Just for completeness: the problem turned out to be the embedded sound card in the motherboard. A VIA chipset (cheap-generic) Encore PCI card solved the issue.

It looks like the line input of the embedded card is too noisy.

Many thanks for the responders.
Fer

Hi.
Thanks for your comment.

>I have run FLDigi with Puppy Linux and a SignalinkUSB >and it is hands down better than running any Windoze >software. No comparison on RX.
That's pretty much like anyone's experiences with fldigi over GNU/Linux (or at least, that's what I've found on the 'net).

Anyway, there are a lot of things involved here, from the hardware to the kernel version, ALSA vs. OSS, et alter.
I think few setups would be similar enough (end to end) to compare each other and draw any conclusions based only on on-the-air observations (unless all results coincide ;-) but if comments like yours represent the average result with fldigi on GNU/Linux, then I hit a unlucky combination of factors/hw/sw.

The most problematic aspect seems to be the fact that on relatively strong signals, without QSB, gmfsk delivers 100% copy flawless on 40m BPSK where on the same signal at the same time fldigi barely reach 70-75% copy. I've made A-B tests (one program running at a time), and simultaneous tests feeding both programs the same audio with ALSA or jackd, and fldigi performed below par every time.

I expect that on the same sw (kernel, drivers, ALSA) same hw and same signals, the performance of both programs should be similar.

ASAP I will build fldigi from source and will try another sound card just to be sure, but I really don't have any hope.

Can you send me your fldigi config files?

Thanks, best regards,
Fer

> What mode(s) have you tested?  I am assuming that you
>have used PSK31 in your tests. Thanks for reporting
>your results.

Yes, mostly PSK31, it seems it's the major part of the activities. Sorry
about not mentioning it.

I've done some testing on RTTY signals but not enough to derive any
conclusion or to really compare results.

Thanks for your answer.

73, Fer

Hi.
I've alternate the use of fldigi and gmfsk31 on my Debian box.
Over several months of carefully testing I'm concluding (not without some
surprise) that gmfsk is nearly always better than fldigi at decoding signals.

What I've found is that for a given signal strength or S/N ratio, gmfsk
consistently shows much less errors or garbage in the received text, and with poor
conditions or weak signals, gmfsk gives me some copy when fldigi only gets garbage.

(I've made the comparison in the same machine, using the same soundcard and a sw
mixer in alsa or jack to feed both programs the same signal at the same time,
running both programs side by side in the same screen at the very same signal,
carefully calibrating the frequency in each program and letting both programs
freely auto center it's frequency (AFS on)).

I saw and occasionally heard comments on the ability of fldigi on capturing
data from signals with poor S/N ratios to the point of not even be visible on
the waterfall. Then I must emphasize that this never happened to me up to this
day, and all of the signals I've tested on were always very clearly visible in
the wf.

Given the good reviews and features that fldigi, I'm somewhat perplexed by this.

Any ideas or similar/complete dissimilar experiences?

(Pse: only Linux/BSD comments, I do not run windows nor mac, and I'm not
interested in doing so).

My setup is:
Official fldigi 3.12.5-1 for Debian Sid on amd64 (Athlon 64 3000+).
Sound card is integrated Nvidia CK804 (mobo is ASUS A8N SLI, NForce4 chipset).
The tcvr interface is homebrew, fully isolated.

73,
Fer

There are several similar classic circuits in the Handbooks and internet.
The SCR-274-N/AN-ARC5 command set transmitters are 2x1625 MOPA rigs. The oscillator is a 1626.

I love that I don't need an antenna tuner... err, wait...

Hello.

IP packets have nothing to do with this. Every packet can be sniffed, it's the same work to sniff one or all packets passing through a link, router, network interface or whatever the device in the middle of the traffic is.

But, you miss the point completely again: the problem is not the sniffers on the *wires* alone, is the very nature of email transfer what makes it vulnerable. In order to get transferred from one machine to another, the email has to be stored in _multiple_ places along the path from sender to recipient. This path may involve one machine if the sender and recipient are both users of the same system or net, but the common scenario involves several boxes at very different places each one with it's own policies and sysadmins.
These days too much companies/sysadmins have the policy of store copies of everything that passes through their servers, in some cases for months, just because they can (storage is cheap). One of those systems compromised, or one unscrupulous sysadmin, and you get a very nice set of email-addresses and verified passwords to try on.

You don't believe, don't you? Well, try this: subscribe to a couple of lists that send reminders like this one, using the password and username/mail address of your internet-banking account, and wait. Drop me a note if you do so, I would like to take a chance ;-)
Bye.

Oh, man.

>>The big reason for a login at cebik.com
>>is to keep out the mass downloaders that
>>eat up bandwidth, up until the death of
>>Dr. Cebik the site was wide open for anyone to
>>visit and learn.

I completely understand and respect the site's owner policy on resource limitation/administration. I do *not* put a word against this. I'm not questioning policy, but mechanism over which the policy is implemented on.

>>If you think your email is sent in one part,
>>they you need to learn about how the internet works.

Yes, the email was sent in one part, and no, it's not my opinion, is a fact.
And if I'm wrong on this, well, there are several respected sources on-line and in dead-tree form that are wrong too, so I would be really happy if you can explain me how the email I've copied in my original post was delivered to my mail box in "parts".

If this is interesting to you, or you're administering a mailman service, please take a look at the mailman documentation and see how this is an old issue, discussed several times on mailman-users list.

Look at this too if you mind:
http://wiki.list.org/display/DEV/200701/13Passwords+done+right

Anyway, I think both you and Stone are trying to kill the messenger, instead of fixing the problem. And believe me, the fix to this flaw takes much fewer keystrokes than the responses you post here or the mail Stone send to me to get me out of cebik.com. In fact, it takes only ONE click on the mailman administrative page.

Cheers.

Hi,

>>the emails they send are not sent in one piece but split up
>>on the way from cebik.com to your isp and to you.

Really? Can you explain me how this is done? Also please tell me how the
pieces are re-assembled before hitting my email client's inbox
folder, since they're in one piece when I've looked there.

e-mail is by no means a secure way to send any sensible data. You're
assuming a lot of things which are not as simple as you think. email has
to travel a lot of different networks, machines, sysadmins policies and
unsecured/untrusted paths to get from the sender's machine to
destination. In every one of this paths data packets can be sniffed or
copied or both.

>>Nothing to worry about.
I'm not worried: _I'm_ not using the same login/password in two sites or accounts. But *if* I were doing such a thing, I would be better very worried:
suppose I've had the bright idea to put up every account to the same
username (very likely, my callsign) and password. Now, a sniffer is run
in the path of this email, so the sniffers gets the *email address* and
the user/name password on it sent in *PLAIN TEXT*. Got it? He is able to
hijack my email address, and with it, he can hijack EVERY other account
I have linked to that address, merely hitting "Forgot your password"
links in sites and getting new passwords sent to this email address.  He
can impersonate me in every site I have account on, and even create new
accounts. Hope I wasn't so stupid to link my bank/credit cards accounts
to the same email address or username/password.

But, if happened to the tap/sniffer to be closer to the sender smtp server,
every address/username/password of the site would be hijacked.
Please see http://en.wikipedia.org/wiki/E-mail#Privacy_concerns

Also, you have failed to note another security trap revealed by the
email:
> http://www.cebik.com/mm2/control_panel.php?email_address=lu2dfm@domain.com&pass=XXXXXXX
Please note the fact that the credentials to the site itself are passed
out as variables embedded in the URL, so if you're login in the site
from a untrusted/public/not-your-own machine, you need to properly
delete the browser history file and cache to at least make a little
difficult to other people to see at your authentication data. Also it's
possible that this URL with embedded credentials to get logged in the
http server log files at the hosting site, because several ISPs use this
info to track down usage statistics, client's IP addresses, etc.  It's
possible to the http server admin, which may or may not be authorized to
manage user authentication information, to get this info.

>You are underestimating the man who runs Antennex.com and cebik.com, Jack Stone.
Oh, yes? Why?  He's not doing the right thing as site sysadmin. Enough.

> why not drop him an email at webmaster@antennex.com
I've replyed to the sender address (listmgr@antennex.com) of the email
I've got, which was signed by another person. And this mail didn't get read by
Jack Stone nor anybody else (although it was delivered to their server
and accepted into them, so the account must exists there).


I received today this email from Jack Stone:
=========================================================================
Date: Fri, 07 Aug 2009 08:35:31 -0500
From jack@antennex.com Fri Aug 07 10: 6:07 2009
From: Jack L Stone <jack@antennex.com>
To: lu2dfm@domain.com
Subject: Your complaints
Reply-To: jack@antennex.com
X-Spam-Status: NO, hits=-10.00 required=4.50
Status: O

Fernando:
It has been noted you have made unwarranted complaints about how we
handle login info at the cebik.com site.

If you have a complaint, suggest you complain to me (us) directly
instead of an unrelated forum.

Clearly you don't know what you're talking about. Until a year+ ago,
there was no login at all and the content on www.cebik.com was wide
open. The login is only an attempt to curtail piracy by attempts at
mass downloading.

Visiting the web sites are a privilege, not a right. Privileges and
rights can both be revoked for cause.

BTW: Other hams have reported your behavior as being willfully
ignorant. You are not doing yourself any favors by such posts.
=========================================================================

So this demonstrates that they are not even capable of setting
administrative address aliases, or non-reply-accepting address in their
servers, so my response to the address from where the email was sent was
not delivered to him by his *own* mis-configured server.

My account was revoked, and I was unsubscribed from their site/list.

Also, this shows how thin-skinned some people is, and how they manage
their mistakes: killing the messenger.


Best regards,
Fer

quote "They send an auto renewal each year
to the email address on file,
if it bounces they close
the account on www.cebik.com."

Yes, but every list/site out there sends me a remind of my subscription once a month or once a year. The point was: there is no need to send login _and_ password information in a unsolicited administrative email in clear text.
This reveals a very limited knowledge of internet and internet technologies, and system administration.
And clearly points out that they have not any concern about user's security or privacy.
A real shame, considering the quality of W4RNL's work.


quote "I agree, good idea to have seperate
email accounts for radio related stuff"

For this site the most clever measure is to put up some special username and password (everybody *is* using different username/password pairs for each site/account, anyway), but dedicated/disposable email address.

73

www.cebik.com site is a quite useful one, widely recommended and linked/cited here on eHam and alike sites over internet.

Today I received this email from antennex.com account/list manager:

===========================================
Date: Fri, 31 Jul 2009 00:06:04 -0500
From listmgr@antennex.com Fri Jul 31 02: 6:27 2009
From: LB Cebik Antennas Website <listmgr@antennex.com>
To: "Fernando M. Maresca" <lu2dfm@somedomain.com>
Cc: jason@i2-services.com
Subject: Your Account is Nearing Expiration: LB Cebik Website
Reply-to: LB Cebik Antennas Website <listmgr@antennex.com>
Status: O

Fernando M. Maresca,

Your account is set to expire on August 7, 2009.

You can access your account using the link below:

http://www.cebik.com/mm2/control_panel.php?email_address=lu2dfm@somedomain.com&pass=
XXPasswordInPLAINTEXT

...and cancel your account and re-register which will renew for a year.
...
==========================================

As you can see, they have a totally relaxed policies about privacy and user identity protection. Login name and password are sent in the email ON PLAIN TEXT to provide for auto-login on the list account management form.

Be sure to put a username and password UNIQUE for this site, and use a password disposable or exclusive for this account.

Hi
>>There is a link to the login page at the top left of >>the page in the green bar.
Yes, I know. But it takes you away from where you are. Since it's trivial to make the successful logging to redirect to where we were previous to logging in, I suspect that something is FUBAR in the site.
But that's not was my point: I was suggesting a little form, user and passwd, in every page, instead of a logging link, with a js script that put me in without to even reload the page where I am.


>>Our new forum software will have all of the features >>you would expect in a full featured normal web forum >>system.
Cool

>>We have looked at the slashcode. In many ways I like >>it but I am not sure we are ready to convert all of
>>our code like DX clusters and other special pages >>over to perl. I really do admire the slashdot >>moderation system.
I'm not know what that code for DX clusters and other special pages does, but it's _very_ little what can't be done in perl. Besides, exists other slash-like codebases in other languages: http://www.slashcode.com/slashalikes.shtml (which I don't know nothing about).

>>I to have wondered if it would work on our site.
Well, that's the big question IMHO. And probably you're the best situated to guess the answer :-)

George: as I said in other thread, you're doing a great job; eham is perhaps the best amateur radio site out there.
My point is not about questioning the site or the codebase of eham, but to point out that taking out the duty of make and maintain the site infrastructure, there will be more time to concentrate in contents, and users experience.

I'm a developer too, and I know that's most times it seems logical or easier to work a lot over something is known, instead of facing a big change to something not known. But sometimes, need is need.

My offer to contribute to the code is standing.

Best regards,
Fernando

In another thread in this same forum I've suggested that eham needs to move to a codebase ala slashdot (slashdot.org).

My suggestion is based in the idea that the slashdot codebase (which is free, see http://www.slashcode.com/) has already much of, if not all, the features we have here on eham, and a _lot_ more. Some of these features are perhaps necessary, like peer moderation and thresholds for post browsing based on the post score.

I'dont know what the eham's user base is, nor what is the real percentage of users actually active in fora, so the critical mass for a site like slashdot to proper get the moderation system to work is here or not. But my guess is that it will be fine (even if it don't work at all, in the very worst scenario, the present moderators will be doing the very same job).

Let me put this straight: I'm not doing any criticism about the codebase of eham, which I don't ever look at, not the ability of eham's developers to get the job done. My point is about not reinventing the wheel, and redirect the efforts that are made to support and develop this site, to another well established codebase, which is FOSS for anyone to use.

Thanks

Hi. Thanks for your reply.

Couple of other suggestions:
1) When trying to post, not logged in on eHam, we're redirected to the Register as a new user form, from where we need to click on a link to go to the logging form for already registered users. I think this is the reverse order to do that, so it would be nice to get this reversed, or, at least, put the two forms in the same page.
Another related improvement would be to provide an user/password logging box in a sidebar/topbar in every page of the site.

2) About fora: a HUGE improvement would be to provide threads and nesting in the forum and the articles; this is a necessity, because about any article gets a number of off-topic, semi off-topic and other unrelated posts, which gets mixed in the middle of relevant ones.
Hey, this site is in the very need to migrate to slashdot code!!

Thanks

Can anyone explain why it's not allowed (and this restriction is not indicated in the forum help nor the site faq) to provide correct HTML markup or tagging to URL references when posting links, like this: <a href=www.google.com>Google</a>?

Why not provide the reader the "normal" means to take the intended action when confronted to such references in poster's comments, e.g. make use of the conventional gestures/keystrokes in his or her browser (enter, ctrl-enter, left-click, middle-click, etc.) instead of make him or her highlight the text, copy it, open a new tab/window, paste the copied line and hit enter, which I frankly found obnoxious and irritant.

BTW, it's about being in the WWW, or not?

I would contribute a patch for doing so if it's needed.

Thanks, Fernando