Re: lame main branch

Very nice indeed Thomas! Thank you!

> PS: How serious are you about cutting decoder delay from mp1 and mp2? I
> left that alone in mpg123 as I don't have any encoder padding values

> anyway. I only added gapless decoding for mp3, where it has a chance to

> work.

I would guesstimate pretty much noone transcodes from MP2, and even less
people use MP1 at all. Besides, I suppose there is no equivalent to the LAME
header in MP1 or MP2, right? And even if there was, any encoder writing it?

Matter of fact, is there any MP1 encoder besides the dist10 sources and
(gasp!) Uzura?

I suppose you can just ignore gapless decoding for MP1/2, and worry about it
in the very unlikely event that someone complains.
 
-- 

Roberto José de Amorim
MSc in Computer Science
Columbia University - New York





________________________________
De: Thomas Orgis <[email protected]>
Para: 
Cc: "[email protected]" <[email protected]>
Enviadas: Quinta-feira, 24 de Agosto de 2017 19:02
Assunto: Re: [Lame-dev] lame main branch



Am Tue, 22 Aug 2017 19:30:33 +0200

schrieb Thomas Orgis <[email protected]>: 


> Well, my plan is this: I'll see what I can come up with until

> (including) Thursday. 


I got a sketchy build of lame from CVS with libmpg123 (verboseness

active with the Note: lines):


$ frontend/lame /projects/mpg123/svn/test/regression/drum.mp3 drum-anew.mp3

Note: Disabling all formats.

Note: Want to enable format 0/3 for encodings 0xd0.

Note: ID3v2.3 rev 0 tag of 1558 bytes

Note: ID3v2 COMM frame of size 60

Note: Storing comment from Latin 1 encoding

Note: ID3 comm/uslt desc of length 11.

Note: ID3 comm/uslt text of length 47.

Note: ID3v2 TIT2 frame of size 24

Note: Storing text from Latin 1 encoding

Note: ID3v2 TIT2 text frame: Test track (some drums)

Note: ID3v2 TPE1 frame of size 13

Note: Storing text from Latin 1 encoding

Note: ID3v2 TPE1 text frame: Thomas Orgis

Note: ID3v2 TYER frame of size 5

Note: Storing text from Latin 1 encoding

Note: ID3v2 TYER text frame: 2010

Note: ID3v2 TALB frame of size 22

Note: Storing text from Latin 1 encoding

Note: ID3v2 TALB text frame: The mpg123 Repository

Note: ID3v2 TRCK frame of size 4

Note: Storing text from Latin 1 encoding

Note: ID3v2 TRCK text frame: 1/1

Note: ID3v2 TCON frame of size 5

Note: Storing text from Latin 1 encoding

Note: ID3v2 TCON text frame: (33)

Note: Xing/Lame/Info header detected

Note: Xing: 1247 frames

Note: Xing: 860640 bytes

Note: Xing: quality = 77

Note: Info: Encoder: LAME3.98r

Note: Info: rev 11

Note: Info: vbr mode 3

Note: Info: peak = 0.000000 (I won't use this)

Note: Info: Radio Gain = -9.8dB

Note: Info: Audiophile Gain = 0.0dB

Note: Encoder delay = 576; padding = 1344

LAME 3.100 (beta 0, Aug 24 2017) 64bits (http://lame.sf.net)

Using polyphase lowpass filter, transition band: 16452 Hz - 17032 Hz

Encoding /data/projekte/mpg123/svn/test/regression/drum.mp3 to drum-anew.mp3

Encoding as 48 kHz j-stereo MPEG-1 Layer III (12x) 128 kbps qval=3


Note: Cut frame 0 buffer on beginning of stream by 1105 samples, fill now 188 bytes.

    Frame          |  CPU time/estim | REAL time/estim | play/CPU |    ETA 

  1200/1247   (96%)|    0:00/    0:00|    0:00/    0:00|   38.687x|    0:00 

-----------------------------------------------------------------------------------------------00:01---

   kbps        LR    MS  %     long switch short %

  1247/1247  (100%)|    0:00/    0:00|    0:00/    0:00|   38.450x|    0:00 

-------------------------------------------------------------------------------------------------------

   kbps        LR    MS  %     long switch short %

  128.0       40.7  59.3        95.2   2.9   1.9

Writing LAME Tag...done

ReplayGain: -9.4dB


This is currently an experimental mpg123 build because I presume I need

to add those hooks for lame to extract the raw ID3v2 data (any further

parsing of that?). It wants to put that into the re-encoded file,

right? Sorry for being a bit incoherent … it's my state of mind trying

to navigate the lame sources.


I guess I could get that into shape. But it dawns on me that this is

only a small part of the work. So far I got the frontend using

libmpg123, but libmp3lame is a whole different thing, isn't it? It has

another interface for the mpglib? How many entry points are there? A

list of files that need hooks replaced would be nice. Also which

configure variables I need to set to slip libmpg123 linking into

libmp3lame … of course I could find out this stuff myself with frantic

searching, but since the project is so alive right now, the answer

might be easily given.


Also I would like to switch on floating point output from libmpg123. Is

there any place in lame that actually wants clipped 16 bit samples?



Alrighty then,


Thomas


PS: How serious are you about cutting decoder delay from mp1 and mp2? I

left that alone in mpg123 as I don't have any encoder padding values

anyway. I only added gapless decoding for mp3, where it has a chance to

work.

------------------------------------------------------------------------------

Check out the vibrant tech community on one of the world's most

engaging tech sites, Slashdot.org! http://sdm.link/slashdot

_______________________________________________

Lame-dev mailing list

[email protected]

https://lists.sourceforge.net/lists/listinfo/lame-dev

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Lame-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/lame-dev