However good down sampling involves interpolation, you're just a stubborn kid who won't admit he's got his terms mixed up. And with that, I'm done replying to this thread.
LOL, coming from your mouth! Look, I used the easiest and most concise examples because they are simple. The examples I used previously clearly point out what the terms you are confusing actually means. However, you want to throw in more terms with more erroneous definitions as if muddying the waters with extra terms makes it seem as if you are right.
Here is what all the terms and definitions really mean from your previous post. Refute if you can, but you are just going to come across like a bigger idiot if you do.
Sampling is to take a measured set of data points. How you get your set of data is more of technique, but the data itself is the sample. So a set of numbers like: 1, 3, 5, 7, 9 is a sample. The sample can be the whole thing, IE you measured everything that can be measured, or it can a subset. For example, the numbers 1, 3, 5, 7, 9 are all the odd integers from 1 to 10. That sample is the complete measurement because it didn't miss a data point. However, it is a smaller set of numbers if one is capturing odd numbers from 1 to 100.
Sub sampling:
Taking a sub set of sample numbers from a previous sample set. This is one form of down scaling, that can be done. Meaning, that there are many ways to damn scaling or down sample, but sub sampling is ALWAYS down scaling/sampling. Here is the simplest mathematical example of sub sampling. Take the previous example of odd numbers from 1 to 10. Lets find the prime numbers from that sample.
Sample set of:
1, 3, 5, 7, 9
becomes
3, 5, 7
The 3, 5, 7 is a sub set of the previous set. It made by taking a smaller sample of data points from the previous set. Thus it is SUB SAMPLING. The fact that the sub sampling was derived by taking prime numbers is the technique for the sub sampling in this example.
Super Sampling does not apply to mathematics in the pure sense unless you want to count matrix math. It is used solely for image purposes as an anti aliasing technique. It is a form of interpolation as it relates to images to add or insert data points into the current sample size.
Again, I am going by very strict definitions here, which you are wanting to fudge around with. This is what causes tons of confusion because people like to mix terms as you are doing Jag87.
Interpolation is not the same thing as down sampling or down scaling. Now in process that is meant to both down scale an image and still have good image quality, interpolation can happen AFTER the down scaling. My example in the previous post illustrates just that. However, one can down sample an image without using interpolation to reach a sample set.
This is because the terms are technically the exact opposite has I have shown and proven in very concise and easy to understand examples. Anyone can read wikipedia or merriam-webster definitions of all the terms to see that what I have typed is correct and what you typed was incorrect.
You are applying a broad term technique of what say NVidia does in terms of anti-aliasing "algorithms" and saying flatly that it is all "interpolation." This in incorrect because as stated, it is algorithm(S), as in plural. As in the fact that many things typically go on which can include interpolation.
Perhaps there are different ways to do Super Sampling. ATI's current method is supposedly what I have said; rendering a higher resolution image, and then down scaling it to the appropriate size.
Facebook
Twitter