Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Eli's Suggested EAC settings for FLAC (Read 482679 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Eli's Suggested EAC settings for FLAC

Reply #25
from what i understand, the "more sophisticated" settings of eac (accurate stream, c2, cache) only serve to speed the process up if the drive supports it. turning accurate stream and c2 off and turning "drive caches audio" on will increase the safety of the rip while making the process slower. what were your other settings when you got different checksum each time while not changing any settings? from personal experience i can tell that c2 error checking with lite-on drives is to be avoided.

Eli's Suggested EAC settings for FLAC

Reply #26
Quote
from what i understand, the "more sophisticated" settings of eac (accurate stream, c2, cache) only serve to speed the process up if the drive supports it. turning accurate stream and c2 off and turning "drive caches audio" on will increase the safety of the rip while making the process slower. what were your other settings when you got different checksum each time while not changing any settings? from personal experience i can tell that c2 error checking with lite-on drives is to be avoided.
[a href="index.php?act=findpost&pid=317922"][{POST_SNAPBACK}][/a]

I was using a Lite On drive , but I was using the safest (slowest) settings:
Secure Mode with Accurate Stream,  Drive Caches Audio checked, C2 not checked.

I was sometimes getting matching CRC's with T&C, sometimes not matching. I did notice the CRC's it came up with would occur again, so it wasnt totally random. Mind you, the drive would slow down considerably at the rough spots of that track.

Eli's Suggested EAC settings for FLAC

Reply #27
Quote
Hi Martin,

very good summary, why I prefer ripping to single tracks, too, no huge image file.

Thank's user.   
Quote
May I copy your text, and implement(&reference to your nickname) it one day to www.High-Quality.ch.vu , if I find time for an update ?

Yes, and thank you very much. 
Quote
Keep in mind the 1st and last track may have different hashes because of a drive's inability to overread into lead in/lead out.

AccurateRip cuts 4 sectors of the start of the first track and 4 sectors of the end of the last track when calculating CRC's to avoid overreading issues...

-Martin.

Eli's Suggested EAC settings for FLAC

Reply #28
Quote
Quote
Hi Martin,

very good summary, why I prefer ripping to single tracks, too, no huge image file.

Thank's user.   
Quote
May I copy your text, and implement(&reference to your nickname) it one day to www.High-Quality.ch.vu , if I find time for an update ?

Yes, and thank you very much. 
Quote
Keep in mind the 1st and last track may have different hashes because of a drive's inability to overread into lead in/lead out.

AccurateRip cuts 4 sectors of the start of the first track and 4 sectors of the end of the last track when calculating CRC's to avoid overreading issues...

-Martin.
[a href="index.php?act=findpost&pid=317936"][{POST_SNAPBACK}][/a]



OK...I am back and need some more help.  As I mentioned earlier in this thread, I am getting good results with EAC to FLAC using AccurateRip.  I am, however, getting some wierd AccurateRip results at times.  According to AccurateRip, the vast majority of CD's that I am ripping are either accurately Ripped, not in the database, or a different pressing.  There are several, however, that are giving me accurate rips on individual tracks and then non-accurate rips on other tracks.  In some cases the non-accurate rips are the first and/or the last tracks on the CD.  In other cases, the non-accurate track is a random track on the middle of the CD and all other tracks are stating that they are accurately ripped.

I assumed that the non-accurate rips on the first and the last track were related to the fact that my drive over-reads into lead-in and lead-out and that this might throw off the results (I get no errors occurred from EAC).  But, what about tracks in the middle of a CD?  There is one more wierd thing...when I take one of the tracks that gets a non-accurate rip from AccurateRip and rip it individually as opposed to ripping it with the whole CD, Accurate now tells me it is accurately ripped.  What is going on?

Eli's Suggested EAC settings for FLAC

Reply #29
Quote
I assumed that the non-accurate rips on the first and the last track were related to the fact that my drive over-reads into lead-in and lead-out and that this might throw off the results (I get no errors occurred from EAC).

AccurateRip cuts 4 sectors from the beginning of the first track, and 4 sectors from the end of the last track when calculating CRC's, to avoid overreading issues...
Quote
There is one more wierd thing...when I take one of the tracks that gets a non-accurate rip from AccurateRip and rip it individually as opposed to ripping it with the whole CD, Accurate now tells me it is accurately ripped.  What is going on?

Re-ripping corrected the error...

-Martin.

Eli's Suggested EAC settings for FLAC

Reply #30
Quote
Quote
I assumed that the non-accurate rips on the first and the last track were related to the fact that my drive over-reads into lead-in and lead-out and that this might throw off the results (I get no errors occurred from EAC).

AccurateRip cuts 4 sectors from the beginning of the first track, and 4 sectors from the end of the last track when calculating CRC's, to avoid overreading issues...

Quote
There is one more wierd thing...when I take one of the tracks that gets a non-accurate rip from AccurateRip and rip it individually as opposed to ripping it with the whole CD, Accurate now tells me it is accurately ripped.  What is going on?

Re-ripping corrected the error...

-Martin.
[a href="index.php?act=findpost&pid=328968"][{POST_SNAPBACK}][/a]


Thanks for the reply.  Sorry for being dense, but I am not following.  Re-ripping only corrects the error when I rip the track individually.  If I rip all the tracks on the CD at once, I get the same error.  Why would I get different results?

Eli's Suggested EAC settings for FLAC

Reply #31
Thanks for the reply.  Sorry for being dense, but I am not following.  Re-ripping only corrects the error when I rip the track individually.  If I rip all the tracks on the CD at once, I get the same error.  Why would I get different results?
[a href="index.php?act=findpost&pid=328969"][{POST_SNAPBACK}][/a]
[/quote]

Have you carried out reripping cd at once 10 times, and the suspicious tarck ripping individually, too, 10 times ?
Only then you could claim, that only ripping track individually leads to accurate rip.
I assume, ther eis  alight scartch/damage on the CD, which would lead in EAC to crc mismatches/accurate rip to inaccurate rip, but which can be corrected, read out correctly by chance, ie. by rerip in EAC (you will notice by comparing crcs in eac logs, that 1 of the crcs happened already in the 1st mismatch rip), or by rerip in accurate rip.

Eli's Suggested EAC settings for FLAC

Reply #32
Only a matching CRC in EAC or a match in AccurateRip will actually tell you if you have ZERO errors in the read. EAC may report no errors but still not have a perfect rip.

Eli's Suggested EAC settings for FLAC

Reply #33
encoding to -5 vs -8, on most modern computers will see little difference in encoding time. Plus encoding is a 1 time thing, so its my opinion that a little space savings is worth it. That being said, -5 is perfectly acceptable


That said, -5 is totally unacceptable. if you are too lazy to wait the almost unnoticable time to have smaller files that will take less disk space on your hard drive and/or DAP then don't even bother in the first place.

Eli's Suggested EAC settings for FLAC

Reply #34

encoding to -5 vs -8, on most modern computers will see little difference in encoding time. Plus encoding is a 1 time thing, so its my opinion that a little space savings is worth it. That being said, -5 is perfectly acceptable


That said, -5 is totally unacceptable. if you are too lazy to wait the almost unnoticable time to have smaller files that will take less disk space on your hard drive and/or DAP then don't even bother in the first place.


I've observed different benchmaks.  The difference between -5 and -8 speed is tremendous.

Are you guys using special optimized compiled builds of FLAC?  Specific CPUs?

On a IBM ThinkPad laptop, I see the following throughputs for FLAC (converting 465 gigabytes worth of WAVs -- about 850 CDs):
  • flac -1 :  8.7 megabytes/sec processed WAVs (14.7 hours total to encode)
  • flac -5 :  6.3 megabytes/sec processed WAVs (20.3 hours total to encode)
  • flac -8 :  1.7 megabytes/sec processed WAVs (73 hours total to encode)




I get slightly faster results on Intel D865PERL 3.2GHz and Intel D975XBX D940 3.2GHz  --- but the performance ratio between -5 and -8 stay the same.  I don't have AMD processors to compare to.  I also don't think the new Core 2 Duo procs will change the ratio of performance between -5 and -8 either.

So the difference between -5 (less than a day to encode 850 CDs) vs -8 (more than 3 days to encode) is more than 2 days!  Waiting 49 extra hours for number crunching seems pretty significant to me.  But I guess it's all relative.

So what's your threshold for "almost unnoticable" ? 

Eli's Suggested EAC settings for FLAC

Reply #35
My experience is not based on such large scale encodings and certainly the time differnce adds up to be significant. However, when ripping and encoding, my encoding is typically able to keep pace with the ripping as the ripping tends to be the limiting factor.

Eli's Suggested EAC settings for FLAC

Reply #36
My experience is not based on such large scale encodings and certainly the time differnce adds up to be significant. However, when ripping and encoding, my encoding is typically able to keep pace with the ripping as the ripping tends to be the limiting factor.


Ah yes... that's the key point I overlooked.  I never encode as I rip.

I also had to encode to FLAC twice:  once as single file FLAC images and once again as FLAC individual tracks so I was hyper-aware of how fast FLAC actually works.  I also have large-scale benchmarks for LAME and NeroAAC ... they're slow too but not quite as slow as flac -8.

Eli's Suggested EAC settings for FLAC

Reply #37
Let me see if I understood this thread:

To get a bit-by-bit perfect CD copy:
(Considering EAC/AccurateRip/keydisc/Offset number are all configured and OK)

#1 - You MUST rip in ANY MODE but it has to be with Test & Copy, and it MUST match the CRCs.
#2 - You MUST have an AccurateRip log saying tracks are Accurately Ripped with, preferably, a very high confidence number (let's say over 30).
#3 - Your MUST have the "Over-read lead in / lead out" feature supported by your drive, and if it doesn't support it, nothing that you've done to this point guaranteed you a actual perfect rip. (Eg. If I have a LEAD-OUT only drive, I am fooling myself to think that my rips are perfect.)

About the Lead-In/Lead-Out feature that is required to achieve 100% perfect CD copies (a feature that only Plextors have got):

I am trying to understand here... You mean "100% PERFECT RIPS" as a PERFECT CD layer copy. Or do you mean PERFECT AUDIO DATA copy?

Does this lead-in/lead-out uncapacity TOUCHES the result of actual AUDIO DATA? if my drive has only the LEAD OUT or LEAD IN can it still extract perfect audio data, but not perfect cd layer copy?

Eli's Suggested EAC settings for FLAC

Reply #38
#1 - You MUST rip in ANY MODE but it has to be with Test & Copy, and it MUST match the CRCs.
-Test & Copy does not guarantee you an error-free copy, regardless of the mode used.
-It is possible to get an error-free copy without having ever generated a test CRC.

#2 - You MUST have an AccurateRip log saying tracks are Accurately Ripped with, preferably, a very high confidence number (let's say over 30).
-This is about all you need, though I'd say a confidence of 1 is adequate if you know that you aren't comparing to your own submission.  If you are comparing to your own submission then this is the same as getting matching CRCs from a Test & Copy.

#3 - Your MUST have the "Over-read lead in / lead out" feature supported by your drive, and if it doesn't support it, nothing that you've done to this point guaranteed you a actual perfect rip. (Eg. If I have a LEAD-OUT only drive, I am fooling myself to think that my rips are perfect.)
-Per your example of the lead-out, this only affects the very end of the last track on the disc.  If the samples that require overreading are silent, this setting makes no difference.

About the Lead-In/Lead-Out feature that is required to achieve 100% perfect CD copies (a feature that only Plextors have got):
-The part of your quote that I put in bold print is false.

I am trying to understand here... You mean "100% PERFECT RIPS" as a PERFECT CD layer copy. Or do you mean PERFECT AUDIO DATA copy?
So far, with the points that you have raised, only perfect audio data.

Does this lead-in/lead-out uncapacity TOUCHES the result of actual AUDIO DATA? if my drive has only the LEAD OUT or LEAD IN can it still extract perfect audio data, but not perfect cd layer copy?
Audio data and hence not a perfect copy, but again only for discs that have non-silent data in the area requiring overreading.

Eli's Suggested EAC settings for FLAC

Reply #39
There is a difference between getting a perfect rip and knowing you have a perfect rip. Much of the guide is dedicated toward KNOWing you have a perfect rip. There are a few ways to judge your rip:

-quality of rip in secure mode
-matching crcs in T&C
-matching AR results 1 is ok, anything >2 (not your submissions) is great and probably as good as anything greater

You can get perfect rips with just about any ripper, but the thing is there is almost know way to judge the quality of your rip with most rippers.

Eli's Suggested EAC settings for FLAC

Reply #40
let me see if I understood this...

Not even Secure Mode (with or without drive features enabled), not even burst mode, not even CRC checks, not even re-ripping and getting that same CRCs, not even testing & copying instead of just copying, etc, NONE of this will ACTUALLY TELL YOU you KNOW you have a perfect rip... this according to what greynol said? 

And the only saviour in this story is AccurateRip if you get like 20 confidences per track (like all my tracks say) will actually tell you you KNOW you have a perfect copy... and apart from that, "PERFECT" as in a PERFECT CD LAYER copy, sector by sector, a 100% physical replication, and if all before is not PERFECT, if you get a sample filled with silence, etc, the ACTUAL AUDIO DATA might be absolutely perfect !!??

Eli's Suggested EAC settings for FLAC

Reply #41
I use Test & Copy in burst mode or in secure mode w/C2 and -usefua using my PX-716A and normally rely on matching CRCs to give me some confidence that there wasn't an error if AccurateRip can't give me a conclusive result.

What Eli is failing to tell you is that errors can be consistent, in which case you'll get matching CRCs, yet the result is that you have an error.  IOW, a pair of matching CRCs is NO GUARANTEE that a rip is without error.

I have seen consistent errors while using secure mode though I have never personally identified a consistent error while using burst mode, though this is certainly still possible.

With AccurateRip, I see no reason why it would take a confidence of greater than one to guarantee yourself a perfect rip if you aren't comparing to a result that you provided previously.  What are the chances that someone else is going to rip a track in error in exactly the same way?

Also, with AccurateRip, be aware that the first 5 audio frames of the first track and last 5 audio frames of the last track are not compared in order to compensate for the lack of overreading between different drives.  In addition to HTOA and perhaps very short tracks (like filler tracks), it is clear that AccurateRip is not set up to compare all the audio data from a CD.

As far as "PERFECT CD LAYER" vs. "PERFECT AUDIO DATA" one is a subset of the other.  My focus here is on the latter.

Eli's Suggested EAC settings for FLAC

Reply #42
What Eli is failing to tell you is that errors can be consistent, in which case you'll get matching CRCs, yet the result is that you have an error.  IOW, a pair of matching CRCs is NO GUARANTEE that a rip is without error.


This is correct, but very very rare and also why I use so many other methods to detect an accurate rip. I have had a few rips that EAC reports as 99.8 or 9 (never 100)% quality, that AR has been able to show as inaccurate. I have been able to idenify maybe 1 or 2 tracks in about 30k+ track rips that have had matching CRCs with T&C (I personally use secure mode with -usefua in T&C).

Eli's Suggested EAC settings for FLAC

Reply #43
This is correct, but very very rare and also why I use so many other methods to detect an accurate rip. I have had a few rips that EAC reports as 99.8 or 9 (never 100)% quality, that AR has been able to show as inaccurate. I have been able to idenify maybe 1 or 2 tracks in about 30k+ track rips that have had matching CRCs with T&C (I personally use secure mode with -usefua in T&C).

I've seen plenty of 100% rips that AR says is inaccurate.  Some have been consistent errors, some haven't been.  When AccurateRip says something isn't accurate it isn't always correct.  You may have a different pressing or the results may be skewed through questionable submissions (the database is not free from CD-R entries or piracy).  When it says a rip isn't accurate, the confidence of the inaccuracy isn't as meaningful as the confidences for the tracks from the disc that it says were accurate.

I've also ripped many thousands of tracks in addition to seeing over a thousand log files for rips I did not make.  It isn't about how many tracks you've ripped, it's about the types of defects your discs have.

Sorry for taking this thread this far off-topic.

Eli's Suggested EAC settings for FLAC

Reply #44
greynol, I think we are saying the same thing. I basically use T&C for when AR does not have the disc in the database. While T&C is not perfect, it is very reliable. Same goes for EAC in secure mode - not perfect but very reliable. My recs basically set up a number of ways to verify: AR - this is the best; T&C - reliable; Secure Mode - reliable

Eli's Suggested EAC settings for FLAC

Reply #45
About used commandline options:
Why -8? If we'll read comparison at http://flac.sourceforge.net/comparison.html, we can see that -8 gives 55min to compression VS 12min at -5, and win 2.5MB from 413MB total. Is here a sense?

Even -5 is too much compared what you win with it. Best ratio between encoding-speed and resulting size brings -4.
My used codecs and settings:
FLAC V1.1.2 -4 / APE V3.99 Update 4 -high / MPC V1.15v --q 5 / LAME V3.97b2 -V2 --vbr-new / OGG aoTuV V4.51 Lancer -q5

Eli's Suggested EAC settings for FLAC

Reply #46
well, I read your explanations.... but the funny thing is that they are telling me exactly what I said...

I came to the conclusion that if AccurateRip does not show a confidence of at least, let's say, 10... you have no way to claim the show EAC performed was accurate, (well I know over 3 is probably a good number of confidence, but let's not give margin to errors...) of course it might look reliable, and you might take a guess that it was reliable. Sorry but that's what I conclude!

Eli's Suggested EAC settings for FLAC

Reply #47
Quick question: does the "Spin Up Drive Before Ripping" option always have to be checked, or is it drive dependent? How could one test whether or not it's causing a problem? (disabling it almost halves the time my Pioneer A05 takes to rip an album..)

Eli's Suggested EAC settings for FLAC

Reply #48
Quick question: does the "Spin Up Drive Before Ripping" option always have to be checked, or is it drive dependent? How could one test whether or not it's causing a problem? (disabling it almost halves the time my Pioneer A05 takes to rip an album..)


no, unchecking it will usually make the ripps faster. In fact I dont use it unless I am having a problem with a rip and have had a number of problem rips that spin up fixed. I made it the default to make the guide easier for new users to follow.

Quote
I came to the conclusion that if AccurateRip does not show a confidence of at least, let's say, 10... you have no way to claim the show EAC performed was accurate, (well I know over 3 is probably a good number of confidence, but let's not give margin to errors...) of course it might look reliable, and you might take a guess that it was reliable. Sorry but that's what I conclude!


You are correct. You cannot say with with 100% confidence that the rip is !00% accurate. But if you have a rip quality of >99.8% and matching T&C you cannot do any better. If you are really uncertain and have multiple drives you can do a Test in one drive and the Copy in a second drive which is basically the same as AR.

Eli's Suggested EAC settings for FLAC

Reply #49
@Eli, yes we agree.  I am more confident with matching CRCs in burst over secure from personal experience.  Of course burst can't do anything for a disc with errors.  If you're paranoid, secure rips with less than 100% quality where rereads don't occur at the end of the track deserve a test CRC, IMO.

@Bourne, how much margin for error is one really giving for an accurate rip confidence of 1 when it was a submission from a different disc?  A consistent error from the same disc is one thing, but from two different discs???