Oversampling LPF question

[earlevel]

"Exactly equivalent" is not synonymous with "contains", even in information theory.

And THIS is where we disagree.

Not going to carry on, we can disagree. But I don't understand how you can say that. I'll the analogy I raised earlier:

An AM broadcast is not exactly equivalent to the program material.

This point is important for the person designing the receiver.

If the person designing a receiver (or trying to fix one) needs help understanding, it's important to be able to tell them the details of the modulation and the frequencies it creates. It does not help them to say it's exactly equivalent to the program material.

[mystran]

An AM broadcast is not exactly equivalent to the program material.

Why?

[earlevel]

An AM broadcast is not exactly equivalent to the program material.

Why?

I don't know if you're asking me to explain how AM works...or this is sport...

But we can't fit analog audio into discrete memory without some sort of encoding, and PCM does it. We can't send audio frequencies through the air very far without AM. Are you denying that AM at radio frequencies is not the same as the original audio used to generate it? They look a bit different on a spectrum analyzer, no?

I think you have a unique philosophy, that if something contains the information, it's the same as the information. I don't think Shannon had that philosophy in his papers on information theory nor sampling theory. I'm fine to leave you with it.

[mystran]

An AM broadcast is not exactly equivalent to the program material.

Why?

I don't know if you're asking me to explain how AM works...or this is sport...

Ofcourse I'm NOT asking you to explain how AM works. What I'm asking is why do you think that an AM modulated signal is not equivalent to the original? How can we even look at the spectrum if a time-domain and frequency domain signal are not equivalent as well, or is Fourier transform except from this rule that when you transform a signal into a different representation it becomes something else?

I think the reason I brought up the whole topic of treating sampled values as a continuous-time band-limited signal is that this line of thinking is precisely what originally helped me to gain deeper understanding of the whole resampling thing. At the time, most documents where suggesting something silly like if you wanted to convert from 44.1kHz to 48Khz you should first upsample by a factor of 480 and then downsample by a factor of 441... and then they'd go on to explain how you can carefully optimize this whole process by eliminating all the multiplies by zero and discarded samples.

But.. what if we want to convert to 48001Hz or perhaps 48003.14159..Hz so the factor isn't even rational. The continuous-time extension where we simply treat each sample as a scaled sinc and add them together gives a straight-forward solution: just lookup the values at the positions you want by adding the values of the sincs at those time instants... and once we understand the general case, polyphase filtering for the integer (or power of two or whatever you want) rates is just a simplified special case. What we can't do when converting from one rate to another is construct a dirac train unless the rate happens to be an integer factor... unless ofcourse we scatter fractional delay kernels, but those are precisely the same sincs again. And if one of the sampling rates isn't uniform, then we have to give up sinc too (since it's just the special case for Lagrange-interpolation in case the sampling instant happen to be equally spaced; I think it was around the time of the original post that I was also actively looking for efficient practical solutions for the non-uniform case).

This is infact somewhat of a "sport" for me: when someone asks a question, I intentionally try to (to the best of my ability) to give them a broader view in order to help them gain a deeper understand. I make somewhat of a point of not handing people "copy paste" solutions, because I'm "evil" in a sense: I want people to think and try to understand things, even though I do realize thinking can be quite frustrating at times.

[earlevel]

Ofcourse I'm NOT asking you to explain how AM works. What I'm asking is why do you think that an AM modulated signal is not equivalent to the original? How can we even look at the spectrum if a time-domain and frequency domain signal are not equivalent as well, or is Fourier transform except from this rule that when you transform a signal into a different representation it becomes something else?

Answered—why are you acting like it wasn't?

Convolving with sinc—yet another reference to lowpass filtering.

I never faulted you for educating the OP—I even said you replied well to him. You erroneously faulted my reply to him—saying the sampled signal is no more than the bandlimited signal exactly equivalent. You weren't educating him, he hasn't posted since page 1, you're attacking my explanation.

And the reason I'm not going to discuss this with you is that you completely ignore anything that you're not going to be able to talk down. You won't address that the data is PCM encoded, even through we routinely call it PCM, because PCM has undeniable sidebands. Meaning there are frequencies you can't account for in your view that the samples are just the bandlimited audio (frequency domain). And PCM, by definition, also says the samples are impulses (time domain). If you deny these two points, you deny PCM, so you pretend I didn't bring it up.

So I'm done.

[Z1202]

To the OP's question, you replied with a very reasonable response. Z1202 didn't address the question, but instead faulted details of your response. I responded to the OP's question, addressing neither of your posts. Both you and Z1202 commenced faulting details of my response. (Meanwhile, the OP thanked me for my detailed answer.) That's what's so typical of the way this forum operates. The original question and resulting help is overshadowed by people finding fault with other helpers' answers. It might feel different if I'd said something that was factually incorrect, but it's hard to come away from these threads and still feel clean.

I'll repeat that I admire you guys and your what you've. It still doesn't feel any better to take x amount of time to help someone, only to spend 8x the time mud wrestling about it.

(Sure someone will say I can bail out. But then I've put my knowledge up here semi-permanently only to accompanied by semi-permanent reminders from credible authorities that what I say can't be trusted.)

I haven't re-read the entire thread or read new posts (someone just gave me a headsup to this discussion). But I have to say that I don't feel the need to comment on smth, if other people can do it, and maybe even in a simpler language. However I feel (or rather felt) the need to refine the details which might be misleading, incomplete, POV-specific, or sometimes plain wrong. Even if the OP might not immediately benefit from such fine details, something would hopefully stick, so that they won't have a deceiving impression that they have gotten a complete answer to their question. Also it was my hope that a number of silent readers might benefit from such refined discussions. However this kind of responses apparently sometimes or maybe often is taken as bringing negative value, just bringing in useless complexity and confusion and at times maybe even as personal attacks. Since I don't feel there is a need for me to say anything else (as mentioned, I believe other people can do a better job of it), I feel my contributions to this forum are becoming progressively less welcome. That's why I stopped frequenting it. I'm not telling everybody what they should like or welcome, but if they don't welcome what I can bring, then no point for me to contribute. "So typical KVR", to use your own words.

[Music Engineer]

I feel my contributions to this forum are becoming progressively less welcome. That's why I stopped frequenting it.

That is very sad. I personally highly value your (and mystran's) contributions to this forum. Your posts are always very high quality and I think that I have learned a lot of things from you two guys. By the way, thanks for your second book, too. ...I actually wanted to post this in the book thread but then got distracted. It was like great christmas gift.

[earlevel]

To the OP's question, you replied with a very reasonable response. Z1202 didn't address the question, but instead faulted details of your response. I responded to the OP's question, addressing neither of your posts. Both you and Z1202 commenced faulting details of my response. (Meanwhile, the OP thanked me for my detailed answer.) That's what's so typical of the way this forum operates. The original question and resulting help is overshadowed by people finding fault with other helpers' answers. It might feel different if I'd said something that was factually incorrect, but it's hard to come away from these threads and still feel clean.

I'll repeat that I admire you guys and your what you've. It still doesn't feel any better to take x amount of time to help someone, only to spend 8x the time mud wrestling about it.

(Sure someone will say I can bail out. But then I've put my knowledge up here semi-permanently only to accompanied by semi-permanent reminders from credible authorities that what I say can't be trusted.)

I haven't re-read the entire thread or read new posts (someone just gave me a headsup to this discussion). But I have to say that I don't feel the need to comment on smth, if other people can do it, and maybe even in a simpler language. However I feel (or rather felt) the need to refine the details which might be misleading, incomplete, POV-specific, or sometimes plain wrong. Even if the OP might not immediately benefit from such fine details, something would hopefully stick, so that they won't have a deceiving impression that they have gotten a complete answer to their question. Also it was my hope that a number of silent readers might benefit from such refined discussions. However this kind of responses apparently sometimes or maybe often is taken as bringing negative value, just bringing in useless complexity and confusion and at times maybe even as personal attacks. Since I don't feel there is a need for me to say anything else (as mentioned, I believe other people can do a better job of it), I feel my contributions to this forum are becoming progressively less welcome. That's why I stopped frequenting it. I'm not telling everybody what they should like or welcome, but if they don't welcome what I can bring, then no point for me to contribute. "So typical KVR", to use your own words.

Sorry to drag you into my rant. As I've said, I respect and admire what both you and mystan bring. I just don't like to get involved with kvr threads because they seem to always follow the same path. Looking at page one, I see the OP asking a question, mystran replies, you challenge what he said. I reply to the OP, both you and mystran challenge what I said. And the rest of the thread is that.

To be clear, you didn't say I was wrong, just that what I described "doesn't have that much usefulness" broadly, and that you "personally see little reason for this way of thinking". But, then, why not address an answer to the OP with a description you think might be helpful?

That's a little rhetorical, I understand you like chatting about these technical aspects, which is not a bad thing. And it's not the person aspect, I can laugh that off, but I don't want subsequent readers to not bother reading and understanding what I wrote because a respected person said it doesn't have much usefulness.

It's just a little frustrating to have one person saying that what I say is not useful...and another says it's wrong (and more that I'll let rest). I didn't mean to imply that you played a big part in this, obviously it was very small, just an example of how these threads turn out to not be so much about the original poster. I posted to the OP, I didn't post anything controversial, I even posted a classic DSP text saying the same thing. If someone doesn't think it was useful for the OP, well, that's an opinion, though frustrating someone would say it without themselves trying to help the OP. But it's not wrong, it's the most accurate mathematical model of what happens with sampling—there's a reason we call it PCM.

I don't want to influence how or how much you post here. This was my rant on why I don't visit kvr often. There's useful stuff here, good people here, it just takes too much out of my life.

[earlevel]

Hate when I make an edit, but accidentally quoted instead...

[mystran]

Seriously don't take anything I say personally, I'm not trying to pick a fight or anything, I'm just chatting to pass some time and frankly I do enjoy being a little bit controversial at times, on purpose.

[earlevel]

Seriously don't take anything I say personally, I'm not trying to pick a fight or anything, I'm just chatting to pass some time and frankly I do enjoy being a little bit controversial at times, on purpose.

Thank you, I appreciate this comment very much. Please excuse the rant. I re-entered the old thread looking for specific counters, such as in the mathematical sense, to what I'd proposed before. The reason is that's I'd been "going to do" a video on it for years, and wanted to have some ideas of how to approach it with the knowledge of what kind of reservations people might have about it. It seemed like a good idea at the time, but I shouldn't have expectations, then get frustrated it didn't go down the path I wanted—my bad. I do enjoy reading your posts, and will take things better in the future knowing your outlook. I think a past experience probably attributed to me not rolling with it.

Well, it a little funny (maybe only to me) so I'll tell the story: The reasons for why people insert zeros came up on the music-dsp mailing list, something like 2002 IIRC [found it—2005]. I gave a detailed explanation, one guy told me I didn't know what I was talking about. Exchanges went a few rounds, with me being very polite and deferential, thinking we could chat out the reasoning, and him basically insulting me each time. I finally asked him, ok, then could you please explain the reason for inserting zeros? One total reply: "Exploit serendipity." So, I had mathematical reasons, but for him it was a happy accident. I finally said, you're just telling me my explanation of sampling is wrong, could you please give me yours? He said sure, but was in the process of moving and would answer later. Well, it's 2024...

Funny side note: Some guy with a DSP book he seemed to always be promoting (stuff like, "no, that's incorrect. buy my book to find out how it really works") joined in and said I was wrong. I quoted JOS saying the same thing, he replied he didn't care if "Julius Cesar" [sic] made the same argument, it's still wrong. I replied ok with a link to a JOS page and said it may be "wrong, wrong, wrong" [being ironic—stating his view], "yet clearly helpful to many". JOS popped into the thread and ask me what's wrong about it, to please clarify—LOL, I felt bad that it appeared I was telling him he was wrong.

So I guess I have a history of frustration with this topic. But it's not tragic, I've always used the opportunities to better understand the process so I can explain it better.