whole morning session went in admin work - db updates, creating a report, emails etc.

Procedure for preparing pre-recorded CDs for broadcast

1. Check content - extremely poor quality stuff and extremely jazzy stuff to be avoided.

2. Save as mono wav files to a folder in computer
(a) Open Nero Burning ROM
(b) Choose menu item Recorder->Save track
(c) Choose Settings, choose mono, choose folder to save in, GO.

3. Rename properly using tools like CKRename
(a) Songs are to be named like SV_AUSTRALIA_WE_ARE_ONE_07_WHY_FEAR.wav
SV_(COUNTRY)_(ALBUM NAME)_(TRACK NUMBER)_(SONG NAME).wav
(b) Bhajans as BV_MOROCCO_GITANJALI_01_GANESHA_JAYA_SRI_GANESHA.wav
BV_(COUNTRY)_(ALBUM NAME)_(TRACK NUMBER)_(SONG GOD)_(SONG NAME).wav
(c) Similarly, instrumentals and concerts have prefixes also
BV = Bhajan Vocal
BI = Bhajan Instrumental
SI = Song Instrumental
SV = Song Vocal
CONCERT = concert
4. Remove silence in excess of 1 second max from beginning and end of each track,
apply fade in/fade out respectively.
(a) Open wav with the option "Operate directly on file" in Sound Forge
(b) Select silence, DEL
(c) Select beginning 1 sec, choose menu Process->Fade->Fade in,
similarly fade out at end.
(d) Save and close.
5. Put all tracks into a batch process to normalize volumes to -10dB
(a) In Sound Forge, choose menu item Tools->Batch...
(b) choose source files and destination folder
(c) Set convertion settings from menu item Options->Convertion Settings,
From the drop down lists, choose
"Process"
"Normalize"
"Normalize to -10dB"
respectively, choose the (+) button to add this to the batch operations.
(d) Run the batch process.
6. Convert the resulting files to mp3.
(a) Open Musicmatch Jukebox, choose menu item File->Convert Files
(b) Choose source and destination folders
(c) Choose wav as source and mp3 as destination, set to 96 kbps.

colinux install

I'll just link to the installation instructions... with the difference in this case being that we already had a full fledged Debian 1 GB with all our required stuff from the old server... so we just zipped that image, transferred it over and hey presto. Of course, with the config changes I mentioned yesterday.

stream.radiosai.org synchronised with published schedule.

adding RSA auth

Copying and pasting from the ssh man page:

ssh implements the RSA authentication protocol automatically. The user creates his/her RSA key pair by running ssh-keygen(1). This stores the private key in $HOME/.ssh/identity and the public key in $HOME/.ssh/identity.pub in the user's home directory. The user should then copy the identity.pub to $HOME/.ssh/authorized_keys in his/her home directory on the remote machine (the authorized_keys file corresponds to the conventional $HOME/.rhosts file, and has one key per line, though the lines can be very long). After this, the user can log in without giving the password.

Now copy/pasting from a mail I wrote before:

Procedure:
Run as whichever user you want to connect as:
ssh-keygen -t rsa

don't enter any passphrase if you want it to be automated.

copy the key file (from .ssh/id_rsa.pub by default) to the next line of ~/.ssh/authorized_keys file of whichever
system you want to connect to. If the file doesn't exist, just copy the key
file to authorized_keys file.

And disabled password based authentication for security:

in /etc/sshd-config, the lines

PasswordAuthentication no
PermitEmptyPasswords no
KerberosOrLocalPasswd no
UsePAM no

checked it by first logging on as sgh from AfriStar playout machine, then after the changes, tried again, it said

Permission denied (publickey,keyboard-interactive)..

Could've just tried as a different user - even the root account on my local machine is denied. :-)

Added the new datacenter's name servers to /etc/resolv.conf

Changed myhostname in /etc/postfix/main.cf

Added stream.radiosai.org CNAME domain record in Plesk

Added the gmail account to the mailer daemon's list of addresses

Found lots of errors in /var/log/mail.err - fatal: open database /etc/aliases.db: No such file or directory
Ran newaliases and restarted postfix as per http://www.howtoforge.com/forums/showthread.php?t=652

Restarting postfix in debian (not service postfix restart as in RH/Mandrake): command is
/etc/init.d/postfix restart

colinux up and running

New streaming server coming online soon. Have set up colinux on it for sourcing the stream. This is an interesting story by itself and will post on it later when time permits.

Needed to add firewall rules (copy paste in /etc/sysconfig/iptables) on our local server to allow the new server to connect - ip-based filtering.

TODO - running sshd (Using the method given in http://pigtail.net/LRP/printsrv/cygwin-sshd.html )

editing dns

Our host uses Plesk. And I was just now changing the domain forwarding. Earlier, it was sssihl.edu.in forwards to sssihl.org. Now they want it the other way. So, deleted and remade the domains. The painful part being the downloading and then uploading of all the files. A few hundred small files, so takes > 20 minutes to ftp.

It seems this giant puppet looks like me or vice versa.

Evening session spent making Album updates with JAlbum (and kwrite for the main page, of course.)

AutoGK + and -

AutoGK settings:
(A good guide is installed by default. It will be in the start menu in the AutoGK->Tutorials folder.)

Open file: - Open the main IFO file if you want to transcode and entire DVD, and choose the appropriate program chain (PGC) by seeing the duration.
(Unfortunately it seems to pick the wrong chain - tho it mentioned 40 minute PGC duration, final O/P was 2 hours long!)

Choose a Subtitle track if you want a subtitle track burnt into the avi.

Select output size as per your requirement. Better to choose file size than the percentage stuff.
In advanced settings, choose the minimum width option and give value as 576.
(So, the video will be resized, with a min width as 576 pixels. In case the output size is big enough, a bigger resolution will be automatically chosen.)

Aspect ratio, frame rate, cropping etc are automatic. If you feel auto crop may not work, press Ctrl-F9 to go to hidden options, and change the number of frames and the begin frame of the auto crop detection. You could also change the "Enable Standalone support" to ESS if you prefer - that is supposed to be more compatible.

"Add job" and add any more jobs to queue if you have more than one PGC to process.

Processing takes roughly 1.5x realtime with an output of 150 MB file for one hour of video (roughly 300 kbits per second!) on a PIV 2.4 GHz.
But this low bitrate makes the output look like VCD quality. . .

Google Earth for Mac

Oh good! Google Earth is now available for Mac too. When is the Web-based or Linux version?

StaxRip

StaxRip needs .NET framework! Sorry, too bulky for me to try out.

mirroring

wget -r -N -l 2 -nr --random-wait --limit-rate=2k -k http://hometheaterhifi.com/masterindex_essays.html

for getting the stuff mirrored. wget -m is equivalent to -r -N -l inf -nr (infinite levels, mirroring) as the man page says.

and the -k does the modification of links for local browsing.

hometheaterhifi.com has a few good feature articles. Informative.

AutoGK

AutoGK needs you to do a little more than the FairUse Wizard, in the sense it will not automatically cache files to hard disk, you have to do it yourself, preferably with DVDDecrypter which can read even somewhat scratched disks. And no pretty progress bars for the "Advanced Option" of Preview, which needs indexing for 5 minutes... and then with progress bar for another 5 minutes... And all that gets done again when you are actually encoding! Not cached.

Only where FairUse Wizard cannot be used do we need to consider AutoGK.

better DVD to DivX tool?

Came up against the limitations of Fair Use Wizard free version - won't allow arbitrary in and out points, and won't allow output file size > 700 MB.

Gordian Knot and AutoGK are well regarded. Will try them out if I've to do any more DVDs.

How time flies

Good old fashioned logging:
07:30 to 07:45 - emails and playlist subscription updates
07:45 to 09:00 - playlist making
09:00 to 09:15 - starting cool edit pro on noise reduction, WM encoding begin
09:15 to 09:45 - fixing 2 hard disks (earlier on Prema) on Sathya
09:45 to 10:10 - trying to get the 80 GB drive recognized, finding that Seagate Disc Wiz might format it, abandoning attempt
10:10 to 10:35 - finding out why WMAppend is crashing - due to ASFTools corrupting the wmv or due to the added change in the encoded file of not cropping? Encode again and find out...
10:35 to 11:00 - start DVD decrypter and WMEncoder, type this out. Cool Edit Pro still running..
11:00 to 11:40 - saving cool edit pro noise reduced wav as 3 files, dvd decrypter not reading 2 DVDs... removed the 80 GB drive from Sathya to try later on saispace2
12:45 to 13:30 - WMAppend not crashing with freshly encoded files - so most probably the ASFTools is the culprit - files joined with ASFTools were not seekable, and it is probably corrupting the source files too...
13:30 to 14:00 - creating batch job for 2 DVDs, screening new CDs for Radio Sai.
14:00 to 14:30 - Flute anns cleanup
15:30 to 16:00 - Convo drama 2005 cleanup
16:00 to 16:30 - taking out songs of convo drama 2005

Time out for bhajan

17:30 to 18:30 - going to studio, getting CDs, asking SSR to ask if we have M. Santanam music rights, etc...
18.30 to 19:15 - collating and renaming the DivX files, adding metadata to WMV files and starting upload to saicast (20th Nov choir)

20:15 to 21:00 - CG's "console clock" wiring.

(the post time does not reflect edits!)

Contact address

Just in case you want to contact me and don't know how to:

hn_88 AT rediffmail.com

This is just to test the image upload. From XP's sample images.

And so it auto resizes to 320x240 - orig was 800x600.

DVD to DivX

Today's experiments were on DVD to DivX with (English) subtitles encoded into the video.

GMencoder looked promising, but the subtitles part was greyed out - maybe some dependency not fulfilled.

transcode was simply too much work for my click-happy self.

dvd::ripper failed some Perl module dependency - Locale::TextDomain

So, finally went back to Windows. Poking around in Videohelp.com came across this nice tool based on VirtualDubMod, I believe, FairUse Wizard. The resolutions it chooses etc are quite nifty. Free version limited to 700 MB files, but can have batches and can choose individual titles or parts of them... On the whole I'm happy.

The power of open source?! Or "The best tool for the job"

Wow! Virtualdub is 10x faster than Adobe Premiere!

Exporting from PAL VCD mpg file (352x288 25 fps) to 320x240 huffyuv encoded avi with "Better resize" in Premiere and Bicubic resize filter in Virtualdub, with a logo being watermarked... Premiere did at approx 16 fps, Virtualdub is doing now at 140 fps!

Of course, I had to do some trial and error and create the logo file separately in Photoshop for Virtualdub, while I could make a title file (.ptl) in Premiere itself for making the logo....

pottering around before sleeping

Loaded the ext2 driver for Windows from fs-driver.org in the Thinkpad too, so now I can see the whole 80 GB in Win2k too. Did some rudimentary listening to Jan 1st morning music prg, wrote todo notes and left it at that, since the compressor in the Adobe Audition loaded in the Thinkpad doesn't have the particular preset I favour.

PS. Posting this in the morning with Firefox 0.8 on Win98 - so three posts on three different browsers and three different OSs. Attilio film dubbing going on ...

And this post is from Mozilla 1.6 on Mandrake 10.

All's well so far. Even copied + pasted text from kwrite comes OK with para breaks. Will stop now and continue later, probably typing on the laptop

First post

This first post is from IE6 on XP. Now I'll restart into Linux and try posting with Mozilla.