IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
Apple Lossless Encoder, that does not require iTunes / QuickTime
spoon
post Mar 19 2006, 23:22
Post #1


dBpowerAMP developer


Group: Developer (Donating)
Posts: 2725
Joined: 24-March 02
Member No.: 1615



After many months of wrangling with issues with the ALAC format (iTunes would read the files fine, but not iPods), Illustrate can proudly announce a worlds first, the first non-Apple dependant Apple Lossless Encoder.

This codec does not require iTunes, or Quicktime and has the advantage of being faster than Apples encoder! (x50 encoding speed on my AMD 64).

Download: ALAC Beta


--------------------
Spoon http://www.dbpoweramp.com
Go to the top of the page
+Quote Post
kurtnoise
post Mar 19 2006, 23:43
Post #2





Group: Members
Posts: 326
Joined: 26-June 02
From: Aix-en-Provence
Member No.: 2400



Thanks spoon for your work... smile.gif This is really a good news.

I suppose that the sources of this encoder are widely closed ?

This post has been edited by kurtnoise: Mar 19 2006, 23:48


--------------------
http://www.unite-video.com/phpbb/viewtopic.php?t=5412 :: An overview of all lossless Audio Formats (in french language ;-)
Go to the top of the page
+Quote Post
spoon
post Mar 19 2006, 23:50
Post #3


dBpowerAMP developer


Group: Developer (Donating)
Posts: 2725
Joined: 24-March 02
Member No.: 1615



The sources are held closed until a certain date, I am currently enquiring when that date is.


--------------------
Spoon http://www.dbpoweramp.com
Go to the top of the page
+Quote Post
Sgt_Strider
post Mar 20 2006, 07:06
Post #4





Group: Members
Posts: 256
Joined: 29-July 02
From: Vancouver, Canada
Member No.: 2854



Do you think the final version of this encoder will compress files even smaller? Or I should just use this?
Go to the top of the page
+Quote Post
xmixahlx
post Mar 20 2006, 09:07
Post #5





Group: Members
Posts: 1394
Joined: 20-December 01
From: seattle
Member No.: 693



/me awaits the source or linux binary smile.gif


later


--------------------
RareWares/Debian :: http://www.rarewares.org/debian.html
Go to the top of the page
+Quote Post
NumLOCK
post Mar 20 2006, 11:40
Post #6


Neutrino G-RSA developer


Group: Developer
Posts: 852
Joined: 8-May 02
From: Geneva
Member No.: 2002



QUOTE (xmixahlx @ Mar 20 2006, 09:07 AM)
/me awaits the source or linux binary smile.gif


later
*

If/when the source is released, I'd be glad to contribute a gentoo package for it smile.gif


--------------------
Try Leeloo Chat at http://leeloo.webhop.net
Go to the top of the page
+Quote Post
Alex B
post Mar 21 2006, 12:31
Post #7





Group: Members
Posts: 1303
Joined: 14-September 05
From: Helsinki, Finland
Member No.: 24472



I tested the performance with my 25 test tracks (the same set I have used previously for bitrate evaluations). It seems that this new encoder works like the other lossless codecs when a faster mode is selected. It is clearly faster, but the compression ratio is not as good.

The source files were in Monkey's Audio (Normal) format. I used the dBpowerAMP iTunes module and the new ALAC encoder. The test PC was a 2.8 GHz P4 on XP.

dBpowerAMP iTunes
Encoding time: 8 min 44 s
Total size: 597.19 MB
A table of the individual file sizes: ALAC_iTunes.htm

dBpowerAMP ALAC
Encoding time: 5 min 39 s
Total size: 611.12 MB
A table of the individual file sizes: ALAC_dBpA.htm

For comparison, the source Monkey's Audio files:
Total size: 570.31 MB
A table of the individual file sizes: MonkAud_norm.htm

I think I have found a problem track too. This small ogg vorbis file produces about 24% bigger file size when converted with the new encoder:
http://www.hydrogenaudio.org/forums/index....showtopic=42702.
It produces these sizes:
dBpowerAMP iTunes: 3.39 MB
dBpowerAMP ALAC: 4.19 MB
Monkey's Audio Normal: 3.11 MB

EDIT

I also tested the ALAC files with foo_bitcompare. All files were fine.

This post has been edited by Alex B: Mar 21 2006, 12:41


--------------------
http://listening-tests.freetzi.com
Go to the top of the page
+Quote Post
spoon
post Mar 21 2006, 13:55
Post #8


dBpowerAMP developer


Group: Developer (Donating)
Posts: 2725
Joined: 24-March 02
Member No.: 1615



My initial tests show that the dBpowerAMP ALAC sometimes beats iTunes ALAC compression, sometimes it does not. Looking at the code there seems to be a fall back encoding method (with less compression), this is used when the audio goes out of range of the predictor, that is my understanding, which could be wrong smile.gif

Of-couse the much improved encoding speed is a bonus...


--------------------
Spoon http://www.dbpoweramp.com
Go to the top of the page
+Quote Post
Alex B
post Mar 21 2006, 14:41
Post #9





Group: Members
Posts: 1303
Joined: 14-September 05
From: Helsinki, Finland
Member No.: 24472



The speed increase is remarkable when the dBpowerAMP GUI is used.

My test file set is not comprehensive enough for being reliable, though I selected the tracks carefully (various music types from various decades and no tracks that typically produce extremely low or high bitrates). A lot more tracks should be included and the set has no classical tracks (guruboolez did that part of testing back then).

The Vorbis file I uploaded contains mostly human voice and of course it is already encoded with a lossy encoder. Can these factors be difficult for the new encoder?

I tested the raw encoding speed too. I converted the test set first to wave format and compared the new dBpowerAMP ALAC conversion with a wave to ALAC conversion made directly inside iTunes. Naturally the speed difference is much smaller without the intermediate decoding and temp file steps:

dBpowerAMP new ALAC: 2 min 18 s
Standalone iTunes: 2 min 44 s


--------------------
http://listening-tests.freetzi.com
Go to the top of the page
+Quote Post
MystikTK
post Mar 24 2006, 07:28
Post #10





Group: Members
Posts: 8
Joined: 7-December 05
Member No.: 26261



As specified in this thread: http://www.hydrogenaudio.org/forums/index....88&#entry374588

Two questions:

1) If I use this encoder to change the said files in the above thread into this encoder, will they in both my iPod and Winamp?

and

2) Will I lose any quality since I would be converting from Vorbis > ALAC > this new version of ALAC?


Cheers.
Go to the top of the page
+Quote Post
spoon
post Mar 24 2006, 10:00
Post #11


dBpowerAMP developer


Group: Developer (Donating)
Posts: 2725
Joined: 24-March 02
Member No.: 1615



If your files are already in Apple Lossless converting again into Apple Lossless is useless, also converting from a lossy (Vorbis) format to a lossless one is questionable...


--------------------
Spoon http://www.dbpoweramp.com
Go to the top of the page
+Quote Post
pepoluan
post Mar 24 2006, 21:13
Post #12





Group: Members
Posts: 1455
Joined: 22-November 05
From: Jakarta
Member No.: 25929



QUOTE (spoon @ Mar 24 2006, 04:00 PM)
If your files are already in Apple Lossless converting again into Apple Lossless is useless, also converting from a lossy (Vorbis) format to a lossless one is questionable...
*
In MystikTK's case, I think converting from Vorbis to ALAC is not questionable... if iPod's the target device. Without the RockBox firmware, iPods can't run Vorbis. To reencode into AAC will degrade quality. Thus if the source is Vorbis and no lossless source, transcoding to ALAC will be preferable.

So that boils down to one question: If he/she transcode into this ALAC version, will it compress better or not?

This post has been edited by pepoluan: Mar 24 2006, 21:15


--------------------
Nobody is Perfect.
I am Nobody.

http://pandu.poluan.info
Go to the top of the page
+Quote Post
Alex B
post Mar 25 2006, 07:52
Post #13





Group: Members
Posts: 1303
Joined: 14-September 05
From: Helsinki, Finland
Member No.: 24472



I happened to have the same 25 tracks encoded with the new Vorbis aoTuV B5 test version at q -1. (~48 kbps)

I converted the Ogg Vorbis files with both ALAC versions. The resulting compression rates were significantly different.

iTUnes: 521.41 MB, 795 kbps

dBpowerAMP: 591.30 MB, 902 kbps


--------------------
http://listening-tests.freetzi.com
Go to the top of the page
+Quote Post
Alex B
post Mar 25 2006, 08:22
Post #14





Group: Members
Posts: 1303
Joined: 14-September 05
From: Helsinki, Finland
Member No.: 24472



Also, I noticed that the file tagging is somehow different with new encoder than with the iTunes module.

Mr QuestionMan 0.701 and Mp3tag 2.35k were unable to read the file information of iTunes encoded ALAC tracks. That didn't happen with the new ALAC encoder.

I resolved the problem by opening the iTunes encoded files with foobar 0.9 and making foobar to write a tag change to the files. After that the mentioned programs could read the file information.


--------------------
http://listening-tests.freetzi.com
Go to the top of the page
+Quote Post
MystikTK
post Mar 27 2006, 04:42
Post #15





Group: Members
Posts: 8
Joined: 7-December 05
Member No.: 26261



QUOTE (pepoluan @ Mar 24 2006, 05:43 PM)
QUOTE (spoon @ Mar 24 2006, 04:00 PM)
If your files are already in Apple Lossless converting again into Apple Lossless is useless, also converting from a lossy (Vorbis) format to a lossless one is questionable...
*
In MystikTK's case, I think converting from Vorbis to ALAC is not questionable... if iPod's the target device. Without the RockBox firmware, iPods can't run Vorbis. To reencode into AAC will degrade quality. Thus if the source is Vorbis and no lossless source, transcoding to ALAC will be preferable.

So that boils down to one question: If he/she transcode into this ALAC version, will it compress better or not?
*



My apologies, I keep being misunderstood in regards to my situation. To clarify, what I was planning on doing is this:

I had already converted the Vorbis files into Apple's version of ALAC and put these files on the iPod. However, I would like to take these files that are encoded in Apple's ALAC and encode them with this new version of ALAC so that I can play the files in Winamp etc.

My question is: Will I lose any quality by converting these Apple ALAC files into this new dbPowerAmp ALAC format? Keep in mind that these files were ORIGINALLY Vorbis.

As for compression, how much smaller will my file sizes be (rough estimate)?

This post has been edited by MystikTK: Mar 27 2006, 04:43
Go to the top of the page
+Quote Post
Alex B
post Mar 27 2006, 10:45
Post #16





Group: Members
Posts: 1303
Joined: 14-September 05
From: Helsinki, Finland
Member No.: 24472



QUOTE (MystikTK @ Mar 27 2006, 05:42 AM)
... My question is: Will I lose any quality by converting these Apple ALAC files into this new dbPowerAmp ALAC format? Keep in mind that these files were ORIGINALLY Vorbis.
*

The decoded output will be bit to bit identical since both formats are lossless. No additional quality loss can occur when a lossless file is re-encoded with a lossless codec unless the codec is buggy.

QUOTE
As for compression, how much smaller will my file sizes be (rough estimate)?

If you mean from Apple ALAC to dBpowerAMP ALAC, the files will probably be bigger (based on the results I have posted in this thread).

It would be useful if you could test that with your files. Could you convert a bunch of Apple ALAC files to dBpowerAMP ALAC and report?

Edit: rephrased

This post has been edited by Alex B: Mar 27 2006, 21:27


--------------------
http://listening-tests.freetzi.com
Go to the top of the page
+Quote Post
MystikTK
post Mar 29 2006, 07:11
Post #17





Group: Members
Posts: 8
Joined: 7-December 05
Member No.: 26261



QUOTE
The decoded output will be bit to bit identical since both formats are lossless. No additional quality loss can occur when a lossless file is re-encoded with a lossless codec unless the codec is buggy.


Thanks.

QUOTE
As for compression, how much smaller will my file sizes be (rough estimate)?
If you mean from Apple ALAC to dBpowerAMP ALAC, the files will probably be bigger (based on the results I have posted in this thread).

It would be useful if you could test that with your files. Could you convert a bunch of Apple ALAC files to dBpowerAMP ALAC and report?




Original Apple ALAC/dbPowerAmp ALAC

File 1 - 69.5 mb/70.5 mb
File 2 - 70.8 mb/71.9 mb
File 3 - 44.4 mb/44.9 mb

These converted files still do not play in Winamp. Any ideas as to what I could be doing wrong?

EDIT: Nevermind, I assume that it is impossible to get these to play in Winamp at the moment.

This post has been edited by MystikTK: Mar 29 2006, 07:20
Go to the top of the page
+Quote Post
guruboolez
post Apr 1 2006, 09:26
Post #18





Group: Members (Donating)
Posts: 3474
Joined: 7-November 01
From: Strasbourg (France)
Member No.: 420



I compared both encoders with my reference set of classical complete tracks (150).

CODE

dBamp iTunes

595,1 597,3 -2,3 kbps
887,2 865,3 21,9 kbps
923,0 918,7 4,3 kbps
907,5 908,6 -1,1 kbps
798,3 776,4 21,8 kbps
364,4 371,4 -7,0 kbps
656,5 651,3 5,2 kbps
619,9 624,4 -4,5 kbps
545,8 547,8 -2,0 kbps
581,7 580,8 1,0 kbps
624,3 626,9 -2,7 kbps
499,8 502,5 -2,7 kbps
794,5 733,3 61,2 kbps
638,5 633,2 5,4 kbps
618,7 600,3 18,4 kbps
687,5 684,7 2,8 kbps
536,6 523,5 13,1 kbps
682,3 682,8 -0,5 kbps
621,6 625,0 -3,4 kbps
599,4 603,9 -4,5 kbps
715,3 720,5 -5,2 kbps
669,2 669,7 -0,5 kbps
540,1 544,1 -4,1 kbps
523,7 525,1 -1,4 kbps
395,2 396,8 -1,6 kbps
519,5 523,4 -3,9 kbps
706,9 707,0 -0,1 kbps
542,8 543,6 -0,8 kbps
500,1 499,9 0,2 kbps
799,0 800,6 -1,6 kbps
618,6 612,7 5,9 kbps
583,1 577,8 5,3 kbps
661,7 661,4 0,3 kbps
668,6 660,9 7,7 kbps
679,7 676,6 3,1 kbps
896,6 897,8 -1,2 kbps
803,2 805,3 -2,1 kbps
803,7 779,6 24,2 kbps
486,3 490,4 -4,1 kbps
723,4 725,4 -2,0 kbps
896,2 901,0 -4,7 kbps
608,4 610,6 -2,1 kbps
587,2 590,2 -3,1 kbps
568,3 571,1 -2,8 kbps
575,3 562,5 12,8 kbps
766,9 768,2 -1,2 kbps
749,5 752,2 -2,6 kbps
735,2 732,4 2,8 kbps
487,7 487,9 -0,2 kbps
878,9 880,1 -1,2 kbps
646,8 648,6 -1,8 kbps
697,9 697,4 0,5 kbps
713,4 716,6 -3,3 kbps
661,6 662,5 -0,9 kbps
923,9 933,3 -9,4 kbps
724,5 726,5 -2,0 kbps
809,8 812,6 -2,8 kbps
777,6 775,0 2,6 kbps
787,1 781,0 6,1 kbps
707,0 709,5 -2,5 kbps
795,4 783,0 12,3 kbps
815,7 811,8 3,9 kbps
881,7 878,4 3,3 kbps
792,0 794,6 -2,7 kbps
821,5 823,8 -2,3 kbps
622,6 585,3 37,3 kbps
542,5 540,6 1,9 kbps
538,2 537,2 1,0 kbps
652,7 649,6 3,1 kbps
627,2 628,0 -0,7 kbps
566,0 563,8 2,2 kbps
782,0 782,4 -0,4 kbps
625,2 616,6 8,6 kbps
599,9 596,1 3,9 kbps
704,2 692,7 11,4 kbps
796,4 799,5 -3,2 kbps
894,3 895,1 -0,8 kbps
998,0 992,5 5,6 kbps
978,0 979,4 -1,3 kbps
785,5 780,4 5,1 kbps
861,4 862,8 -1,4 kbps
700,9 698,2 2,7 kbps
544,4 544,0 0,5 kbps
546,4 548,2 -1,9 kbps
680,1 680,4 -0,3 kbps
560,2 556,7 3,5 kbps
693,6 693,4 0,2 kbps
589,5 592,0 -2,5 kbps
592,7 590,0 2,7 kbps
540,9 533,6 7,3 kbps
417,4 421,2 -3,7 kbps
454,7 452,4 2,3 kbps
497,6 501,0 -3,4 kbps
459,7 462,7 -3,1 kbps
436,7 432,1 4,7 kbps
436,7 432,1 4,7 kbps
692,5 689,4 3,1 kbps
597,3 589,5 7,7 kbps
736,4 738,2 -1,8 kbps
842,5 824,3 18,2 kbps
514,0 517,6 -3,7 kbps
698,1 686,8 11,3 kbps
620,2 619,2 1,0 kbps
495,0 483,8 11,2 kbps
686,0 693,7 -7,7 kbps
585,6 582,6 3,0 kbps
556,5 549,0 7,5 kbps
600,0 598,5 1,5 kbps
779,6 777,4 2,2 kbps
426,9 425,0 1,9 kbps
506,5 503,1 3,4 kbps
568,7 566,9 1,7 kbps
589,4 590,6 -1,2 kbps
541,1 527,7 13,4 kbps
795,4 794,8 0,6 kbps
650,7 641,6 9,1 kbps
721,2 724,0 -2,9 kbps
532,1 521,5 10,6 kbps
405,4 399,6 5,8 kbps
516,9 514,4 2,5 kbps
672,7 676,0 -3,2 kbps
669,8 670,5 -0,7 kbps
285,1 279,7 5,4 kbps
859,7 859,3 0,3 kbps
634,0 636,1 -2,1 kbps
777,1 779,4 -2,3 kbps
678,3 677,1 1,2 kbps
469,4 470,8 -1,4 kbps
735,2 728,0 7,2 kbps
752,2 746,1 6,1 kbps
596,1 582,9 13,2 kbps
680,4 681,2 -0,8 kbps
765,5 743,7 21,8 kbps
688,6 690,0 -1,3 kbps
653,3 654,3 -0,9 kbps
626,0 623,3 2,7 kbps
493,5 499,3 -5,8 kbps
670,0 675,5 -5,4 kbps
608,9 605,9 3,0 kbps
678,2 680,6 -2,4 kbps
814,7 801,1 13,6 kbps
617,8 613,6 4,2 kbps
618,7 611,3 7,4 kbps
884,6 883,2 1,3 kbps
620,3 618,0 2,3 kbps
618,6 616,1 2,5 kbps
750,4 747,0 3,4 kbps
567,4 568,8 -1,4 kbps
757,7 751,9 5,8 kbps
657,0 659,5 -2,5 kbps
--------------------------------
658,09 655,45 2,6 kbps
dBamp iTunes


On average, both encoders are very close but iTunes has a slight avantage (-2.6 kbps). It may be interesting to note that dBPowerAmp improves the compression up to 9.4 kbps compared to iTunes encoder with some tracks but also lower the efficiency up to... 61 kbps (the next one is +37 kbps)!

EDIT: for reference, additional encoders' results are available here.

This post has been edited by guruboolez: Apr 1 2006, 09:51
Go to the top of the page
+Quote Post
spoon
post May 13 2006, 08:17
Post #19


dBpowerAMP developer


Group: Developer (Donating)
Posts: 2725
Joined: 24-March 02
Member No.: 1615



Updated today, includes a tagging fix (iTunes would not read the tags previously).


--------------------
Spoon http://www.dbpoweramp.com
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 25th April 2014 - 02:50