Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
It reverses an LPF. I would call it a high-pass filter. Read the reference I provided if you don't understand.
The D800E has a single LPF which creates a blur in the vertical dimension. It also has a second filter (call it what you will) that refocuses the blur to make it sharp again. Sharpening is a high-pass operation, not a low-pass operation.
The D800 has two LPF filters, one which blurs vertically, and one which blurs horizontally.
High Pass filter is Photoshop-speak for sharpening. A high pass filter would attenuate lower frequencies and a low pass filter would attenuate higher frequencies. Combine the two and you have neither low pass nor high pass filter, but a band pass filter where everything but frequencies within a range is attenuated. Is that what is happening in D800e?
In both, D800 and D800e, the first layer (LPF) attenuates high frequencies in one direction. It is the next layer where the two differ. So, let us investigate.
In D800, the second layer attenuates high frequencies in a direction that is perpendicular to the first. While we at it, let us re-visit a claim that both Floyd and you have made about a second filter only boosting the effect of the first. That is a wrong assumption when you consider that while the second LPF is adding to the effects of the first, it is actually adding it in another dimension but not adding to the first to make it "more split" in the same direction. The idea, with two LPFs, has to be in attenuating certain high frequencies in horizontal and vertical directions.
In D800e, the second layer is not doing that. But is it re-generating higher frequencies that were previously blocked from low frequency signal? Or, is it blocking lower frequencies than were allowed to pass through? The latter is how a high pass filter would work, and effectively the two layers combine to give a band pass filter instead where the extreme high and extreme low frequencies are attenuated.
Quote:
Originally Posted by Floyd_Davidson
When you find technical documents from Nikon that support your statements, please do provide them.
I've cited a list of credible technical sources. You didn't read a single one of them.
So you have none, and were making an excuse on "technical grounds" and "lost in translation".
The D800 has two LPF filters, one which blurs vertically, and one which blurs horizontally.
That is perhaps valid in layman's vernacular (which should be sufficient here, so I don't want to be overly critical of that usage), but technically there is one low pass filter, and the layers being referred to are components of the filter, but are not low pass filters in themselves.
There are two layers of birefringent material that act as beam splitters, with a polarity shift and a phase shift for one of the beams. That is not in and of itself a low pass filter. It doesn't cause any amplitude distortion until the beams are combined. That happens when the light is measured by a sensor. Because the sensor has spatial sampling, and the width of the beam split is equal or less than the space between two sensor locations, it then and only then has the effect of removing higher spatial frequencies and not removing lower spatial frequencies (amplitude distortion).
Hence the device shown in the diagram Nikon has provided is a single low pass filter. It has multiple components that work together to form a low pass filter. There are two layers of birefringent material, there is a sensor that samples in the spatial domain, and there appears to be at least one other layer that has not been defined by Nikon and may or may not be significant to the technical discussion (even though it is not significant for a marketing blurb). It is interesting that the unspecified layer is illustrated with different colors for the D800 and the D800E, which suggest that it might be two distinctly different components (such as something which shifts either polarity or phase).
To put it mildly, the diagram leaves many more unanswered questions than answers. And those reading technical definitions for "Low Pass Filter" into are in error.
^^ I see what you're saying, Floyd. To me, a single layer of birefringent material could be viewed as a low pass filter for a single axis. But it's not a complete LPF when viewed spatially by the sensor. I think this is in the realm of semantics, and whether a tree falling in the forest with no one around makes a noise or not. But I can't really argue with your point.
High Pass filter is Photoshop-speak for sharpening.
So now "Photoshop-speak" is your technical reference for definitions? Actually "sharpen" tools, in any image editor, are an implementation of a DSP high pass filter to the spatial domain. It has nothing at all to do with Photoshop, and in fact Adobe has misused the term "High Pass" to name one specific algorithm, as opposed to others that are also "high pass" filters. It's much like some people who distinguish between Unsharp Mask and an high pass sharpen too by calling the latter "convolutional sharpening". Not much of a distinciong, because convolution is a way to find edges, not a way to sharpen them, and is often used by USM algorithms too!
Quote:
A high pass filter would attenuate lower frequencies and a low pass filter would attenuate higher frequencies.
At last you got at least one thing correct!
Quote:
Combine the two and you have neither low pass nor high pass filter, but a band pass filter where everything but frequencies within a range is attenuated. Is that what is happening in D800e?
That does not necessarily produce a band pass filter, but that is the way a band pass filter is accomplished. It can also produce a band rejection filter, or just a simple attenuator.
The D800E is not using any of the above. There is no low pass filter used for anti aliasing. Instead there is a beam splitter that also causes a phase shift, and then another one that shifts the phase in exactly the oppose amount. It probably does result in at least some loss of light, so it would be an attenuator. But it is not a combination of a low pass filter plus either another low pass filter nor a high pass filter.
Quote:
In both, D800 and D800e, the first layer (LPF) attenuates high frequencies in one direction. It is the next layer where the two differ. So, let us investigate.
If you will simply do some actual research on birefringent materials the first thing you'll discover is that the first layer is not actually a low pass filter and does not attenuate high frequencies. It is a beam splitter than shifts the polarity and the phase of one of the two beams.
Quote:
In D800, the second layer attenuates high frequencies in a direction that is perpendicular to the first.
The beam is split in a different direction. There is no attenuation of high frequencies at that point, just as there was none at the first beam splitter.
Quote:
While we at it, let us re-visit a claim that both Floyd and you have made about a second filter only boosting the effect of the first. That is a wrong assumption when you consider that while the second LPF is adding to the effects of the first, it is actually adding it in another dimension but not adding to the first to make it "more split" in the same direction. The idea, with two LPFs, has to be in attenuating certain high frequencies in horizontal and vertical directions.
Please cease with the techie nonsense. "Adding it in another dimension" is nonsense. It's a beam splitter... there is no attenuation of frequencies at that point. There are two beams, one of which has the polarity shift and has a phase shift.
Quote:
In D800e, the second layer is not doing that. But is it re-generating higher frequencies that were previously blocked from low frequency signal?
Regenerating signal that was removed? Wow, I suspect you also have a perpetual motion machine in your closet? That's nonsense. It cannot regenerate anything.
Quote:
Or, is it blocking lower frequencies than were allowed to pass through? The latter is how a high pass filter would work, and effectively the two layers combine to give a band pass filter instead where the extreme high and extreme low frequencies are attenuated.
It does neither. It shifts the phase of one of the two beams. No loss at all until the beams are summed at the sensor, and then only if the signals are out of phase.
Quote:
So you have none, and were making an excuse on "technical grounds" and "lost in translation".
I have cited several credible technical sources that explain it, and you have cited nothing other than a single diagram provided by the marketing people at Nikon. What a hoot...
Last edited by Floyd_Davidson; 02-15-2012 at 12:00 PM..
^^ I see what you're saying, Floyd. To me, a single layer of birefringent material could be viewed as a low pass filter for a single axis. But it's not a complete LPF when viewed spatially by the sensor. I think this is in the realm of semantics, and whether a tree falling in the forest with no one around makes a noise or not. But I can't really argue with your point.
But there is absolutely no amplitude distortion in the birefringent layer. None. If there were, there would be no way to cancel the effect with a similar second layer! If it is a Low Pass Filter it by definition must result in amplitude distortion. And of course not only is there no loss of higher frequencies in the first layer, there is no regeneration of higher frequencies in the second layer!
The birefringent layers are beam splitters and phase shifters, not low/high pass filters.
So now "Photoshop-speak" is your technical reference for definitions?
No, I provided a more technical explanation later.
Quote:
That does not necessarily produce and band pass filter, but that is the way a band pass filter is accomplished. It can also produce a band rejection filter, or just a simple attenuator.
You have a habit of repeating someone's words, immediately after disagreeing with them. Repeat after me: such combination of a high pass filter plus a low pass filter gets you... a band pass filter. We can call it band rejection filter if it were designed to reject a band, which would not apply here.
Quote:
The D800E is not using any of the above. There is no low pass filter used for anti aliasing. Instead there is a beam splitter that also causes a phase shift
How is this "beam splitter" (since you're allergic to calling it an LPF) is different from that in D800?
Quote:
It probably does result in at least some loss of light, so it would be an attenuator.
Probably? Likely. But I'm not into complications that real world challenges to designing anything when trying to focus on the idea behind the design itself. To that effect, what is being attenuating is the key. I said, certain frequencies. Do you disagree?
Quote:
But it is not a combination of a low pass filter plus either another low pass filter nor a high pass filter.
We should have this resolved as soon as you provide us with the "technical diagram" you boast about relying upon.
Quote:
If you will simply do some actual research on birefringent materials the first thing you'll discover is that the first layer is not actually a low pass filter and does not attenuate high frequencies. It is a beam splitter than shifts the polarity and the phase of one of the two beams.
Funny how you challenged my argument on "Photoshop-speak" earlier and then say this. What is the difference between "splitting" versus "blurring"? Which of the two does the first layer accomplish, but not the other?
Quote:
Regenerating signal that was removed? Wow, I suspect you also have a perpetual motion machine in your closet? That's nonsense. It cannot regenerate anything.
Ain't that funny? I presented two scenarios, which would you cling on to if not this? This was an argument placed to counter the idea that the second layer is a high pass filter to negate the effects of the previous (low pass filter).
Quote:
It does neither. It shifts the phase of one of the two beams.
Two beams? Having said that, recall that I brought up noise cancellation headphones earlier. They are based on phase-shift of wave forms at certain frequencies. This second layer appears to be working with the same premise. So, we have something in common here. Where we differ is your idea of splitting a beam into two beams whereas what is really happening is some blurring of the light using a LPF layer followed by de-blurring of that light with a completely out of phase shift in the second but using pretty much the same concept as the first. This is the point I was trying to make by taking exception to the idea of high pass filter.
Quote:
I have cited several credible technical sources that explain it, and you have cited nothing other than a single diagram provided by the marketing people at Nikon. What a hoot...
Should I blame your sources for your ideas of "splitting the beams" as opposed to blurring?
But there is absolutely no amplitude distortion in the birefringent layer. None. If there were, there would be no way to cancel the effect with a similar second layer! If it is a Low Pass Filter it by definition must result in amplitude distortion. And of course not only is there no loss of higher frequencies in the first layer, there is no regeneration of higher frequencies in the second layer!
The birefringent layers are beam splitters and phase shifters, not low/high pass filters.
Ok, by your logic the birefringent layers aren't LPF's at all in of themselves. That actually didn't sit right with me either although I didn't articulate it. They are contributing components of an LPF filter. So really, the D800 has a single LPF whereas the D800E has none.
No, I provided a more technical explanation later.
Ha ha, funny story.
Quote:
You have a habit of repeating someone's words, immediately after disagreeing with them. Repeat after me: such combination of a high pass filter plus a low pass filter gets you... a band pass filter. We can call it band rejection filter if it were designed to reject a band, which would not apply here.
No, I disagreed with you, and pointed out that it does not necessarily result in a band pass filter. It sounds as if you don't know that a "band pass filter" is not the same as a "band rejection filter".
Quote:
How is this "beam splitter" (since you're allergic to calling it an LPF) is different from that in D800?
Nikon has not indicated how, or if, it is different, other than apparently being rotated by 180 degrees.
Quote:
Probably? Likely. But I'm not into complications that real world challenges to designing anything when trying to focus on the idea behind the design itself. To that effect, what is being attenuating is the key. I said, certain frequencies. Do you disagree?
I disagree, and so does Nikon. They have specifically said there is no affect on "certain frequencies". No anti-aliasing low pass filter. None, Zilch. Nada.
Quote:
We should have this resolved as soon as you provide us with the "technical diagram" you boast about relying upon.
Funny how you challenged my argument on "Photoshop-speak" earlier and then say this. What is the difference between "splitting" versus "blurring"? Which of the two does the first layer accomplish, but not the other?
Ain't that funny? I presented two scenarios, which would you cling on to if not this? This was an argument placed to counter the idea that the second layer is a high pass filter to negate the effects of the previous (low pass filter).
Two beams? Having said that, recall that I brought up noise cancellation headphones earlier. They are based on phase-shift of wave forms at certain frequencies. This second layer appears to be working with the same premise. So, we have something in common here. Where we differ is your idea of splitting a beam into two beams whereas what is really happening is some blurring of the light using a LPF layer followed by de-blurring of that light with a completely out of phase shift in the second but using pretty much the same concept as the first. This is the point I was trying to make by taking exception to the idea of high pass filter.
Should I blame your sources for your ideas of "splitting the beams" as opposed to blurring?
Look, this has been explained to you several times, and sources have been cited. READ THE CITED REFERENCES. And stop making up nonsense from your imagination.
bi·re·frin·gence (br-frnjns)
n.
The resolution or splitting of a light wave into two unequally reflected or transmitted waves by an optically anisotropic medium such as calcite or quartz. Also called double refraction.
Ok, by your logic the birefringent layers aren't LPF's at all in of themselves. That actually didn't sit right with me either although I didn't articulate it. They are contributing components of an LPF filter. So really, the D800 has a single LPF whereas the D800E has none.
No, I disagreed with you, and pointed out that it does not necessarily result in a band pass filter. It sounds as if you don't know that a "band pass filter" is not the same as a "band rejection filter".
Your ears might be playing tricks with you. Use your brain, as you read my post on the subject.
Quote:
Nikon has not indicated how, or if, it is different, other than apparently being rotated by 180 degrees.
Rotated by what? Phase-shift sounded much better.
Quote:
I disagree, and so does Nikon. They have specifically said there is no affect on "certain frequencies". No anti-aliasing low pass filter. None, Zilch. Nada.
Well, we're still waiting on you providing your "technical source" towards that effect (no, not blogoshere, but straight from the horses mouth), since you've been in denial mode with whatever Nikon has provided.
Quote:
Look, this has been explained to you several times, and sources have been cited. READ THE CITED REFERENCES. And stop making up nonsense from your imagination.
bi·re·frin·gence (br-frnjns)
n.
The resolution or splitting of a light wave into two unequally reflected or transmitted waves by an optically anisotropic medium such as calcite or quartz. Also called double refraction.
Damn, you can find meanings in online dictionaries and post links here.
But, how does it explain what Nikon is doing?
Last edited by EinsteinsGhost; 02-15-2012 at 02:55 PM..
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
