and you said the zeroes are at the end of the file
2007-09-06 24948, 2007
ojnkpjg
yeah
2007-09-06 24956, 2007
ojnkpjg
which is why it doesn't make a lot of sense
2007-09-06 24914, 2007
ojnkpjg
but it's kind of interesting that's the only track with a gap (other than 1) and that's the only difference
2007-09-06 24916, 2007
hawke
but cdparanoia reads the track starting after the pregap. -- I could see a CD drive returning nulls when you try to read past the end of the CD
2007-09-06 24927, 2007
ojnkpjg
on the last cd i compared, there were no diffs, and no tracks had pre-gaps other than 1
2007-09-06 24942, 2007
canidae
hawke: but the file size is the same for both eac and cdparanoia
2007-09-06 24949, 2007
ojnkpjg
hawke, but then you'd expect to see differences elsewhere
2007-09-06 24957, 2007
canidae
if eac "detects" this gap and skip it, the file should be smaller
2007-09-06 24911, 2007
canidae
so there's something fishy here :p
2007-09-06 24926, 2007
canidae
ojnkpjg: is that track the only track with pre-gap?
2007-09-06 24929, 2007
ojnkpjg
yep
2007-09-06 24933, 2007
ojnkpjg
other than 1
2007-09-06 24934, 2007
hawke
canidae: right -- (perhaps) eac gives pregap+track, while cdparanoia gives track+nulls the length of the pregap
2007-09-06 24949, 2007
hawke
This is of course mostly speculation
2007-09-06 24903, 2007
ojnkpjg
i think its presence must be messing some other calculation up
2007-09-06 24926, 2007
ojnkpjg
might be a simple bug, but i've got to go now
2007-09-06 24937, 2007
canidae
i gotta go too, meeting in a minute
2007-09-06 24946, 2007
canidae
i'd love to look at this later tonight, if i get the time
2007-09-06 24908, 2007
canidae
just don't talk down on cdparanoia ;)
2007-09-06 24944, 2007
brianfreud
I guess the real question is, is that very very minor rip difference between the two even audibly or spectrographically noticable, esp if you're not one who stores in flac/ape/shn?
2007-09-06 24927, 2007
brianfreud
*s/noticable/significant, if you like
2007-09-06 24929, 2007
hawke
I believe the answer to that would be "no"
2007-09-06 24910, 2007
hawke
since it will only affect the last 80 bytes of the last track on a CD (or perhaps any track with pre-gap)
2007-09-06 24918, 2007
dholmes
Yes, but if you're ripping FLAC in the first place, that's not necessarily what concerns you
2007-09-06 24905, 2007
hawke
dholmes: oh?
2007-09-06 24913, 2007
brianfreud
that's my point - unless you're storing in the most insane FLAC settings, that 80 byte rip difference is utterly insignificant
2007-09-06 24936, 2007
dholmes
You could accomplish songs that are not audibly noticeable with high-bitrate MP3s, too
2007-09-06 24939, 2007
hawke
even if you are, it's insignificant
2007-09-06 24948, 2007
brianfreud
exactly
2007-09-06 24904, 2007
hawke
dholmes: my point is that it's not going to affect the whole song
2007-09-06 24931, 2007
dholmes
I rip FLAC because I want to be able to throw the CD in the trash, so I like to be 100% confident that every bit is being copied correctly
2007-09-06 24941, 2007
hawke
even at the most extreme, if it made a loud pop at the end of every affected track or something...it wouldn't affect the audio quality of everything else
2007-09-06 24944, 2007
dholmes
Obviously I won't notice it looking at it, but I want it there anyways
2007-09-06 24951, 2007
hawke
dholmes: oh, in that case can you mail me your CDs? ;-)
2007-09-06 24909, 2007
dholmes
That would be copyright infringement. tch tch :P
2007-09-06 24931, 2007
brianfreud
dholmes, are you also testing the FLAC decompressor to ensure it's decoding audio that is bit for bit identical?
2007-09-06 24934, 2007
dholmes
I don't actually throw them away atm; they just get buried in boxes
2007-09-06 24951, 2007
hawke
It is anyway -- you're transferring your physical copy to someone else (the trash collector, or the city). ;-)
2007-09-06 24955, 2007
brianfreud
lossless and bit-for-bit identical don't have to mean the same thing :P
2007-09-06 24903, 2007
hawke
brianfreud: they don't?
2007-09-06 24913, 2007
hawke
brianfreud: I think that's kind of the definition of lossless
2007-09-06 24924, 2007
hawke
that it doesn't lose any data
2007-09-06 24925, 2007
dholmes
I was under the impression that in this particular case, they do
2007-09-06 24935, 2007
brianfreud
depends on what definition of lossless is being used
2007-09-06 24958, 2007
brianfreud
bit-for-bit identical, or "doesn't throw away any audio info", etc
2007-09-06 24909, 2007
dholmes
I'm fairly certain that wav->flac->wav will result in the same audio, bit-for-bit, and the FLAC codec has a mode for verifying this
2007-09-06 24938, 2007
brianfreud
I'd have to look at how flac exactly compresses
2007-09-06 24923, 2007
Kerensky97
Well if it's not reversable it's not really lossless is it?
2007-09-06 24939, 2007
Kerensky97
Because something is lost as some point.
2007-09-06 24943, 2007
hawke
yeah
2007-09-06 24951, 2007
dholmes
Well, depending on the starting format, there could be two representations of equivalent audio data, but that's not the case here
2007-09-06 24903, 2007
hawke
dholmes: eh?
2007-09-06 24914, 2007
dholmes
I just mean at a hypothetical level
2007-09-06 24917, 2007
hawke
you mean like wav vs. au or something?
2007-09-06 24926, 2007
hawke
or wav vs. CDDA?
2007-09-06 24952, 2007
dholmes
Nah, I mean like FLAC at one compression level vs. FLAC at another level
2007-09-06 24910, 2007
dholmes
But the data is equivalent either way, and in wav will be the same
2007-09-06 24929, 2007
brianfreud
that's my point - unl;ess it's using different compression routines themselves, how can you have multipe compression levels?
2007-09-06 24901, 2007
dholmes
FLAC has all manner of levels
2007-09-06 24923, 2007
dholmes
For compression/speed trade-offs
2007-09-06 24941, 2007
brianfreud
from the flac site, this seems to indicate varying degrees of bit-for-bit reversability - note that there is an approximation in use, even if they try to work around it:
2007-09-06 24943, 2007
brianfreud
"Once the model is generated, the encoder subracts the approximation from the original signal to get the residual (error) signal. The error signal is then losslessly coded. To do this, FLAC takes advantage of the fact that the error signal generally has a Laplacian (two-sided geometric) distribution, and that there are a set of special Huffman codes called Rice codes that can be used to efficiently encode these kind of signals
2007-09-06 24957, 2007
dholmes
I'm not an expert on how the codec works, but I am 100% sure that it is 100% lossless
2007-09-06 24954, 2007
dholmes
I'm assuming that the approximation combined with the error produces the original
2007-09-06 24940, 2007
brianfreud
except they're generating the error number from the approximation to then be able to reverse the formula to return to the approximation
2007-09-06 24919, 2007
brianfreud
but that implies they can't get back to better than the FLAC approximation, rather than 100% bit for bit identical
2007-09-06 24952, 2007
dholmes
If that were the case, the entire format would be close to useless. It would defeat the purpose.
2007-09-06 24956, 2007
hawke
yeah
2007-09-06 24907, 2007
hawke
I think flac is more equivalent to gzip for audio...
2007-09-06 24914, 2007
Russss
lossless == bit-for-bit identical
2007-09-06 24914, 2007
hawke
and the compression ratios back that up
2007-09-06 24943, 2007
dholmes
"Lossless: The encoding of audio (PCM) data incurs no loss of information, and the decoded audio is bit-for-bit identical to what went into the encoder. Each frame contains a 16-bit CRC of the frame data for detecting transmission errors. The integrity of the audio data is further insured by storing an MD5 signature of the original unencoded audio data in the file header, which can be compared against later during decoding or testing
I wonder why one of those tracks is 2:20 and the others are 3:33
2007-09-06 24919, 2007
brianfreud
no idea, but mine is 3:28, so I'd guess the 2:20 is wrong
2007-09-06 24919, 2007
FauxFaux wonders if it's legal to execute people who release two (or more) versions of the same cd yet.
2007-09-06 24952, 2007
dholmes
Unfortunately that might include most people who release music these days
2007-09-06 24913, 2007
FauxFaux
The practice'll die out after they spot the trend, I guarnatee it.
2007-09-06 24957, 2007
FauxFaux
Hehe at the SQL Server 2005; "Hilton analyzes 1.4 million records a day"
2007-09-06 24909, 2007
FauxFaux
Someone go count(*) from trm? :p
2007-09-06 24926, 2007
ruaok
9846433
2007-09-06 24942, 2007
ruaok waves as he tries to wake up
2007-09-06 24958, 2007
FauxFaux remembers that the site actually offers the stats panel, which quoth 1,442,4710. Heh heh heh.
2007-09-06 24901, 2007
FauxFaux
Hi ruaok!
2007-09-06 24903, 2007
outsidecontext
hi ruaok
2007-09-06 24907, 2007
FauxFaux
Wait, I fail at commas.
2007-09-06 24918, 2007
ruaok
hi kids!
2007-09-06 24920, 2007
Amberrock has quit
2007-09-06 24916, 2007
hawke
FauxFaux: so is that 14.4 million, or 1.4 million?
2007-09-06 24947, 2007
FauxFaux
14424710, 14 million, it seems.
2007-09-06 24910, 2007
warp
hm, missed interesting bits about cdparanoia and eac.
2007-09-06 24928, 2007
dholmes
Got anything interesting to add?
2007-09-06 24945, 2007
warp
i was under the impression that for accuraterip to work, you need to take care of some offset each drive has when reading audio. i'm wondering how ojnkpjg managed to compare a cdparanoia rip with an eac rip with the accuraterip database.
2007-09-06 24938, 2007
Aankhen``
AR doesn't work at all if you're not in the US.
2007-09-06 24947, 2007
warp
Aankhen``: hm?
2007-09-06 24947, 2007
Aankhen``
(Or rather, buying CDs from the US.)
2007-09-06 24930, 2007
warp
Aankhen``: i'm in the netherlands, the popular cds i've tried matched.
2007-09-06 24937, 2007
Aankhen`` shrugs.
2007-09-06 24945, 2007
Aankhen``
Maybe it just doesn't work at all if you're in India. :-)
2007-09-06 24950, 2007
Aankhen``
(No matching pressings for any album.)
2007-09-06 24908, 2007
warp
Aankhen``: i don't really need every CD to match, I do want some CDs to match so that I know that the process of ripping is reliable.
2007-09-06 24924, 2007
Aankhen``
Every CD doesn't need to match, but I never got a single match.
2007-09-06 24935, 2007
Aankhen``
True, I only tried about 100 CDs.
2007-09-06 24940, 2007
warp
Aankhen``: you did try really popular artists? :)
2007-09-06 24944, 2007
Aankhen``
Aye.
2007-09-06 24903, 2007
warp
i didn't get matches on obscure jpop, but iirc bjork and madonna worked.
2007-09-06 24916, 2007
Aankhen``
Plenty of entries for the album, but nothing that matched, IIRC (it's been a while).
2007-09-06 24932, 2007
warp
oh, weird :S
2007-09-06 24918, 2007
warp
i whish those ppl would be a bit more open though.
2007-09-06 24908, 2007
srotta
Aankhen``: I've had matches on Finnish artists, and my guess is they're not that big in US.
2007-09-06 24937, 2007
srotta
Not for every album, true. And I'm not convinced about the idea in general. But anyway.
2007-09-06 24941, 2007
mbrain
there's a bug in EAC, which sometimes makes the accuraterip not match any discs
2007-09-06 24939, 2007
mbrain
I had the same situation before (accuraterip not matching any key discs), but if I recall correctly, I got it working by making the match in dBPowerAmp and copying the dll file to EAC's folder..
2007-09-06 24927, 2007
canidae is back
2007-09-06 24959, 2007
canidae
just of curiousity, afaik multiple cd's may have the exact same toc, how can accuraterip distinguish them?
2007-09-06 24933, 2007
FauxFaux
Surely all it does is hashes all the audio data that you get off the disc?
2007-09-06 24951, 2007
Aankhen`` has quit
2007-09-06 24955, 2007
FauxFaux
Use the freedb-style-id to get a small set of things to compare it again?
2007-09-06 24905, 2007
canidae
but isn't it so that the data you read may be wrong because of gaps or whatever? or does that only apply to audio data?
2007-09-06 24927, 2007
canidae
or well, that wouldn't matter
2007-09-06 24929, 2007
FauxFaux doesn't know, I don't care /that/ much. :)
2007-09-06 24925, 2007
canidae
frankly, me neither. if the last 40 samples of the last track of my cd's are "wrong" i wouldn't be able to tell
2007-09-06 24905, 2007
canidae
well, unless it made a real nasty static
2007-09-06 24920, 2007
FauxFaux
Just cut them off! <3
2007-09-06 24945, 2007
canidae
still, i'm quite curious about the differences in the last 40 samples from eac and cdparanoia
2007-09-06 24958, 2007
canidae
and if it applies to every cd, and whether it always is 40 samples
2007-09-06 24953, 2007
canidae
and well, to me it's more important that a damaged cd is ripped without jitter than that the last 40 samples are zeroed out since i can't be able to tell unlike i can with a rip that's full of jitter errors
2007-09-06 24957, 2007
trollomat joined the channel
2007-09-06 24927, 2007
FauxFaux
My drive has a tendency to fail the rip and put the worst noise I've ever heard into my rips.
2007-09-06 24912, 2007
FauxFaux
I know there's one somewhere on this double-cd, give me two hours. :p
