Excellent lecture and Well done. You demonstrated and explained in an exhaustive and best way possible how to use the tool and also explained the theory behind everything you described. Hope you have more such lectures as you said at the end of this exhaustive lesson.
Thank you for your support. You will find sequels by my colleagues and myself in the 2020 European GNU Radio Days tutorials, including ruclips.net/video/tAVqLL2E4hI/видео.html and ruclips.net/video/Zu0NgIfXn4c/видео.html
The numbers for fs and LO at 1:13:11 are switched, I believe. Local oscillator for tuning is between 50MHz and 1600MHz, the sampling frequency (fs) is between 1 MS/s and 2.4 MS/s (for a cheap rtl dongle). Very nice explanation on time domain vs frequency domain in filtering!
Indeed it is my belief that GNU Radio is an invaluable teaching tool with hands on experiments on sometimes plainfully abstract concepts of discrete time digital signal processing. For more teaching material on GNU Radio for teaching basic concepts of digital communications: jmfriedt.free.fr/#L3
Very interesting video. It's the best one I have seen discussing these topics. A question did occur. Can GNU radio be used for implementations of Direction Finding methods?
Thank you for the positive feedback. GNU Radio is only a library easing prototyping of discrete time signal processing algorithms, DoA being a phase analysis of the signal being received by multiple separate antennas connected to a coherent (same local oscillator) receiver, GNU Radio is indeed well suited. See Balint Seeber's past presentations or T. Collins' @ ruclips.net/video/_UBPVi1vp2s/видео.html
I would really like to understand this whole session like a real English speaker so much that see explains well, prof J.M Fred. Bamako, National School of Engineering
Jean-Michel you are a light house . Thank so much. If you have any video of raw data acquisition from a SDR (hackrf foe example ) with GNU Radio, please leave a link here.
If running the flowgraph through Python is not limiting your bandwidth, then saving raw data is as simple as connecting a Source (OsmoSDR, USRP) to a File Sink. If multiple streams are to be interleaved, then Sources -> Interleave -> File Sink. The drawback of this approach is 1/ the scheduler might limit the bandwidth and 2/ the floating point number storage will increase the file size (8 bytes/complex number), which might be an issue when saving high datarates to RAMdisk. You will find an example at github.com/jmfriedt/sentinel1_pbr/ of streaming 8-bit I/Q data from 2 channels to a file saved as 2-bytes/sample (I and Q), optimizing storage size and bandwidth, but not using GNU Radio (raw UHD access).
@@kmoraitakis40 I realize I was mistaken when mentioning interleaving in the source to file program I mentioned above. *Either* we are using UHD source and sinking straight to file without relying on GNU Radio (my example, inspired from github.com/EttusResearch/uhd/blob/master/host/examples) *or* using GNU Radio with all its processing functions with datastream organized by the scheduler and the associated overhead: in this case gr-uhd must be used as shown at github.com/gnuradio/gnuradio/tree/master/gr-uhd/examples/c%2B%2B Apologies for the confusion.
Tutorials 2 and 3 should be uploaded prior to the actual conference to be held June 24th, 2021. The tutorials are mostly independent and the numbering only follows the program schedule: tutorial 4 can be tackled before 2 and 3 are uploaded. Best, JM
could you point to the timestamp in the video you are referring to? I would not dare compare GPS and FM broadcast signals, so I must have misrepresented my thoughts in the discussion, apologies for the confusion, I hope I can clarify what I meant if pointed to the right part of the presentation.
@@reddamasio OK I understand the confusion. I am addressing here the "ideal" SDR architecture -- connecting an antenna straight to the ADC input -- with the practical SDR architecture of antenna-mixer for frequency transposition-low pass filtering-ADC input. I am justifying why the "SDR dream" of sampling the whole RF and microwave band with a single ADC might not be wise due to the limited dynamic range, since the weak GPS signal (as an example) would be overwhelmed by the strong FM band signal when sampling from DC to 1.57542 GHz (hence including the 88-108 MHz FM band), whereas transposing the GPS band to baseband by mixing with 1575.42 MHz, then low pass filtering (eg. 1.023 MHz low pass filter for GPS L1 but anything that will remove the frequency components MHz to 10s of MHz from the GPS frequency) and then sampling will cancel strong signals and compensate for the limited quantization capability of SDR. So ADC quantization limits the dynamic range between the strongest and weakest signal in the sampled band, and justifies the practical architecture of transposing-low pass-sampling as opposed to the most flexible approach of sampling the antenna output. Anyway an antenna acts as a bandpass filter and would not allow all signals from DC to 1.57542 GHz from being sampled, but this is not addressed in this discussion (except if using atoms as antennas as described in the fascinating www.nist.gov/publications/multi-band-rydberg-atom-based-receiverantenna-amfm-stereo-reception but that is well beyond this discussion)
Best GRC tutorial I've seen - clear, concise, real and SDR world, and maths - will be checking out the other sessions for sure.
Jean-Michel's lectures are amazing!
Excellent lecture and Well done. You demonstrated and explained in an exhaustive and best way possible how to use the tool and also explained the theory behind everything you described. Hope you have more such lectures as you said at the end of this exhaustive lesson.
Thank you for your support. You will find sequels by my colleagues and myself in the 2020 European GNU Radio Days tutorials, including ruclips.net/video/tAVqLL2E4hI/видео.html and ruclips.net/video/Zu0NgIfXn4c/видео.html
le tutoriel est juste parfait pour commencer à prendre en main GNU radio
thank u so much I've been looking for such content
Thanks for punishing one of the only GNU radio tutorials that ties the software blocks to real RF theory. Fantastic
Thank you, I love RF and electronics
Merci Monsieur j'ai appris Beaucoup avec vous.
excellent , thanks alot for your effort and wonderful explanations
The numbers for fs and LO at 1:13:11 are switched, I believe. Local oscillator for tuning is between 50MHz and 1600MHz, the sampling frequency (fs) is between 1 MS/s and 2.4 MS/s (for a cheap rtl dongle).
Very nice explanation on time domain vs frequency domain in filtering!
Absolutely, thank you for correcting, I should have been more careful when proofreading the video.
Thank you very much for the video!!!
Superb explanation
Great starting lecture. My professors should use GNU Radio for teaching.
Indeed it is my belief that GNU Radio is an invaluable teaching tool with hands on experiments on sometimes plainfully abstract concepts of discrete time digital signal processing. For more teaching material on GNU Radio for teaching basic concepts of digital communications: jmfriedt.free.fr/#L3
Very interesting video. It's the best one I have seen discussing these topics. A question did occur. Can GNU radio be used for implementations of Direction Finding methods?
Thank you for the positive feedback. GNU Radio is only a library easing prototyping of discrete time signal processing algorithms, DoA being a phase analysis of the signal being received by multiple separate antennas connected to a coherent (same local oscillator) receiver, GNU Radio is indeed well suited. See Balint Seeber's past presentations or T. Collins' @ ruclips.net/video/_UBPVi1vp2s/видео.html
I would really like to understand this whole session like a real English speaker so much that see explains well, prof J.M Fred.
Bamako, National School of Engineering
Jean-Michel you are a light house . Thank so much. If you have any video of raw data acquisition from a SDR (hackrf foe example ) with GNU Radio, please leave a link here.
If running the flowgraph through Python is not limiting your bandwidth, then saving raw data is as simple as connecting a Source (OsmoSDR, USRP) to a File Sink. If multiple streams are to be interleaved, then Sources -> Interleave -> File Sink. The drawback of this approach is 1/ the scheduler might limit the bandwidth and 2/ the floating point number storage will increase the file size (8 bytes/complex number), which might be an issue when saving high datarates to RAMdisk. You will find an example at github.com/jmfriedt/sentinel1_pbr/ of streaming 8-bit I/Q data from 2 channels to a file saved as 2-bytes/sample (I and Q), optimizing storage size and bandwidth, but not using GNU Radio (raw UHD access).
@@europeangnuradiodays1445 Thanks for your reply. Is there any sourch that combines DSP and various methods with GNU radio?
@@kmoraitakis40 I realize I was mistaken when mentioning interleaving in the source to file program I mentioned above. *Either* we are using UHD source and sinking straight to file without relying on GNU Radio (my example, inspired from github.com/EttusResearch/uhd/blob/master/host/examples) *or* using GNU Radio with all its processing functions with datastream organized by the scheduler and the associated overhead: in this case gr-uhd must be used as shown at github.com/gnuradio/gnuradio/tree/master/gr-uhd/examples/c%2B%2B
Apologies for the confusion.
Thank you for your magnificent efforts
I cant seem to find tutorials 2 and 3 eventhough tutorial 4 is there?
Tutorials 2 and 3 should be uploaded prior to the actual conference to be held June 24th, 2021. The tutorials are mostly independent and the numbering only follows the program schedule: tutorial 4 can be tackled before 2 and 3 are uploaded.
Best, JM
Hey man, can you explain a little better what you wanted to say about the amplitude of GPS signal using FM band? Couldn't understant it...thanks
could you point to the timestamp in the video you are referring to? I would not dare compare GPS and FM broadcast signals, so I must have misrepresented my thoughts in the discussion, apologies for the confusion, I hope I can clarify what I meant if pointed to the right part of the presentation.
@@europeangnuradiodays1445 04:08
@@reddamasio OK I understand the confusion. I am addressing here the "ideal" SDR architecture -- connecting an antenna straight to the ADC input -- with the practical SDR architecture of antenna-mixer for frequency transposition-low pass filtering-ADC input. I am justifying why the "SDR dream" of sampling the whole RF and microwave band with a single ADC might not be wise due to the limited dynamic range, since the weak GPS signal (as an example) would be overwhelmed by the strong FM band signal when sampling from DC to 1.57542 GHz (hence including the 88-108 MHz FM band), whereas transposing the GPS band to baseband by mixing with 1575.42 MHz, then low pass filtering (eg. 1.023 MHz low pass filter for GPS L1 but anything that will remove the frequency components MHz to 10s of MHz from the GPS frequency) and then sampling will cancel strong signals and compensate for the limited quantization capability of SDR. So ADC quantization limits the dynamic range between the strongest and weakest signal in the sampled band, and justifies the practical architecture of transposing-low pass-sampling as opposed to the most flexible approach of sampling the antenna output. Anyway an antenna acts as a bandpass filter and would not allow all signals from DC to 1.57542 GHz from being sampled, but this is not addressed in this discussion (except if using atoms as antennas as described in the fascinating www.nist.gov/publications/multi-band-rydberg-atom-based-receiverantenna-amfm-stereo-reception but that is well beyond this discussion)