On Friday 05 March 2010 16:23:22 Indranil Das Gupta wrote:
> Hi all,
>
> Last night Dr Mandar Mitra (of ISI Kolkata and IOTA board member) -
> an old friend and long-time FOSS user, sent out a rebuttal to the
> claims of unfitness of Audacity as a audio rec / edit tool. In his
> email he spoke about how he would do his recordings and what has
> been his experience so far with it [1]. He also invited Shri
> Anindya Banerji, the multimedia expert in question to further
> clarify his comments and remarks about Audacity vs Audition.
>
> A while back Sri Banerji, responded back with his clarifications.
> I've put up his response on my blog -
> http://indradg.randomink.org/blog/archives/155
>
> Mr. Banerji's response makes for rather interesting reading and I
> would request input from all, but particularly (based on my limited
> knowledge) from Dr Nagarjuna, Niyam Bhushan, jtd and others who
> have come in with specific inputs so far on this topic.
>
> thanks
>
> references :
>
> [1] http://indradg.randomink.org/blog/archives/154

Ok. From the response one can conclude two points
1) Closed formats
2) The consultant is sticking to a single tool when he has undefined 
jobs.
3) The consultant does not understand the real issues of audio 
processing using computers.
4) This requirement is not for professional high definition audio 
recording, - the type you would have in a SciFi movie or a modern 
heavy metal band - but the usual government ration bhashan editing.

1) Dont use closed formats EVER, irrespective of any arguments. If you 
need quality use analog tapes (and incur the cost of maintaining 
those - remember it is very expensive). For archiving use raw 196Khz 
sampled data. For other archiving purposes use flac. For consumption 
uses ogg.
All parameters of an ogg encoding can be customised. So depending on 
the audience (mp3 player / FM radio / Internet radio / hifi  etc ) 
you can set the sampling rate, sample size etc providing suitable 
compromise on quality v/s size/bw. 
Most important - use of closed formats will make everyone incur a 
penalty while playing back. If the government wanted to stream a 
closed format, they will directly be paying a fat sum for closed 
streaming codecs, instead of using VLC / Apache+ icecast/ shoutcast / 
some taken-for-granted FLOSS tool chain.

2) When you have undefined batch jobs like converting multiple input 
formats / quality to a single predefined format, use sox (and ffmpeg 
for video or audio container).
One has to spend a few hours understanding the workings o these tools. 
Even if one builds a gui, this requirement would not go away - 
undefined problems have a way of styming everything except the CLI.

3 and 4) One does not record sitting on your desk. You require an 
anechoic (or a room with a well defined acoustic print) and fixed 
high performance microphones. And you would have a table load of 
equipment to check dynamic range, TIM distrotion, frequency response 
etc. 
Also clock jitter and drift shows up as distortion. Clock jitter is 
never checked in a pc and quite a bit of drift occurs as the ambient 
and chip temperature changes. Thus pcs will have markedly different 
distortion depending on the difference in the source data clocks and 
the sound card clock. The older sound cards used an external crystal 
(which was good). The new onboard sound uses the motherboards 14.318 
xtal and a pll to generate the 24.576 Mhz clock. They are hence 
exposed to the jitter of he 14.318 Mhz signal (besides their own). 
Hence you use a PRO quality sound card like the delta 1010 with "word 
clock I/O for sample accurate device synchronization" or an external 
DSP box.
  
Finally live recording / dubbing of audio at a desk is used mainly for 
porn, shaadi-happybirthday scenes. "quickly fixing recording errors 
by punching in corrections on-the-fly " a must for my kids first 
classic rendering of "Three little kittens".

As usual while the software requirements are PRO everything else in 
the tool chain is distinctly amateur - or to give the benefit of 
doubt completely unspecified.
IMO Dont waste money on studying the workflow and quality and 
performance. Any old software can be used with a little bit of 
training.


-- 
Rgds
JTD
_______________________________________________
network mailing list
[email protected]
http://lists.fosscom.in/listinfo.cgi/network-fosscom.in

Reply via email to