Re: [Flexradio] IQ correction, future developments
Jim Just thought I'd lighten it up a bit (smile). Thanks for your contributions to the knowledge base and How To. Happy Thanksgiving all! Eric2 -Original Message- From: Jim Lux [mailto:[EMAIL PROTECTED] Sent: Wednesday, November 23, 2005 1:13 AM To: ecellison; 'Mark Amos'; '[EMAIL PROTECTED] Biz' Subject: Re: [Flexradio] IQ correction, future developments At 07:35 PM 11/22/2005, ecellison wrote: >Mark > >Just remember you should always include some almost abosolute statement in >your messages so that Jim Lux can catch you! (smile) You bet... What are amateurs for if not to push the envelope when someone says something like "nobody will ever communicate across the ocean with wavelengths shorter than 200m" >Neat stuff! > >Eric2
Re: [Flexradio] IQ correction, future developments
At 07:35 PM 11/22/2005, ecellison wrote: Mark Just remember you should always include some almost abosolute statement in your messages so that Jim Lux can catch you! (smile) You bet... What are amateurs for if not to push the envelope when someone says something like "nobody will ever communicate across the ocean with wavelengths shorter than 200m" Neat stuff! Eric2
Re: [Flexradio] IQ correction, future developments
Mark Just remember you should always include some almost abosolute statement in your messages so that Jim Lux can catch you! (smile) Neat stuff! Eric2 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Mark Amos Sent: Tuesday, November 22, 2005 7:37 PM To: '[EMAIL PROTECTED] Biz' Subject: Re: [Flexradio] IQ correction, future developments I don't know about the rest of the lurkers on this list, but I find this kind of thing really inspiring. I hope there's something, some time, somehow that I can contribute to this process, but in the mean time, I'll have to be content to just be amazed and inspired by you guys. Mark -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Robert McGwier Sent: Tuesday, November 22, 2005 9:56 AM To: [EMAIL PROTECTED] Biz Subject: [Flexradio] IQ correction, future developments The correct way to do IQ correction is to do it with a pseudo-inversion filter in the frequency domain and as a part of (convolved with) the bandpass filter. When we are finally done, the frequency domain computations will be a really huge thing in the radio (as if they weren't already). We will do block operations for noise reduction, notch filtering, IQ correction, etc. etc. It is going to be very loaded. I have recently made a very large improvement in the efficiency of the dsp. I managed to get SSE and floating point and version 3.0.1 of FFTW to work under WINBLOWS and Visual.NET and our code. This means that we will be using the SSE capabilities to do filtering and soon after the cordic oscillators. The code for dsp has mostly stabilized. The release of the SSE.FFTW float version of the code will be major step forward. Eric has a prototype DirectX version of the video routines running (I can run it here). With stability will come speed and optimization work. Richard Allen has already reported a large increase in the performance of the timers for the keyer. Expect all of this to be the theme for the next several months. I sent Frank and Edson a MASSIVE upgrade to the linux code, incorporating all of the new features that are in the windows code and already implementing the FFTW 3.0.1/SSE version of the code. We will test it there first. Frank and I get together next week and we hope to do a major upgrade to the version that is available via cvs on sourceforge. As Willi is fond of saying, every time you download it is a new radio. I know that some are not accustomed to this and INCORRECTLY view this (IMHO) as "it isn't ready". Bull. It is never going to be "ready" if what you mean by that is it will stabilize and never changed. VAC is not perfect. WDM-KS is not perfect. But it is a new capability and it will get better. We will like replace VAC since we need for that channel to support WDM-KS and we need to be able to use SSE versions of the resampling code so all of that will fly. Dale has agreed to host SVN so we can ALL contribute bits of code and not step on each other. Stay tuned. These features will be at the heart of 1.5.0 when released early next year. Bob -- Laziness is the number one inspiration for ingenuity. Guilty as charged! ___ FlexRadio mailing list FlexRadio@flex-radio.biz http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz ___ FlexRadio mailing list FlexRadio@flex-radio.biz http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz
Re: [Flexradio] IQ correction, future developments
Ditto -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Mark Amos Sent: Tuesday, November 22, 2005 7:37 PM To: '[EMAIL PROTECTED] Biz' Subject: Re: [Flexradio] IQ correction, future developments I don't know about the rest of the lurkers on this list, but I find this kind of thing really inspiring. I hope there's something, some time, somehow that I can contribute to this process, but in the mean time, I'll have to be content to just be amazed and inspired by you guys. Mark -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Robert McGwier Sent: Tuesday, November 22, 2005 9:56 AM To: [EMAIL PROTECTED] Biz Subject: [Flexradio] IQ correction, future developments The correct way to do IQ correction is to do it with a pseudo-inversion filter in the frequency domain and as a part of (convolved with) the bandpass filter. When we are finally done, the frequency domain computations will be a really huge thing in the radio (as if they weren't already). We will do block operations for noise reduction, notch filtering, IQ correction, etc. etc. It is going to be very loaded. I have recently made a very large improvement in the efficiency of the dsp. I managed to get SSE and floating point and version 3.0.1 of FFTW to work under WINBLOWS and Visual.NET and our code. This means that we will be using the SSE capabilities to do filtering and soon after the cordic oscillators. The code for dsp has mostly stabilized. The release of the SSE.FFTW float version of the code will be major step forward. Eric has a prototype DirectX version of the video routines running (I can run it here). With stability will come speed and optimization work. Richard Allen has already reported a large increase in the performance of the timers for the keyer. Expect all of this to be the theme for the next several months. I sent Frank and Edson a MASSIVE upgrade to the linux code, incorporating all of the new features that are in the windows code and already implementing the FFTW 3.0.1/SSE version of the code. We will test it there first. Frank and I get together next week and we hope to do a major upgrade to the version that is available via cvs on sourceforge. As Willi is fond of saying, every time you download it is a new radio. I know that some are not accustomed to this and INCORRECTLY view this (IMHO) as "it isn't ready". Bull. It is never going to be "ready" if what you mean by that is it will stabilize and never changed. VAC is not perfect. WDM-KS is not perfect. But it is a new capability and it will get better. We will like replace VAC since we need for that channel to support WDM-KS and we need to be able to use SSE versions of the resampling code so all of that will fly. Dale has agreed to host SVN so we can ALL contribute bits of code and not step on each other. Stay tuned. These features will be at the heart of 1.5.0 when released early next year. Bob -- Laziness is the number one inspiration for ingenuity. Guilty as charged! ___ FlexRadio mailing list FlexRadio@flex-radio.biz http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz ___ FlexRadio mailing list FlexRadio@flex-radio.biz http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz
Re: [Flexradio] IQ correction, future developments
I don't know about the rest of the lurkers on this list, but I find this kind of thing really inspiring. I hope there's something, some time, somehow that I can contribute to this process, but in the mean time, I'll have to be content to just be amazed and inspired by you guys. Mark -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Robert McGwier Sent: Tuesday, November 22, 2005 9:56 AM To: [EMAIL PROTECTED] Biz Subject: [Flexradio] IQ correction, future developments The correct way to do IQ correction is to do it with a pseudo-inversion filter in the frequency domain and as a part of (convolved with) the bandpass filter. When we are finally done, the frequency domain computations will be a really huge thing in the radio (as if they weren't already). We will do block operations for noise reduction, notch filtering, IQ correction, etc. etc. It is going to be very loaded. I have recently made a very large improvement in the efficiency of the dsp. I managed to get SSE and floating point and version 3.0.1 of FFTW to work under WINBLOWS and Visual.NET and our code. This means that we will be using the SSE capabilities to do filtering and soon after the cordic oscillators. The code for dsp has mostly stabilized. The release of the SSE.FFTW float version of the code will be major step forward. Eric has a prototype DirectX version of the video routines running (I can run it here). With stability will come speed and optimization work. Richard Allen has already reported a large increase in the performance of the timers for the keyer. Expect all of this to be the theme for the next several months. I sent Frank and Edson a MASSIVE upgrade to the linux code, incorporating all of the new features that are in the windows code and already implementing the FFTW 3.0.1/SSE version of the code. We will test it there first. Frank and I get together next week and we hope to do a major upgrade to the version that is available via cvs on sourceforge. As Willi is fond of saying, every time you download it is a new radio. I know that some are not accustomed to this and INCORRECTLY view this (IMHO) as "it isn't ready". Bull. It is never going to be "ready" if what you mean by that is it will stabilize and never changed. VAC is not perfect. WDM-KS is not perfect. But it is a new capability and it will get better. We will like replace VAC since we need for that channel to support WDM-KS and we need to be able to use SSE versions of the resampling code so all of that will fly. Dale has agreed to host SVN so we can ALL contribute bits of code and not step on each other. Stay tuned. These features will be at the heart of 1.5.0 when released early next year. Bob -- Laziness is the number one inspiration for ingenuity. Guilty as charged! ___ FlexRadio mailing list FlexRadio@flex-radio.biz http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz
[Flexradio] IQ correction, future developments
The correct way to do IQ correction is to do it with a pseudo-inversion filter in the frequency domain and as a part of (convolved with) the bandpass filter. When we are finally done, the frequency domain computations will be a really huge thing in the radio (as if they weren't already). We will do block operations for noise reduction, notch filtering, IQ correction, etc. etc. It is going to be very loaded. I have recently made a very large improvement in the efficiency of the dsp. I managed to get SSE and floating point and version 3.0.1 of FFTW to work under WINBLOWS and Visual.NET and our code. This means that we will be using the SSE capabilities to do filtering and soon after the cordic oscillators. The code for dsp has mostly stabilized. The release of the SSE.FFTW float version of the code will be major step forward. Eric has a prototype DirectX version of the video routines running (I can run it here). With stability will come speed and optimization work. Richard Allen has already reported a large increase in the performance of the timers for the keyer. Expect all of this to be the theme for the next several months. I sent Frank and Edson a MASSIVE upgrade to the linux code, incorporating all of the new features that are in the windows code and already implementing the FFTW 3.0.1/SSE version of the code. We will test it there first. Frank and I get together next week and we hope to do a major upgrade to the version that is available via cvs on sourceforge. As Willi is fond of saying, every time you download it is a new radio. I know that some are not accustomed to this and INCORRECTLY view this (IMHO) as "it isn't ready". Bull. It is never going to be "ready" if what you mean by that is it will stabilize and never changed. VAC is not perfect. WDM-KS is not perfect. But it is a new capability and it will get better. We will like replace VAC since we need for that channel to support WDM-KS and we need to be able to use SSE versions of the resampling code so all of that will fly. Dale has agreed to host SVN so we can ALL contribute bits of code and not step on each other. Stay tuned. These features will be at the heart of 1.5.0 when released early next year. Bob -- Laziness is the number one inspiration for ingenuity. Guilty as charged!
