).

More examples[edit] Online audio example[edit]

The qualitative effects of aliasing can be heard in the following audio demonstration. Six sawtooth waves are played in succession, with the first two sawtooths having a fundamental frequency of 440 Hz (A4), the second two having fundamental frequency of 880 Hz (A5), and the final two at 1760 Hz (A6). The sawtooths alternate between bandlimited (non-aliased) sawtooths and aliased sawtooths and the sampling rate is 22.05 kHz. The bandlimited sawtooths are synthesized from the sawtooth waveform's Fourier series such that no harmonics above the Nyquist frequency are present.

The aliasing distortion in the lower frequencies is increasingly obvious with higher fundamental frequencies, and while the bandlimited sawtooth is still clear at 1760 Hz, the aliased sawtooth is degraded and harsh with a buzzing audible at frequencies lower than the fundamental.

 

 

 

[MarkFMB]

I think we all know very well what aliasing is, thank you. The whole point is that we, with this problem, are getting insane amounts of it (at least temporal aa), almost like no AA is applied at all no matter how many settings we enforce. You've seen the video during the 1080p seconds. No, that's not normal aliasing with AA enabled. Every texture in the distance is crawling back and forth and shimmering at the speed of an olympic athlete non-stop. Go compare it to other gameplay videos you find around. See if it's the same. I also said multiple times with my old card, none of that was present and no it's not a case of what has been seen cannot be unseen, cause i wasn't using dsr or tweaks or anything on my old card so it should have played like the 1080p version, which would have been impossible not to notice since it basicly looks like a ps2 game. It was just not present or present to such a small degree as to be unnoticeable in more than 40 hours of gameplay. Now it has "infected" that card as well.

I'd argue while in 4k dsr a distracted or unconcerned gamer might not notice it, cause it looks a lot better (even if it's still pretty visible). But 4k dsr is not how the game is meant to be played, most rigs couldn't handle it, my own pc with all the nice penny i spent on it can't maintain 30fps everytime in it and i'm still using it cause othetwise the game looks like a hot fuming pile of dung (you can very well see it in 1080p). Jesus my old pc had 6 years old cpu and ram and still managed it on high settings without any of those artifacts. That, right there, is not normal.

[Zathel]

Basically, this whole thread I've been trying to say why I shouldn't have to believe him, and saying that I'm suspicious of foul play, as if someone is trying to slander NVIDIA. There's not a whole lot of evidence to support that, but given the way the OP of this thread has been trying to religiously convert us to his cause, it's only been reported by a select few people as you mentioned, and keeps trying to claim that his screenshots are sufficient evidence, I can't help but wonder...and good look trying to find a way to reproduce this. I've asked multiple times, and whilst the OP has claimed that there is nothing stopping me from doing so, I have never experienced this issue and neither have my brother or sister, and since he has said nothing of a way to induce it, I don't know HOW I'm supposed to test this for myself. 

Have you not seen any of the videos we are posting? 

[Zathel]

I think we all know very well what aliasing is, thank you. The whole point is that we, with this problem, are getting insane amounts of it (at least temporal aa), almost like no AA is applied at all no matter how many settings we enforce. You've seen the video during the 1080p seconds. No, that's not normal aliasing with AA enabled. Every texture in the distance is crawling back and forth and shimmering at the speed of an olympic athlete non-stop. Go compare it to other gameplay videos you find around. See if it's the same. I also said multiple times with my old card, none of that was present and no it's not a case of what has been seen cannot be unseen, cause i wasn't using dsr or tweaks or anything on my old card so it should have played like the 1080p version, which would have been impossible not to notice since it basicly looks like a ps2 game. It was just not present or present to such a small degree as to be unnoticeable in more than 40 hours of gameplay. Now it has "infected" that card as well.

I'd argue while in 4k dsr a distracted or unconcerned gamer might not notice it, cause it looks a lot better (even if it's still pretty visible). But 4k dsr is not how the game is meant to be played, most rigs couldn't handle it, my own pc with all the nice penny i spent on it can't maintain 30fps everytime in it and i'm still using it cause othetwise the game looks like a hot fuming pile of dung (you can very well see it in 1080p). Jesus my old pc had 6 years old cpu and ram and still managed it on high settings without any of those artifacts. That, right there, is not normal.

I've seen videos and playthroughs from famous youtubers that have the same problems.

 

Anyone who has SW Battlefront: Put high FXAA and watch at the wings of the TIE fighter in the menu. Look how they shimmer as it slowly rotates. Now, go to options and put TAA as anti aliasing method. Watch once again the wings of the TIE fighter in the menu. You can't be noticing the difference.

[GSTARR]

Okay. Well, I don't know what you individually can do. Maybe see comparisons with the new nVidia GPUs VS their AMD counter-parts? Maybe it's the same.

Click the link he posted. Did you read

[Sauron]

I have difficulties telling the images apart, what exactly should I be looking for?

[GSTARR]

Good day everyone

<snip>

well this certainly changes the game eh?

I guess my rig will have to wait for a gpu as more research has to be done

[GSTARR]

Ok. First things first:

Where are the people who have this problem from? Is it US mainly? Asia? Europe?

Now I don't have this problem myself but for those who do, have you tried flashing all the bioses and things like that? Even modems should be flashed and see if the problem persists.

I find it very intriguing that graphics cards that work well on a shstem stop working when placed on a system that is already compromised.

How is your gpu utilisation?

How old is your electrical system?

How do you connect to the internet?

What kind of port do you use for display?

It seems the problem is less noticeable in recorded videos so the display connection can be at fault. Is there a difference from recording through shadowplay and fraps?

Cmon people feed us data. And before we go into tvs (I mean digital signal corruption is a bitch! The minimal interference through heat or other electronics can make a lot of difference) lets focus on the pc side of things.

To those with working and not working systems, what is different between those systems?

On the link he provided, although he didn't change the gpu nor the cable he used (HDMI, displayport, whatever it was) he changed almost every single component in his computer and the problem persisted.

 

There's another post where a guy changed his psu and the same thing happened with multiple people agreeing with him saying they had just upgraded as well, so electrical problems are out

[xAcid9]

I have difficulties telling the images apart, what exactly should I be looking for?

 

Hope you're not on mobile. More noticeable aliasing at the grandstand and road on the right side for 980 screenshot, it's more visible in motion.

 

Can't be unseen.

[Sauron]

Hope you're not on mobile. More noticeable aliasing at the grandstand and road on the right side for 980 screenshot, it's more visible in motion.

 

oh, that. could be just a random artifact of the precise screenshot

[xAcid9]

oh, that. could be just a random artifact of the pricise screenshot

 

Nah, i think it's the game engine and how modern GPU driver were "optimized".

[Denisix]

http://www.youtube.com/watch?v=OVjmTseHv10

http://www.youtube.com/watch?v=qCvh5Z5M3Ww

http://www.youtube.com/watch?v=KsYpBQhcg7U

http://www.youtube.com/watch?v=1-Qs3cBYGFw

http://www.youtube.com/watch?v=oa1c0PSoLT8

[leandrocm86]

http://www.youtube.com/watch?v=OVjmTseHv10

http://www.youtube.com/watch?v=qCvh5Z5M3Ww

http://www.youtube.com/watch?v=KsYpBQhcg7U

http://www.youtube.com/watch?v=1-Qs3cBYGFw

http://www.youtube.com/watch?v=oa1c0PSoLT8

 

The first two videos and specially the last one make it clear that we have a problem. The 3rd and 4th I guess are related to other graphical issues.

For those who still think this aliasing is normal, check the last video above, and I also recomend the one Jackiy posted some pages ago:

 

 

The quantity of aliasing we are talking about cannot be considered normal. It is just huge!

 

This really look like that LOD bias problem, but still can't fix it. The thread I mentioned in the beginning of this page looks pretty much dead.

[Denisix]

The first two videos and specially the last one make it clear that we have a problem. The 3rd and 4th I guess are related to other graphical issues.

For those who still think this aliasing is normal, check the last video above, and I also recomend the one Jackiy posted some pages ago:

 

 

The quantity of aliasing we are talking about cannot be considered normal. It is just huge!

 

This really look like that LOD bias problem, but still can't fix it. The thread I mentioned in the beginning of this page looks pretty much dead.

Wrote nvidia said it's not our problem!! I don't understand how the graphics in one day will change..With the many challenges faced but it's out of my League!!Question here on the website just a little bit closer to a solution???

[vanished]

Just restating for any newcomers here, we've been looking into this and so far determined that it is not nvidia's problem; this problem exists on nvidia, amd, and intel graphics, and also occurs on many platforms (windows Linux, etc.)

[Xenift]

Does video evidence bring justice?

I have noticed that when i want to capture screenshots for witcher 3, the LOD is not there, as if a more blurred image is being taken.

[Coaxialgamer]

If i read correctly, you claim that nvidia has a feature in their cards that causes crashes and downgrades image quality.

WHY would they do that?

Even if it where true, why would they simply kneecap perfectly working cards?

I could believe that they do pretty much everything in their power to kill off last gen gpus and push maxwell, but this?

[Coaxialgamer]

Just restating for any newcomers here, we've been looking into this and so far determined that it is not nvidia's problem; this problem exists on nvidia, amd, and intel graphics, and also occurs on many platforms (windows Linux, etc.)

Thanks, ive been kind of lost looking through posts...

[MarkFMB]

Yeah to be honest a change of the thread's title would do the thread more justive by now.

[jjohnthedon1]

I think we all know very well what aliasing is, thank you. The whole point is that we, with this problem, are getting insane amounts of it (at least temporal aa), almost like no AA is applied at all no matter how many settings we enforce. You've seen the video during the 1080p seconds. No, that's not normal aliasing with AA enabled. Every texture in the distance is crawling back and forth and shimmering at the speed of an olympic athlete non-stop. Go compare it to other gameplay videos you find around. See if it's the same. I also said multiple times with my old card, none of that was present and no it's not a case of what has been seen cannot be unseen, cause i wasn't using dsr or tweaks or anything on my old card so it should have played like the 1080p version, which would have been impossible not to notice since it basicly looks like a ps2 game. It was just not present or present to such a small degree as to be unnoticeable in more than 40 hours of gameplay. Now it has "infected" that card as well.

I'd argue while in 4k dsr a distracted or unconcerned gamer might not notice it, cause it looks a lot better (even if it's still pretty visible). But 4k dsr is not how the game is meant to be played, most rigs couldn't handle it, my own pc with all the nice penny i spent on it can't maintain 30fps everytime in it and i'm still using it cause othetwise the game looks like a hot fuming pile of dung (you can very well see it in 1080p). Jesus my old pc had 6 years old cpu and ram and still managed it on high settings without any of those artifacts. That, right there, is not normal.

15 gold stars

[Miguel Batista]

I would stilllike to know more. Knowing that a ups xan solve some people's problems is a good starting point but we need more info. What do these people have in common? That is the right question to ask imho. Also I would like people who have this issue and multiple rigs, it would be interesting to see some more testing.

Someone said that cards that did not have this issue, once put on a rig that has, began to have it. What happens when you put them back in the original rig? Does the problem persist?