How much CMRR is a goodly amount? Whilst I love the concepts behind whitlock's bootstrapping to give 85dB even with an imbalance, I would assume that, for vinyl, as long as you can get 30dB that's hum dealt with and any more is just for the enjoyment of it?
This will be one of my next exercises
Yes, the FFT peak forest at my FFTs is telling.
http://www.diyaudio.com/forums/analogue-source/298896-digitizing-vinyl-18.html#post4882181
And where does this 60Hz interference comes from? 😕 (Greece-USA about 6000km apart 😀)
George
Attachments
You know you have something when you get 16.66Hz from some of the rail lines. At your service for unexplained noise... ELF-Signals and Ground Currents
Curiousity piqued from another thread I had a look at the original schematics Dick Burwen did for the Cello Palette (which cello then munged with magic parts). It has an interesting clipping circuit in there. Now this was pro gear and would appear Dick considered overload on op-amps a big no no, but the circuit is relatively simple and elegant (at least to my eyes) .
Now Scott has found that the clipping from ticnpop is not an issue, but given we cannot guarantee that all ADC and their front ends will act this way is it worth looking at in motr depth? My only real concern is if a flat topped click will be harder for the software to spot, but if most of these events are one or two samples long I guess I am over thinking and should just get empirical on the issue.
Now Scott has found that the clipping from ticnpop is not an issue, but given we cannot guarantee that all ADC and their front ends will act this way is it worth looking at in motr depth? My only real concern is if a flat topped click will be harder for the software to spot, but if most of these events are one or two samples long I guess I am over thinking and should just get empirical on the issue.
I have found a few great references on the whole issue of ticks and pops. It turns out that a lot of work has been done in forensic science.
SugarCube
The semi-auto vinyl digitizing engines designed by SweetVinyl are now an official Indiegogo campaign:
https://www.indiegogo.com/projects/sweetvinyl-sugarcube#/
Unlike many crowd-funded vapor products this one has already been shown for a year in prototype form to good reviews, so it seems real enough. As I stated here, if successful the SC-2 model would take the drudge out of much of the process, and perhaps even make it possible for WXYC to begin work on putting their LPs to server. If implemented correctly the 192k/24bit conversion should do justice to most of the vinyl I expect it to see.
An excellent RIAA preamp is still needed, so I am very interested in what you geniuses come up with here! In the meanwhile the station can continue to use their PR&E TX990s. Though not the absolute quietest preamps they do have accurate RIAA eq and superb overload recovery.
Cheers!
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
WXYC Chapel Hill, North Carolina - 89.3 FM
The semi-auto vinyl digitizing engines designed by SweetVinyl are now an official Indiegogo campaign:
https://www.indiegogo.com/projects/sweetvinyl-sugarcube#/
Unlike many crowd-funded vapor products this one has already been shown for a year in prototype form to good reviews, so it seems real enough. As I stated here, if successful the SC-2 model would take the drudge out of much of the process, and perhaps even make it possible for WXYC to begin work on putting their LPs to server. If implemented correctly the 192k/24bit conversion should do justice to most of the vinyl I expect it to see.
An excellent RIAA preamp is still needed, so I am very interested in what you geniuses come up with here! In the meanwhile the station can continue to use their PR&E TX990s. Though not the absolute quietest preamps they do have accurate RIAA eq and superb overload recovery.
Cheers!
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
WXYC Chapel Hill, North Carolina - 89.3 FM
I noticed the veritable stampede of supporters. A miniShark is all that you should need for this and it turns out the all the algorithms are published and in the public domain. The statistical method for identifying ticks is fairly easy I need to read up on the recursive interpolation algorithm the references I found were a little heavy on the maths.
Last edited:
Really, not even an FPGA? Come to think of it folks were very happy with some of the old solutions where they just blanked the pops. They had no way of getting a true delay now it's a no brainer.
Last edited:
I'm sure I heard of one one circa 1980 that used a BBD chip that the output was switched to instead of just blanking.
There was one but there were others, BB's had poor THD. I found even a simple third order spline was not audible.
Gnome Wave Cleaner implements pop removal using algorithms from some AES papers. They look very robust and the source code is a free download.
Ace fine there Scott. A perfect RPi project. Still wish I knew why DeNoiseRT was canned but sure something like this could be turned into real time code by someone who is good with that sort of stuff.
Gnome Wave Cleaner from Redhawk.org
http://www2.ece.ohio-state.edu/~schniter/ee597/handouts/restoration_chapter.pdf
Thank you Scott
George
>Edit:
Godsill, Rayner and Cappe work can be found also in this book (ch.4)
http://link.springer.com/book/10.1007/b117882
http://www2.ece.ohio-state.edu/~schniter/ee597/handouts/restoration_chapter.pdf
Thank you Scott
George
>Edit:
Godsill, Rayner and Cappe work can be found also in this book (ch.4)
http://link.springer.com/book/10.1007/b117882
For some years my vinyl-listening goes like this:
- diy analog linear first stage (rewired tonearm to balenced, john curl style double symmetric j-fets input stage with quite some gain and some lametta)
- ADC is an ancient RME ADI-2 with balanced inputs
- processing the riaa with the software sox with biquad-coefficients provided by Scott here in this forum
- DAC back through the RME-device
I'M overall pleased with the sound. Outstanding is the reproduction of the bass-region, firm, tight and rich at the same time.
I do not fight klicks and pops as they don't distrct me from the essence of the music subjectivly.
I guess the icing would be to perform the upper riaa pole in the analog first stage simply through capacitivly loading it with a good audio grade cap, but I'm not able to perform to calculate the then needed biquad coefficients for it. The overload margin would profit a lot.
Next upgrade would be more modern adc/dac, because subjectivly the ADC-computer-DAC route is not transparent soundwise in my setup when looping analog singals through, even without extra processing them.
Rüdiger
- diy analog linear first stage (rewired tonearm to balenced, john curl style double symmetric j-fets input stage with quite some gain and some lametta)
- ADC is an ancient RME ADI-2 with balanced inputs
- processing the riaa with the software sox with biquad-coefficients provided by Scott here in this forum
- DAC back through the RME-device
I'M overall pleased with the sound. Outstanding is the reproduction of the bass-region, firm, tight and rich at the same time.
I do not fight klicks and pops as they don't distrct me from the essence of the music subjectivly.
I guess the icing would be to perform the upper riaa pole in the analog first stage simply through capacitivly loading it with a good audio grade cap, but I'm not able to perform to calculate the then needed biquad coefficients for it. The overload margin would profit a lot.
Next upgrade would be more modern adc/dac, because subjectivly the ADC-computer-DAC route is not transparent soundwise in my setup when looping analog singals through, even without extra processing them.
Rüdiger
Last edited:
- Status
- Not open for further replies.