Re: [SoX-users] SoX Stats: add peak position? (N Arora)

[nitin a. <nit...@gm...>]

Howdy,

We don't have a problem with backups (either locally or with our backup
network) but I don't know many people who inherently trust that things can't
go really wrong, someday, somehow. Obviously, the plan is to have good
backups and to migrate to new formats if a file type's sustainability begins
to get questioned; but ultimately this comes down not to definitive uses for
peak position, etc. but rather potential uses of data that can be collected
in seconds and will add little in terms documentation costs (considering
that max/min position won't take up much space in a text record). If it
takes seconds at virtually no cost (minus up front costs of devising a
methodology to get this info), than it seems like it's better to grab that
data than not. Ideally it languishes unused, but I don't think it hurts
anything if it's easy/cheap to collect.

"Which you no longer have; it's metadataabout something that
no longer exists; throw it away."

Can't necessarily just throw stuff away if you still have some record
relating to it, i.e. knowing something was once there is valuable
information in itself.

"Of course. If you had to re-create file2 (which shloud never happen),
you also have to create its catalogue metadata; which I suppose
is almost fully automatized anyway, right?"

There's a lot of human activity in terms of describing the item itself -
especially with more or less unique artifacts since you can't just gather
the data from some other database with the work already done by another
institution or company.

"So you want to re-create the data itself so that it fits with
the metadata you have? As opposed to re-create the data and
generate new metadata for it? Why would you want to do that?"

Mainly, I think it comes down to cost and the opportunity of trying to
sustain the prior work that was done. I think if one can find a way to get
things rolling again with the ways things were without doing all the work
all over again, it would be more cost effective.

"This is the first thing that makes me believe you really need
to have it synced. However, if you create file2 the same way you did
with file1 (which is not necessarily possible, yes, and should never be
needed anyway), wouldn't the difference (the un-sync) be just
a constant shift (making it easy to correct the sync info)?"

If the analog was played back a different speed it wouldn't just be a matter
of offset.

""Perhaps I could augment my collection of Beatles FLACs
with metadata about "when exactly is each song loudest",
or even "what's the highest note of each guitar solo".
Would this be an example of the "unique technical data"
of individual recordings that you speak about. How would
I ever use that?""

That comes down to perceived value. If you though that was valuable data and
could justify the cost, then sure.
ps: highest guitar note might be cool.
:-)
That potentially has use for students looking for songs to learn within the
context of what notes/positions they can comfortably play on the guitar, or
if it indicates a higher than standard number of frets, etc. I wouldn't
consider that "technical" but more what we call "descriptive" data, but
still.
But like I said it's a value/cost thing. That would be a ton of manual work
(unlike max/min position). If one was making a collection really tailored to
guitarists, then all of a sudden there's more justification for considering
it.

peace,
-- 
Nitin Arora
nitaro74 (at) gmail (dot) com
"Hope always, expect never."

humaneguitarist.org
blog.humaneguitarist.org