33 ms·

I suppose it's a sign of how good her articles and videos are that the only stuff you can find that're wrong are details. In the spirit of being pedantic (don'

I suppose it's a sign of how good her articles and videos are that the only stuff you can find that're wrong are details. In the spirit of being pedantic (don't take too seriously): "There's also beamforming and MIMO" No. Beamforming attempts to increase apparent power by changing parameters. One could just as easily say "moving sender and recipient closer". MIMO is also manipulation of sending and receiving antennae, and therefore irrelevant to the discussion about transmitting through a channel with a fixed noise level. She did make a mistake about the number of orders of magnitude, though.

ec109685 4y ago
Why isn’t beamforming a way of increasing effective bandwidth?
- johnklos 4y ago
  It is! But she's talking about "a channel with a fixed noise level". Making new antennae, changing current antennae, moving them closer, aiming them differently, replacing them with an ethernet cable are all ways of increasing effective bandwidth, but those are outside of what she's talking about.
  - cycomanic 4y ago
    The receiver is part of the channel. So if you do beamforming your power at the receiver is higher, i.e. your SNR is higher. Any other interpretation does not make sense, because it's completely irrelevant to the discussion.
    - johnklos 4y ago
      Sigh. The receiver is part of the channel. Exactly. Changing the parameters of the channel means you're no longer commenting on her statement - you're now making a new discussion. Her statement wasn't wrong, technically or otherwise, just because you can think of ways to increase information by changing parameters of the channel. I can think of ways to increase information, but absolutely none of them, suggested by you or by me, can change or invalidate her statement: "If you want to transfer more information through a channel with a fixed noise-level" Here, since this seems tough for you: "If you want to execute instructions more quickly in a CPU, you can increase the clock speed." You're saying, "but if you put in a new CPU that has higher IPC"...
cycomanic 4y ago
> "There's also beamforming and MIMO" No. Beamforming attempts to increase apparent power by changing parameters. One could just as easily say "moving sender and recipient closer". MIMO is also manipulation of sending and receiving antennae, and therefore irrelevant to the discussion about transmitting through a channel with a fixed noise level. Yes beamforming changes the SNR at the receiver, by shaping the beam to be more "concentrated", i.e. focusing onto the receiver. It is highly relevant because yes by increasing the SNR at your receiver you do increase the throughput (albeit only logarithmically). MIMO is very relevant to the discussion, because instead of increasing bandwidth you increase dimensionality of your channel, which has the same effect (i.e. you increase the term in front of the log in Shannon's formula). I don't know why this should be irrelevant for a channel with fixed noise level (also it's probably more correct to say fixed SNR).

ec109685 4y ago

Why isn’t beamforming a way of increasing effective bandwidth?

johnklos 4y ago
It is! But she's talking about "a channel with a fixed noise level". Making new antennae, changing current antennae, moving them closer, aiming them differently, replacing them with an ethernet cable are all ways of increasing effective bandwidth, but those are outside of what she's talking about.
- cycomanic 4y ago
  The receiver is part of the channel. So if you do beamforming your power at the receiver is higher, i.e. your SNR is higher. Any other interpretation does not make sense, because it's completely irrelevant to the discussion.
  - johnklos 4y ago
    Sigh. The receiver is part of the channel. Exactly. Changing the parameters of the channel means you're no longer commenting on her statement - you're now making a new discussion. Her statement wasn't wrong, technically or otherwise, just because you can think of ways to increase information by changing parameters of the channel. I can think of ways to increase information, but absolutely none of them, suggested by you or by me, can change or invalidate her statement: "If you want to transfer more information through a channel with a fixed noise-level" Here, since this seems tough for you: "If you want to execute instructions more quickly in a CPU, you can increase the clock speed." You're saying, "but if you put in a new CPU that has higher IPC"...

cycomanic 4y ago

> "There's also beamforming and MIMO" No. Beamforming attempts to increase apparent power by changing parameters. One could just as easily say "moving sender and recipient closer". MIMO is also manipulation of sending and receiving antennae, and therefore irrelevant to the discussion about transmitting through a channel with a fixed noise level. Yes beamforming changes the SNR at the receiver, by shaping the beam to be more "concentrated", i.e. focusing onto the receiver. It is highly relevant because yes by increasing the SNR at your receiver you do increase the throughput (albeit only logarithmically). MIMO is very relevant to the discussion, because instead of increasing bandwidth you increase dimensionality of your channel, which has the same effect (i.e. you increase the term in front of the log in Shannon's formula). I don't know why this should be irrelevant for a channel with fixed noise level (also it's probably more correct to say fixed SNR).