From: Patent Tactics, George Brock-Nannestad


there is now a kink in the thread.

It is not a question of hearing clicks but of separating clicks from the
signal before hearing.

The anti-alias filter for 44.1 kHz has a useful energy bandwidth to 20 kHz,
which means that it is very steep - multipole. It has to reduce signal
components at 22.05 kHz to below LSB. This filter smears any high frequency
energy, such as that in a click, over a wide frequency band below 20 kHz. The
higher sampling rates can and do use much less steep filters.

The Shure M-44 was standard in transfer work for many years, because it would
tolerate a high stylus pressure without bottoming, but obviously an Ortofon C-
head would do a much better job.

The reason that clicks have such a lot of high frequency energy is that
magnetic and dynamic pickups are velocity detecting devices. A small
protruberance gives off a positive-going pulse immediately followed by a
negative-going pulse, a fact that explains why many preamplifiers do not have
sufficient headroom, adding distortion and feedback delay to the pulse

I think that I must add that I know one archive in which a technician would
transfer to DAT at such a high level that noise pulses consistently went into
"the red". He claimed that he was using the converter as a limiter, thereby
limiting the effect of the noise pulse. Oh, recovery!

Kind regards,



Don Cox wrote:

> On 25/06/05, Goran Finnberg wrote:
> > Don Cox:
> >
> >> The more detailed recording makes it easier for declicking software
> >> to distinguish between clicks and music.
> >
> > With a Shure M44 as a source?
> >
> > So what is the source giving us above 20 kHz??
> If you were seriously digitizing from LPs, you would look for a
> cartridge with a frequency response extending beyond 20KHz.
> For example, this page shows the Audio-Technika range (the first page
> Google came up with). One cartridge has a response up to 50KHz. No use
> for the music, but ideal for detecting clicks.
> Why use an M44?
> Regards
> --
> Don Cox
> [log in to unmask]