RSS feed reader

98 Guy wrote in :

One secure method immediately suggests itself if a router exists.
Set the computer to always go to the router's IP, then set the
router to point to the external DNS servers.

That is DNS relaying, and it's just plain stupid.

Because unless you're always accessing your router's administrative
settings via http (and how many people do that) - then you'll never know
if some malware went in and changed the DNS settings in the router.


Well, I wouldn't do it (I actually set the router only as a THIRD option in
case the two external DNS servers failed. (The router is using one of the
ISP's DNS servers, but I was using DirectNIC's as I'd got a domain with them
till recently)).

There are two ways to hide: in the shadows, or in plain sight. Using a router
might suit those who prefer to use the shadows to remain out of easy reach,
but the majority would likely prefer plain sight, because it's easier.
Neither method is inherently stupid, unless people don't make a choice, stick
with it, and be vigilant, What IS stupid is to trust the machine too much.

Lostgallifreyan wrote:
"Bill in Co" wrote in
:

But I think I've heard that said about noise reduction software, too.
Yes, theoretically that is essentially what is automatically being done
by the noise reduction software or plug-ins, but still, I think the
computer algorithms can often analyze and process it more effectively
than me. Otherwise, there wouldn't be much use for noise reduction tools
(including such plug-ins), unless you solely believe it just for
amateurs, who don't want to invest the time doing it with a multiple
band equalizer, which I don't believe is the only reason. :-)


In the case of NR, it's a contant print made with fine detail, to remove
periodic or constant noise whose variations are on extremely short time
scales. That is best done automatically, but FSE (and more importantly the
underlying filter process) need human input. FSE isn't just frequency
domain,
it is also time domain, otherwise it would be no more than a static EQ.

True enough. And I really appreciate it's general effectiveness, in so many
cases. In fact, if I were isolated on an island somewhere, with a choice
of only a couple of audio tools, I'd have to pick at least one with FSE
capability, in addition to Sound Forge (with some plug-ins!). But I agree
with you that editing in Sound Forge is truly a pleasure - when compared to
so many other audio editors. I haven't found any other audio editor I'd
rather work with.

The main reason for FSE is the removal of transients that affect only part
of
the spectrum, and stand some chance of clean reduction if we can isolate
them. This can only be as good as the underlying filter. So long as I can
get
control of a filter like that I don't miind if I can't see the
time/frequency
plot, time is enough, with the second dimension being used to plot a
simple
curve or two, such as cutoff frequency and slope. Many filters set those
parameters statically, but my best shot is likely to be finding one that
lets
me draw dynamic curves the same way that some Cakewalk and SoundForge
plugins
do it. A good dynamic filter will have some phase offset correction I
hope. I
have fixed problems in bass signals with a static EQ and found a delay
that I
had to correct before merging the fixed sound with the original in a
cross-
fade. I don't need full FSE, I just need some dynamic filter that lets me
avoid some of the tedium that results from making up for static methods'
deficiencies.

And that's the tedium I wish to avoid. :-)

(I wonder if FSE might lead to many newcomers having no awareness of
filters and their methods and limits. If so, it may lead to unrealistic
expectations because any audio cleaner is only as good as its filters).

Probably, and not only newcomers; me too, I'd bet. Obviously, the one more
works with any tool, the better one gets at it, and that includes
discovering more and more of its nuances ... along the path to
enlightenment. :-)

none wrote:
"98 Guy" wrote in message ...
Lostgallifreyan wrote:

For example, if I open an instance of FireFox, that instance will be
using what-ever version of the hosts file that existed at the time
firefox was started. If I edit or delete the hosts file while that
instance is running, it has no effect on that instance. Firefox
continues to operate as if the original hosts file is still present.
Even if I open more tabs and browse to new sites, the pre-existing hosts
file is still in effect. If I keep this instance of FF running and open
a new instance, the new instance will behave according to what-ever
changes I've made to the hosts file.

So it seems to be that each application has it's own version of the
hosts file cached for it somewhere in memory, which is loaded at the
time the application is started.


Been a l-o-n-g time since viewing or writing to this NG. Good to see it
active and you guys writing about something worthwhile and interesting
(unlike many other NG's).

Indeed!

none wrote:
"98 Guy" wrote in message ...
Lostgallifreyan wrote:

For example, if I open an instance of FireFox, that instance will be
using what-ever version of the hosts file that existed at the time
firefox was started. If I edit or delete the hosts file while that
instance is running, it has no effect on that instance. Firefox
continues to operate as if the original hosts file is still present.
Even if I open more tabs and browse to new sites, the pre-existing hosts
file is still in effect. If I keep this instance of FF running and open
a new instance, the new instance will behave according to what-ever
changes I've made to the hosts file.

So it seems to be that each application has it's own version of the
hosts file cached for it somewhere in memory, which is loaded at the
time the application is started.


Been a l-o-n-g time since viewing or writing to this NG. Good to see it
active and you guys writing about something worthwhile and interesting
(unlike many other NG's).

Regarding HOSTS and browsers; I too still use FF2 (Bon Echo v2.0.0.22)
on W98 and a large HOSTS file. There is NO perceivable performance hit,

I don't think I've seen a performance hit either, with a fairly decent size
HOSTS file

and
while it's true FF opens the HOSTS file one time (then keeps it open),
I've
noticed Opera (v10.63) does not! I can change the HOSTS file with Opera
browser open, UNblock a URL, and that UNblocked page is now accessible
(simply with a refresh). So HOSTS performance/effectiveness is browser
dependant.

Have you come across any site compatibilities or warnings while still using
FF 2.0? Or maybe that's why you're also running a newer version of
Opera.(?)

"Bill in Co" wrote in
m:

Probably, and not only newcomers; me too, I'd bet. Obviously, the one
more works with any tool, the better one gets at it, and that includes
discovering more and more of its nuances ... along the path to
enlightenment. :-)



I was kind of hoping you might have some dynamic filters to suggest. I
never did look into that thoroughly enough. (Got to run as DX plugins on W98,
but at a pinch I can adapt a VST..)

Incidentally, I guess a noise reduction system is a group of very narrowband
filters, grouped to match the print. But again, such a tool, with a simple
dynamic curve control, will get very close to FSE, while saving a great deal
of coding and CPU usage because it won't have to do the time/frequency domain
display. I'm sure such tools exist, I just never hunted specifically for
dynamic filters aimed at audio cleaning.

Lostgallifreyan wrote:
"Bill in Co" wrote in
m:

Probably, and not only newcomers; me too, I'd bet. Obviously, the one
more works with any tool, the better one gets at it, and that includes
discovering more and more of its nuances ... along the path to
enlightenment. :-)



I was kind of hoping you might have some dynamic filters to suggest. I
never did look into that thoroughly enough. (Got to run as DX plugins on
W98, but at a pinch I can adapt a VST..)

I haven't been using any (of the types I think you're thinking of), so I
don't know of any.

Incidentally, I guess a noise reduction system is a group of very
narrowband
filters, grouped to match the print.

Sounds about right to me. :-) And it's done with the computer algorithms
smartly analyzing it, and with a bunch of narrowband filters at its
disposal, from what I gather.

In some of these noise reduction plug-ins, you can change the FFT window
size, the FFT window overlap, the attack speed, release speed, and all that
stuff. (When using the Sony Noise Reduction plug-in, I've left my FFT
window size set at 2048, and there is indeed a real tradeoff in selecting
that value.

Using too many narrowband bandpass filters will degrade the result, due to
what I'd sloppily call time domain smearing - and you can hear it). So
Less Is More (down to a point). (Using too few filters isn't great,
either, and you can hear those artifacts too).

But again, such a tool, with a simple
dynamic curve control, will get very close to FSE, while saving a great
deal
of coding and CPU usage because it won't have to do the time/frequency
domain display. I'm sure such tools exist, I just never hunted
specifically for
dynamic filters aimed at audio cleaning.

Fortunately I haven't run into any problems with CPU usage so far. Even on
the Win98SE computer, although granted, it doesn't have as much of the later
audio restoration stuff.

"Bill in Co" wrote in
:

In some of these noise reduction plug-ins, you can change the FFT window
size, the FFT window overlap, the attack speed, release speed, and all
that stuff. (When using the Sony Noise Reduction plug-in, I've left my
FFT window size set at 2048, and there is indeed a real tradeoff in
selecting that value.


I found that varying it seemed to make interesting differences rather than
useful ones. Long windows work on long tonal structures, but short ones do
less harm to short atonal sounds like percussion and consonants. I ended up
with 1024 samples as my default. 2048 is good but I wanted the extra edge in
HF for a lot of things. A shorter window processes faster too. A much more
useful copntrol to vary was the amount of cut along with the new control
added to SF's v2 NR. I can't remember its name right now, but if you shift it
toward the lighter end of its effect, you can increase the main cut with less
damage to transients, and it's easy to adjust the two for a compromise that
works on a specific case. The new version had a switch for three different
types of NR, two of which matched settings in the original. I think I ehter
chose the new one, or the second form of the old. Yet another useful trick is
to slightly shift the entire noiseprint graph. That helped with the original,
but the new controls in v2 meant I didn't have to do that anymore.

"Bill in Co" wrote in
:

Fortunately I haven't run into any problems with CPU usage so far. Even
on the Win98SE computer, although granted, it doesn't have as much of
the later audio restoration stuff.


It's relative... Less CPU is always good, I hate waiting for NR to finish
before I can do HR. It's ok if I know what HR I'll do already, I can just
chain them and do something else, but on any new material I never know the HR
choice until NR is done.


About lots of filters, I think if the calculations get too complex, phase
shifts might get confusing. I never grasped the business of filtering very
well. I find that a few simple ones usually do ok. Also, I think that while
digital methods can do things analog stuff can't, it often makes sense to
stay with methods that analog means can do, because otherwise we can end up
with un-natural stuff our brains won't interpolate for. (Think of some of
those REALLY weird mobile phone artifacts heard on some news reports, and you
get the idea..)

"Bill in Co" wrote in news:AZ-dnV3B-
:

I haven't been using any (of the types I think you're thinking of), so I
don't know of any.


Ok. If anyone else is following this, and knows of dynamic filter plugins for
DirectX, running on W98 and aimed at accurate control rather than special
effects or VCO simulation, please post, any useful leads welcome..

Lostgallifreyan wrote:
"Bill in Co" wrote in news:AZ-dnV3B-
:

I haven't been using any (of the types I think you're thinking of), so I
don't know of any.


Ok. If anyone else is following this, and knows of dynamic filter plugins
for
DirectX, running on W98 and aimed at accurate control rather than special
effects or VCO simulation, please post, any useful leads welcome..

Maybe you could give an example of one (or one that might run on some other
windows version), so one might get a better idea.