#musicbrainz

I dont think download AR is appropriate for an artist

you download their releases or individual songs

aCiD2: that page says in the first line "most compact and quiet enclosure in the market"

doesn't mean we should believe them, but they did mention that before anything else

true

but I guess, if you're buying a big enclosure, money is not sooo much of an issue

and I'd say it goes "price, volume of noise, size of enclosure"

$200 is a lot to pay when I could just buy 3 enclosures for $100 and tape them together...

especially since I doubt I will be using RAID

but I am out of USB ports on my stupid iMac

no raid with THAT much space?

seems a little foolish

Some other odd ones here: under track-url, pretty much all the misc ARs seem to duplicate what could be found in the artist's "official homepage" for artist-track ARs... http://musicbrainz.org/edit/relationships/link_...

I'd consider it, but I'd have to get all this crap off somehow in order to raid the drives

aCiD2: Largest I saw to buy off the shelf when I last sold that stuff (January) was a 2.4 TB external

cost around $2k US

single?

iirc, it was a 4 drive RAID 5

ah

think the manufacturer was Yellow Box

see, when you enter that much space, you're just wasting money not buying raid, imo

hehe you should see my desk.. there are 3x250GB USB2 drives sitting under my desk

Can anyone tell me the point of some of these misc track-url ARs? Like, for example, "track has artist & repertoire support by url"?

I have 1 250gb in my "server" and that has 100gb free

k time to study, gonna try writeroom

And I'm off to bed

is it my imagination or did the edit page just change?

I don't see anything different...

hmm, the timestamp was in the front of my edit note instead of after it. and there were links to "open edits" and "all edits" next to the artist and release name on the edit

You sure you're not editing on test?

at least that first (timestamp) sounds just like something I saw on test earlier...

\o/

Hi

Any reason track lengths on MusicBrainz are rounded *down*? When adding all track lengths of an album, the total is often off (smaller) by several seconds because of rounding errors

To MB's credit, all players I've tried seem to do the same

...is it approx. 2 seconds per track?

no, why?

rounded down means the difference is always smaller than 1 second per track

just wondering if it was the whole 2 seconds of silence between tracks thing adding up...

no I'm excluding the usual 2 seconds pregap

besides that would make a difference on the total, not per track

well, if the album is imported via CD lookup or from freedb, the lengths are in 'sectors', not seconds

luks: the sector values are good

so summing all lengths in seconds and the real CD length will differ

I'd imagine that the main reason is that there's no real interest in keeping track of the fractions of a second for each track, and it just gets converted straight to an integer number of seconds, rather than being rounded off to the nearest second...

it's the conversion into minute:second that's rounded down

hm

do you have an example

?

sure

any TOC will do though

if a track lasts 149 sectors (2 seconds minus one sector), the duration will be rounded down to 1 second

hm, not sure if I understand

the lead-out track is 160948

(160948/75)%60 = 45

(160948/75)/60 = 35

I'm talking about individual track durations

ah

ah, right

sub(track-times) - album-time != 0 ?

err... sum(track-times) != album-time

it probably does (int)(X/75) instead of (int)(X/75 + 0.5)

I don't see how (int)(X/75 + 0.5) would be better?

even if it did do that, it would probably still not have the sum of the track times being the same as the album time.

int discards the fractional component;

+0.5 is a mathematical rounding

0.2 + 0.5 = 0.7 ~= 1, not 0

(149/75.0+0.5) = 2

int(149/75.0) = 1

yes, but that case fits

god gawd i should know this

its math

but it dont

:[

if x/75 = 10.2, adding 0.5 will round it up to 11, instead of 10 which would be more accurate

it will not

int(X) will take it down to 10

(int) always truncates?

int(10.7) = 10

yes

ah ok

ug

i think i understand copper

an ia gree with ihm

mo: about the track-times not adding up?

or about int being silly?

yes

if its 10,7 it sould be 11 not 10

(int)(X+0.5) is a common trick in programming for round(X)

because some programming languages doesn't have round :x

right

then yes that would be better indeed

i think i nee a cuppa tea to wake up

then again, from a practical standpoint, durations wouldn't match what users see in their players, because players truncate too

how lame players are

actually, they would

er, no they wouldn't

I'd think you'd still need to add a fudge factor somewhere if you wanted the track times and the album time to add up the same...

but the sums of track durations would match the final CD duration

which is probably more important

MartinRudat: it would still be imprecise and might not always match the total duration, but it would be more precise