The ultimate rumble filter - far more effective than just a high pass filter!

[DF96]

I have just estimated the stray capacitance you might need to upset stereo width. I was surprised at the result.

Assuming an output impedance of 1k, and cross-talk of -60dB at 20kHz, 8pF would do it. Some switches could have capacitance in that region. Reducing the output impedance would solve the problem. -60dB crosstalk ought to be measurable, although perhaps not with just an oscilloscope. Of course, a pickup cartridge might only have -40dB separation anyway.

So be careful how you introduce a mono switch. It can be done, but like everything else in audio electronics it has to be done properly which probably means doing some sums first.

[andyr]

Originally Posted by DF96

So be careful how you introduce a mono switch. It can be done, but like everything else in audio electronics it has to be done properly which probably means doing some sums first.

Correct, DF96 - my point was simply to illustrate that at these low signal levels, anything superfluous will probably degrade the signal (compared to it not being there) ... so this "super rumble filter" will undoubtedly do so. And having to decide whether you need to plug it in, or out, of your signal chain, record by record, will spoil the enjoyment of listening.

So, yes - great for a specific task - digitising old LPs ... but not otherwise.

Regards,

Andy

[monty78pig]

Andyr - I challenge you to build and ABX test the circuit on your setup if you believe it will produce noticeable artifacts! Unless you believe that the ABX switchbox will generate some horrific 'multi phasic frequency inhibited shifts'...

Consider the following facts:

• Even the very best phono cartridges have 30dB of cross channel rejection.
• Your speakers and room will not have more than 20dB.
• 60dB of cross channel leakage will not be perceptible at 20KHz, or indeed at any frequency, if you don't believe me then I can send you a few couple of samples (one without the leakage and one with it)! You probably heard the effects of the series resistors causing changes to the cartridge loading, by putting a series resistance in front of the capacitive load in the preamp (if you actually heard anything).

DF96, what about the capacitance between channels in the cartridge, it's usually a lot greater than 8pF. If we are going to play the maths game then let me say that at 20KHz my circuit has a theoretical cross channel rejection of 84dB, at 10KHz 72dB, at 1KHz 32dB. Don't forget than in order to achieve 32dB cross channel rejection with the IDEAL cartridge the stylus has to be no more than 1.4 degrees to the perpendicular of the playing surface at any time .



Don't forget, that there is NO capacitive coupling between channels (and using a mono switch does NOT hurt channel rejection unless you put a cap across it). Measurements can pick up EVERYTHING, to say otherwise is simply operating outside the realms of reality . This circuit is designed to work at line level, not at the ultra low MM/MC levels where loading really matters.


This circuit does nothing to degrade the signal and theres no point in trying to save the separation in the bass that isn't there on any vinyl releases. Also good quality digital audio is far superior vinyl .

[DF96]

I wasn't 'playing maths games', merely doing what all engineers do: estimating whether a claim is plausible. 1k is a reasonable estimate of source impedance. -60dB was probably too low, because of other issues (one of which I mentioned).

I have no idea what the interchannel capacitance is in a cartridge, but simple internal screening could ensure that it stays low. My conclusion that a claim is plausible does not necessarily mean that I accept it; it merely means that I am prepared to consider it.

[monty78pig]

Sorry DF96 if I seemed a little sharp there, I was merely a little exasperated with andyr's somewhat pseudoscientific approach to this concept . What I really meant was that the effects of stray switch capacitance can (generally) be considered irrelevant because with a decent toggle switch in direct contact with the chassis, the capacitance is less than negligible. I was simply comparing using the same methods, the channel rejection of my circuit in contrast to the set of conditions that you described.

[Pano]

Originally Posted by monty78pig

I don't think I have any problem with tonearm resonance as the LF noise doesn't seem to peak anywhere on an FFT on most recordings, it seems rather flat (within it's limits) to me.

Thanks, I was just wondering because I've had problems with that in the past. It really made the woofers jiggle! But it you aren't seeing a peak on the FTT, it's likely not the problem.

Originally Posted by andyr

..a Project Debut 3 with a Shure M97XE is at the low end of the vinyl delivery scale.

And yet that cart is well liked in blind tests. Go figure.

[andyr]

Originally Posted by monty78pig

Sorry DF96 if I seemed a little sharp there, I was merely a little exasperated with andyr's somewhat pseudoscientific approach to this concept

I can't see how you can call my approach "pseudoscientific" ... as all I said was that, based on an experiment I had done (putting a mono switch at the output of my phono stage - which had a deleterious effect on the signal), I thought it was likely your rumble filter would degrade the sound, compared to it not being there.

But in the context of your original post - which was that you built the filter to assist you when you digitise old recordings - it undoubtedly does an excellent job ... and the signal "quality" on these old recordings is probably not to a level of quality such that you can hear the degradation caused by the circuit. But I wouldn't use the filter for good LPs.

Don't get me wrong - I am in awe of your expertise in designing that circuit ... I just wouldn't keep it in place for all your LP listening.

Regards,

Andy

[monty78pig]

Andy, I already use a software implementation when doing restoration. What I plan to use this circuit for is live playback for all discs as the resultant intermodulation that even a little rumble causes in the loudspeakers and headphones will be far outweighed by any 'degredation' caused by the circuit, using even the very best setups. This intermodulation is highly undesirable and, even with very small amounts of rumble, causes the stereo image to 'wobble' (in my experience with several setups), anything that can reduce it is a good thing . I stated this in my original post.

I just can't see any valid reason to not use this circuit in the signal chain. I would reconsider this if it caused any deterioration to the audio present in the disc, but with my experiments in audacity, subtracting the output of the process from the input show that all that is missing is just rumble and distortion - all undesirables. None of what is removed has any musical relevance to the content that was originally cut to the disc.

Further research into mastering practice confirms that mastering engineers deliberately removed the separation at lower frequencies for vinyl mastering to limit vertical excursion, they also use subsonic filters when cutting (and a digital delay too after about 1978 to optimise groove pitch between loud and soft passages). This makes my filter circuit a completely transparent and beneficial process, all that is rejected is substantial noise, seeing as surface noise is mainly composed of lower frequencies, my filter brings down this surface noise at these frequencies admirably too.

You are correct that saying that everything causes some deterioration to the audio signal, but the effects of this circuit are well, well below any of the effects that the rest of even the very best hifi will cause. I cannot overemphasise the lack of change to the original audio recorded to the disc that this filter will cause .

I referred to your approach as pseudoscientific for a couple of reasons, the main being that the series resistors that you used caused an adverse effect to the cartridge loading (I am 100% certain that this was the cause of the problem), if it was a switch on it's own and you blind tested it, then I would take your findings seriously, but there were too many psychological/other factors present for your experiment to be valid. All the experiments I have performed while tweaking my rolloff values have been performed double blind for best results.

If you know any technical reason why my circuit will cause deterioration to the sound quality, then please let me know! No hard feelings, though eh?

[monty78pig]

Also, with the 'everything causes deterioration' approach, one might as well argue that the RIAA curve should be abolished . Having restored acoustic recordings in the past, I can tell you that you definitely don't want to go down this route! This circuit works on similar principles!

[sreten]

Hi,

IMO what you want is a Cauer or elliptical filter, its far simpler
if you want something over a simple 12dB/18dB highpass filter.

3rd order high pass Cauer can be done around a single op-amp.
Basic 6dB high pass followed by a 2nd order notch filter, it can
emulate far higher order filters in the initial stopband depending
on the notch Q, and you match the notch Q to the problem.

Its far more elegant than brute force high order highpass,
especially when the high order high pass function is not
needed below the problem area you are addressing.

sreten.