Thank you for the amazing video Leslie. I had one question on one of the slides (slide 124 of 165 or time 22:40); I noticed that the zero crossing of each subcarrier don't overlap with the peak of the other subcarrier which means there is ICI (Inter Carrier Interference). I assume the Rx image is just to show attenuation and the ICI issue is a minor image mistake (e.g. Looking at the peak of the orange subcarrier in the Rx image, the purple/red/yellow zero crossing should all overlap at a single point exactly below the orange peak otherwise there is ICI). Please feel free to correct me if I am mistaken.
I’m trying to code my own OFDM solution right now and I’m quite confused by the IFFT and FFT. If I’m using 64 sub carriers, is the IFFT applied to sub carrier 1 then sub carrier 2 etc, or is it applied to the first symbol on each sub carrier, computed, then the second symbol on each sub carrier etc.
Where in the system is the sinc spectrum shape applied? What I'm seeing everywhere is that there is just a set of frequency spikes that represent the data. In my system, I tried to shape my frequency spectrums for each bin like a relatively detailed sinc function, but this of course results in a time domain function that is limited by a rect function, so with a lot of dead signal. I don't see this anywhere, so I have to assume that no OFDM system is really orthogonal. What am I missing here?
I realised my mistake. To avoid the rectangular shape, engineers generally sample the sinc spectrum with frequency steps equal to the distance between zero crossings. That results in the sinc spectrum being expressed as 0 vectors, with 1 vector the same size as the main lobe, which looks just like the singular vector in the video
16QAM translates 4 bits into a constellation point, which is 1 complex number, which describes the amplitude and phase of your frequency component. That 16QAM constellation point is put into 1 single subcarrier. So the n-QAM modulator size does not have any influence on the IFFT and FFT
@@bilalhamdan2287 So let's say you want to use 16-QAM in your system. This means that every subcarrier uses their own 16-QAM modulation. With this modulation, you can send 4 bits per subcarrier. That means that if you have 8 subcarriers, you can send 32 bits per OFDM symbol. For every subcarrier, you can translate its 4 bits with the 16-QAM into a point on something that's called a "constellation diagram". This constellation diagram is a grid of complex numbers centered around the origin. The amount of dots is equal to the Nth-QAM which you are using. So the modulation translates your 4-bit symbol to a complex number, which will be the size of a subcarrier vector. The magnitude of the vector translates to the amplitude of the resulting sine wave, and the argument of the complex number translates to its phase. So as you can see, you only need as many points as you have subcarriers. Modulation does not have anything to do with this.
@@bilalhamdan2287 There are no bits entering the ifft. It's a complex number that is made from the bits using your QAM modulation. You can implement this in your VHDL as a simple 2x1 vector with a real and imaginary value
![แฟนแนวใด๋ - ยูริ โตเกียวมิวสิค [OFFICIAL MUSIC VIDEO ]](http://i.ytimg.com/vi/4_XVMuhhbsU/mqdefault.jpg)
Finally, I found a complete and well-explained video on this topic. Many Thankss
Probably the best explanation of OFDM on TH-cam (and maybe on the internet in general)
I'm trying to implement OFDM with my USRP devices, and this video helps a lot!!! Thanks for the content.
I don't know why I learned OFDM from another person while there is a great explain like this, really thank you very much.
Such a complicated topic explained beautifully, eloquently. Thank you in t- f- spatial and all other domains :)
Thank you for making these videos! you saved me on many tests :,)
Thank you for these videos. You’re awesome!
Thank you so much for this outstanding explanation of an OFDM transmission. You`re saving my thesis!
thank you for creating such a resourceful video. much appreciated.
Tks u a lot Ms Leslie, the video helps to understand the relation of signal in time and frequency domains.
This is the explanation I’ve been looking🙌🏾
this video is great! thanks
thanks!! great content
YOU ARE LEGEND
Thank you very much!
Very clear explanation on OFDM
Thanks you help me a lot
Thank you for the amazing video Leslie. I had one question on one of the slides (slide 124 of 165 or time 22:40); I noticed that the zero crossing of each subcarrier don't overlap with the peak of the other subcarrier which means there is ICI (Inter Carrier Interference). I assume the Rx image is just to show attenuation and the ICI issue is a minor image mistake (e.g. Looking at the peak of the orange subcarrier in the Rx image, the purple/red/yellow zero crossing should all overlap at a single point exactly below the orange peak otherwise there is ICI). Please feel free to correct me if I am mistaken.
Great introduction
I’m trying to code my own OFDM solution right now and I’m quite confused by the IFFT and FFT. If I’m using 64 sub carriers, is the IFFT applied to sub carrier 1 then sub carrier 2 etc, or is it applied to the first symbol on each sub carrier, computed, then the second symbol on each sub carrier etc.
hi dr where can i get the powerpoint presentation from?
Where in the system is the sinc spectrum shape applied? What I'm seeing everywhere is that there is just a set of frequency spikes that represent the data.
In my system, I tried to shape my frequency spectrums for each bin like a relatively detailed sinc function, but this of course results in a time domain function that is limited by a rect function, so with a lot of dead signal. I don't see this anywhere, so I have to assume that no OFDM system is really orthogonal. What am I missing here?
I realised my mistake. To avoid the rectangular shape, engineers generally sample the sinc spectrum with frequency steps equal to the distance between zero crossings. That results in the sinc spectrum being expressed as 0 vectors, with 1 vector the same size as the main lobe, which looks just like the singular vector in the video
Dear Sir, can you explain my wondering. As I know, output of N-IFFT is just in baseband, where do the carrier appears in the overall OFDM system?
The N-FFT is converted from digital to analog using a DAC. The analog output is mixed with an local oscillator at the carrier frequency.
Each s carr is 15khz
So 20000/15 is about 1200 carriers in DTV
Hello mam i want explanation on DFT-S-OFDM. Please reply
Hmm nice try but things are quite complicated then what is presented...but this lays out a very good foundation
and if we are using a 16 QAM modulator then the ifft and fft processors must be of 16 points?
16QAM translates 4 bits into a constellation point, which is 1 complex number, which describes the amplitude and phase of your frequency component. That 16QAM constellation point is put into 1 single subcarrier. So the n-QAM modulator size does not have any influence on the IFFT and FFT
@@Pep95 i have a project in my university about ofdm implementation using vhdl coding and i am very confused about these concepts and the vhdl codes
@@bilalhamdan2287 So let's say you want to use 16-QAM in your system. This means that every subcarrier uses their own 16-QAM modulation. With this modulation, you can send 4 bits per subcarrier. That means that if you have 8 subcarriers, you can send 32 bits per OFDM symbol. For every subcarrier, you can translate its 4 bits with the 16-QAM into a point on something that's called a "constellation diagram". This constellation diagram is a grid of complex numbers centered around the origin. The amount of dots is equal to the Nth-QAM which you are using. So the modulation translates your 4-bit symbol to a complex number, which will be the size of a subcarrier vector. The magnitude of the vector translates to the amplitude of the resulting sine wave, and the argument of the complex number translates to its phase. So as you can see, you only need as many points as you have subcarriers. Modulation does not have anything to do with this.
@@Pep95 okay so how many bits are entering the ifft part?
@@bilalhamdan2287 There are no bits entering the ifft. It's a complex number that is made from the bits using your QAM modulation. You can implement this in your VHDL as a simple 2x1 vector with a real and imaginary value