Hi all.
I suposse that the high-CPU-consuption in the encoding
audio processes is one of the problems we find as time
passes and the encoding algorithms get much and much
complex...

Testing and trials sure that consume a lot of time of
processing only with experimental pourposes waiting only
for the computer to encode...

This week I had an idea accelerate the encoding times...!
Network Audio Encoding.

In the same way that several 3D graphics programs can
render over several computers conected to a LAN (3DSMax,
and many other render programas have this option...)

OFF-TOPIC:
I'm a graphic designer, and I use network rendering at
night over the company's network to perform final "high-quality"
renders...

Maybe will be interesting an application that can
"distribute" the audio encoding process over the computers
of a LAN Network?

Following the paralelism with the 3d graphics progs...
All that we need is a "Server programa" who distributes the
files to encode (even maybe the encoder executable), and a client
program that get the file to encode, the encoder executable, and
the encoding options... does the works and sends back the resulting
file!

Also exists some network rendering programs (as Mental Ray) that go
further far, and are capable of distibute 1 single image over several
computers... I don't know if this is possible to do encoding WAV files...

I mean slice a single audio track in parts (as many as computers in the
network), send each part to each corresponding computer, the "slave"
cumputer, encodes this part and sends back the MP3 file resulting...
and the server program "glues" the MP3 parts in a single MP3 file.

Well here is my idea...
Hope that someone finds it interesting...

Best regards to all.

b:.

Well, it sounds like an idea for transcoding a music library with 10000+ files, but audio encoding isnt nearly as time consuming than 3D rendering.

Well...
I usually read here about people that are thinking in "encode" their
whole CD collection... a 1000 CD collection at 10 tracks per CD =
10.000 files!



Maybe you're right... It will be more interesting talking about great file
numbers... but I think Its interesting.

Best regards
b:.

If I had about 10 computers and about 1000 CDs I'll copy on each of them about 100 CD's and manualy run RazorLame session on each. It's a theory. But how can you grab 1000 CD's in one day to leave them for night?

doesn't everyone have a 5000 disc jukebox cdrom drive?

Rob Fantastic MP3 Encoder: it uses networked machines
to speed up the encoding process.
I've seen it around for over 2 years, and the app looks like outdated now.
I didnt found any speed improvement by using it.(at least on my 3 machines home network).
see it here

LIF

If you have 1000 CDs and want to network encode them, how do you tempstore the uncompressed data? Ever thought about that?

I was encoding some oggs (a relatively intensive encode) last week at 10x realtime on my 1.5GHz PIII. I realized that the encode time compared to the time required to get a high-quality rip is essentially trivial on a modern processor, so I'm not sure this would be worthwhile. Now xvid encoding, *that* might be worth it!

I think GRIP does support this network ripping & encoding.

I've taking a look to "A Rob Fantastic MP3 encoder" and does exactly was I said. But the program is seriously outdated... an lack of some serious limitations as:

- It doesn't have an option tu put "command line" options...

- It doesn't support external encoders...

With this limitations...
And it comes with an old version of lame_end.dll, without options to encode at VBR.

So it's impossible to encode the files with my "beloved" --alt-preset standard.

It's a pity. I still think that in some situations may be usefull.
Thanks all for your comments.

b:.