TR Forums

Well, due to various threads in the recent past, plus problems that i've had recently with Musicmatch, my previous program of choice, I decided to do a little bit of testing, and figured that i would share with ya'll my endeavours in CD ripping with EAC.
My roomate has around 200 cd's that he wanted me to to put in mp3 for him, so I began this task

I thought that my DVD-RW drive would be the best choice, b/c of some previous thoughts that I had, however I was to be very much mistaken.

The test system: see sig
Windows XP Pro SP2
Software: Exact Audio Copy V0.95 prebeta 5 and lame-3.96.1 ("--alt-preset cbr 128") (NOTE: the purpose is not for music quality of the mp3's. I do not have that good of ears, hence I rip everything at 128kbps. Also it helps conserve the hard drive space.)
Test drives:
NEC 8X Black Dual Layer DVD+/-RW Drive, Model ND-2510A
CDRW 52X32X52|SONY CRX230E
DVDROM 16XMSI MS-P8216S
and an old 52x drive I pulled out of the closet

Test Audio CD: Jars of Clay: The Eleventh Hour, 42 minutes 26.30 seconds (NOTE: the cd is quite new and so is not likely to have many, if any errors. That is not the purpose of this test. It is almost for the pure speed aspect. I did, however listen to the mp3's that were created, and they sound flawless)

Here's screenshots of the configuration pages:















So anyway, Notice on the general tab, there is the option to have multiple threads run, well I did a test on that to see if it would make a difference.
This was on my CD-RW drive. The first time is the time until EAC reports the extraction complete, the second is (via my stopwatch) the time until the last LAME encoding is done. (I also tested with the threads disabled, just for reference)



As you can see, the only difference is when going from 1 to 2 simultaneous threads, as it keeps the CPU filled (also note I do not have dual processors, just a overclocked AthlonXP). By the time the CD drive is on the 10th track, all 4 of the threads are in use, so the last little bit is pure CPU speed, making up much of the difference between the extraction time and the encoding time.
Needless to say, I used the 4x threads in all the drive tests.

Screenshot of multiple LAME threads:



Another feature/note. I tested the cd ripping to both of my hard drives to see if the raptor would be any faster than the 7200rpm at holding the temporary data. (NOTE: these results were not done during the main run, so values can not be directly compared) The results:



Needless to say, there is very little difference in between the drives, so we know that that is not going to be a bottleneck. I went with the 120GB as that is where the files are going to be stored in the long run.

Now for the actual test. In the drive configuration, I chose the maximum multiple that was available, as well as the "allow speed reduction during extraction" option. For the drives, the maximums were:



As you will see, these values correlate very well with the times that are achieved. Not a linear scale, per se, but indeed the general trend will continue.

So anyway, the values that you have been waiting for. (NOTE: tests were not repeated due to my laziness and lack of overall belief that anything would change with a different CD or different setting)



So there you have it: The CD-RW and the DVD-ROM outpace the other drives and tie for the fastest ripping of my cd at 3 minutes and 39 seconds start to finish. Impressive, to say the least.




Feel free to comment or ask questions about the test, or whatever you wish.

Comments: Much thanks to derFunkenstein for help with configuring EAC

EDIT: updated links 03/23/07

just an update, LAME 3.96.1 has started wigging out on me...giving me illegal operation or something almost every CD, so I'm now using the 4.0a10 BETA version and that seems to be working fine...a little bit faster, too

Heh... didn't realize they were up to 4.xx on LAME. I'm still using the 3.93 version (which has not given me any trouble).

EAC rules, and YOU rule for putting in the work with screenies and all. Thanks.

I'm curious why you chose to use fast mode, rather than burst mode in conjunction with the Test and Copy option. Was this only for benchmarking purposes, or are these the settings you always use for ripping?

Test and Copy reads each track twice, generating a CRC after each read and only writing the file on the second read. While this sounds slower, burst mode is so fast that it still doesn't take much time, and if the CRCs match you can be certain that the track was ripped properly. If the CRCs don't match, you have a problem track, and can drop down to one of the more secure modes to re-rip it.

Without the Test and Copy function, you have no way of knowing if the track was successfully extracted. The use of fast mode makes it less likely that you'll have errors, but if you do you won't know it until you listen to the track.

Finally, I haven't used fast mode enough to comment with certainty on its speed relative to burst mode, but IIRC burst mode is fast enough that even with the dual reads it is typically quicker than fast mode. You may achieve different benchmark results using burst mode instead of fast mode as well.

Apologies if you already knew all this, but IMO the Test and Copy function of EAC is what makes the program the best ripper out there, and I would recommend always using that function when ripping tracks.

Or you can be like many of us and always use secure mode exclusively. After all, I'll take quality over speed every time.

Dr. Evil wrote:

I'm curious why you chose to use fast mode, rather than burst mode in conjunction with the Test and Copy option. Was this only for benchmarking purposes, or are these the settings you always use for ripping?

Test and Copy reads each track twice, generating a CRC after each read and only writing the file on the second read. While this sounds slower, burst mode is so fast that it still doesn't take much time, and if the CRCs match you can be certain that the track was ripped properly. If the CRCs don't match, you have a problem track, and can drop down to one of the more secure modes to re-rip it.

Without the Test and Copy function, you have no way of knowing if the track was successfully extracted. The use of fast mode makes it less likely that you'll have errors, but if you do you won't know it until you listen to the track.

Finally, I haven't used fast mode enough to comment with certainty on its speed relative to burst mode, but IIRC burst mode is fast enough that even with the dual reads it is typically quicker than fast mode. You may achieve different benchmark results using burst mode instead of fast mode as well.

Apologies if you already knew all this, but IMO the Test and Copy function of EAC is what makes the program the best ripper out there, and I would recommend always using that function when ripping tracks.

To be honest, I had never touched EAC until I started this endeavour and thus know little about the program...the advice I was given by derFunkenstein was to use the "fast" mode...i'll try the others and let you (as a collective TR community) know.
Thanks for the suggestion

EDIT: spelling

ok well here's an extension for ya'lls reading pleasure... Due to the suggestion by Dr. Evil, I tested the 3 different read modes using the copy, and the Test&Copy functions.
The results below are not directly comparable with the ones from the initial review. I used my DVD-ROM drive w/ the read speed set at 39.9x, allowing speed reduction.
I also used a different CD this time around. dc Talk, Supernatural. It is 56:50.65, comprised of 14 tracks totalling 573.80mb.
Other than that, almost all settings are analogous to those from the initial review, and I'm using LAME 4.0a10 BETA

anyway, the results:



As you can see, the test mode takes a good bit longer, but ensures a job well done. I may think about going that route from now on. Also, the extraction takes more time, thus the encoding doesn't get backed up, like it does when just copying. So there you have it. If you're willing to take another 40% of your time, then the test© is for you. Also notice that 'Burst' mode is almost a minute faster than 'Fast', so the settings of choice seem to be Test & Copy, using the Burst mode.

questions/comments/whatever feel free to leave.

Dag, yo, that was a helluva post you put together. I'm not that ambitious.

I told him to use Fast over Burst because Burst actually has run slower on my LiteOn SOHW-812S and on my Toshiba 4x DVD-RW (I forget the exact model), though it Burst was fairly useful on my Sony 52x CD-RW. You might also be able to get away with not spinning up the drive before you rip, but I've just observed the "x" factor when ripping to be lower when you do, though there's not the pause in between tracks, so it's a tradeoff. Seems to vary by drive

Ok well on that idea, i went back and tested my DVD-RW. This is with the same dc talk cd, so the results are comparable. I only did the the copy, figured the test would scale fairly linearly...so yeah



Well there ya have it.
Burst increases the speed a great deal...I'd go back and test my other drives, but I already put them away and it's a pain to have to open my case and unroute my cables and stuff, so yeah..i'm not that dedicated
One note, i disabled the 'drive caches audio data' from the DVD-RW b/c the wizard said that the drive didn't support that, so yeah...don't know if that makes a difference, but yeah

Man...what I'm getting from this is that I need to go out and buy a DVD-ROM or CD-RW to get the best speed out of ripping. My DVD-RW is dang slow compared to the speeds you're posting here...

A quick question about EAC... The "Test and Copy" option mentioned is called "Secure Mode" in the "Extraction Method" tab of the "Drive Options" menu, right? So to use both burst mod and T&C/Secure mode you just check both boxes?

-Rick

So... if hard disk space is a big consideration for you, why not use 128 ABR instead of CBR? You'll get considerably better sounding mp3s at the exact same size. I cannot think of a single reason to ever use CBR other than 320. Also, I would not recommend using fast or burst mode with EAC, it will save you some time now, but you'll end up with a bunch of mp3s with annoying pops in them.

Take some advice from someone who has had the misfortune to re-rip the same cd collection 3 times. Do it right the first time. You will get pickier over time, and then you'll just be stuck doing the job again. I only rip to --preset insane or use lossless encoding now, and I always use secure ripping.

yeah you really should be using VBR.

i never use anything but secure mode with EAC also. i've had a few CDs with scratching on a few songs and EAC managed to extract them perfectly!! it took a while though lol.

LAME's "--preset standard" option is really the way to go, unless you're trying to pack as much music as you can onto a flash-based portable (in which case you'll want to use a lower quality VBR setting), or you're really picky about quality (in which case "--preset extreme" would be a better bet).

128kbit CBR is far from optimal, even for low bitrate encoding.

I figure I might as well put my two cents in also. You should have normalize enabled. It'll keep you from having to adjust the volume from song to song (if you play the songs randomly)/

I guess the main reason to choose cbr is that it makes the math easier...trying to figure out how much I can fit on a CD (MP3-cd player, I have or on my 128mb mp3 player...just know that it's xxkbps, that translates to xx mb/minute...so to find out how much I can put on the player, just divide by the mb's and i can know how many minutes..then I just get a playlist in winamp and get one that is the exact right length to use up 99% of space and yeah.....not really any real reason, i probably will change to the --preset normal or whatever for the remainder of the cd's per ya'll suggestion....and about the normalized...I did that in musicmatch, and to me non-audiophile ears, it sounded like the quality was lessened a little bit...was probably just me though..I may do that again, but odn't really have a problem adjusting volume anyway, you know remote-control's and all...just from my experiences, it hasn't been a worthy trade-off in the past

and RickB...no, the test and copy is not a setting in the options. It is a way to do the ripping..you go to Action-> "Test and Copy selected tracks" instead of the "copy selected tracks" mode

[quote="Usacomp2k3"]and RickB...no, the test and copy is not a setting in the options. It is a way to do the ripping..you go to Action-> "Test and Copy selected tracks" instead of the "copy selected tracks" mode[/quote]

Thanks, I would never have spotted that. I always just click on the "MP3" button on the left side of the program window.

-Rick

RickB wrote:

Usacomp2k3 wrote:

and RickB...no, the test and copy is not a setting in the options. It is a way to do the ripping..you go to Action-> "Test and Copy selected tracks" instead of the "copy selected tracks" mode

Thanks, I would never have spotted that. I always just click on the "MP3" button on the left side of the program window.

-Rick

I didn't either...thank Dr. Evil for pointing it out to me

ok well now I'm having other problems...My computer seems to like freezing on my now. This is only when I'm using EAC...the temp never gets about 53, but it just freezes mid-rip sometimes...I'm pretty sure the problem is with EAC, but I don't know for sure...I don't think it's LAME, but could be...any suggestions/thoughts?

BTW, should this thread be in echo vale or general software?...it's not really hardware, except for testing the drives

well I'm finally getting around to finishing this process....so I'm currently running 3 different copies of the same program, all ripping off of different drives (my DVD-ROM, DVD-RW, and CD-RW)...they all go at the same time, so averaging it out, I'm churning through a cd ever minute 15 sec or so

too bad my poor cpu is underclocked....it gets backed up pretty quickly

EDIT: never was really able to get this to work stabilly for more than 1/2 a cd..bummer