Re: [hylafax-users] Some TIFF's have bad pages

[Lee Howard <faxguy@deanox.com>]

On 2002.04.26 06:46 Dustin Puryear wrote:
> At 08:32 AM 4/26/2002 +0200, Bernd wrote:

>> I do not know how hylafax detects errors and how it reacts on
>> detected errors.

Errors in copy quality are detected by cqtest (when copy quality checking 
is enabled - not in the modem).  The reaction is either to send MCF or 
RTN.  When the modem is performing copy quality checking (Class 2/2.0) 
then the checking is done by the firmware (Class2PTSCmd).  Again the 
reaction is to either send MCF or RTN.

>> This also depends on the protocoll and how the protocol
>> is implemented on the other side. AFAIK the sender just drops the line
>> and tries again some minutes later.

This all depends on how RTN is handled by the remote.  But one would think 
that if RTN was sent and the remote disconnected that they would retry the 
fax if it were important.  Some bulk faxers do not retry after receiving 
RTN because of costs.

> That is what I had gathered from this list. However, I was looking (and 
> still am) for supporting documentation

CCITT T.30 is where you're going to find documentation about RTN/MCF.

> though because I've been told that "this didn't happen before Hylafax."

Well, that's interesting because I've had CQ problems (scrambled faxes, 
"unsupported compression", corruption, etc) when using any of a horde of 
Windows applications as well as when just using a "plain fax machine".  
cqtest does have some imperfection, as well as many Class 2/2.0 firmwares, 
and in cases where there have been problems, passing the TIFF through 
tiffcp or ImageMagick's convert resolved the issue.

> So, I'm in the situation where the users believe the problem is either 
> Hylafax or the Multitech modem cards.

Some modems' firmware can cause problems with CQ, true.  I've seen it 
happen on USR Sportsters, USR Couriers, some Rockwell implementations, and 
some MultiTech modems.  Again, in these cases switching to HylaFAX-based 
checking (switching to Class 1) helped, and then using tiffcp or convert 
corrected any further problems.

If you're having tiffcp generate corrupted tiffs from a corrupted tiff, 
then you should bring the issue up with the libtiff folks.

Lee.

____________________ HylaFAX(tm) Users Mailing List _______________________
  To subscribe/unsubscribe, click http://lists.hylafax.org/cgi-bin/lsg2.cgi
 On UNIX: mail -s unsubscribe hylafax-users-request@hylafax.org < /dev/null