Welcome Guest ( Log In | Register )

how to get the lower bitrates?
post Jan 12 2002, 19:46
Post #1

Group: Members
Posts: 159
Joined: 12-January 02
Member No.: 992

So, I would like to have mp3 files close to 192 kbps. I listen mostly metal (various types - heavy, power, black, ...) and --alt-preset standard seems to create mostly bitrates of 200-250 (tested with several albums). I consider this to be too high, how should I create the files to get lower rates? Should I use ABR with --alt-preset [rate] or maybe some VBR command line (for example, edited -r3mix), using -V[n] to adjust the bitrates? Any suggestions?
Go to the top of the page
+Quote Post
Start new topic
post Mar 12 2002, 03:30
Post #2

Group: Developer
Posts: 2797
Joined: 22-September 01
Member No.: 6

Originally posted by ErikS
No. Sorry for beeing a bonehead, but i don't understand. Why isn't it possible to lower the other scalefactors (<21), when you add to the global scalefactor, if the only purpose is to increase resolution in sfb21? Ok, i admit that i haven't read a single line of code. I'm sure there is a reason why it's not done, but it would be interesting for me to know.
*Cough*, my try to explain the problem with VBR coding and high freq content:
There is a certain masking threshold created by psycoacoustics across the full frequency band (21 scalefactor bands with long blocks), and the encoder (quantization-noiseshaping loop) must try to obey this threshold so, that enough bits (resolution) is used for every SFB so that there wouldn't be audible quantization noise. Quantization-noiseshaping loop does this by adjusting scalefactors which are used to reduce the stepsizes of every SFB - except SFB21 because there's no scalefactor for it. Smaller stepsize of certain SFB means higher resolution in that SFB (more bits used).

Now, suddenly there is lots of loud high frequency content and the q-ns loop notices for SFB21 that damn - I need higher resolution for this scalefactor band number 21, what to do? q-noiseshaping loop can't adjust the scalefactor to decrease the q-stepsize (increase the resolution) of SFB21 in order to comply with masking threshold - simply because there's no scalefactor for SFB21.
So, how can q-ns loop adjust the q-stepsize of SFB21 to comply with SFB21 masking threshold which demands higher resolution? There's still one way:
Global gain value defines the largest q-stepsize value used. So, obviously a proper global gain value must be used (it's now used for SFB21) so that it complies with resolution demanded by SFB21 masking (of course this is not the same global gain which is used in loudness/amplitude tweaking by mp3gain).
Too bad is that the smaller the stepsize of SFB, the more resolution (bits) will be used in that SFB. So, in order to comply with resolution demanded by SFB21 masking threshold, global gain (largest stepsize value which in this case will be used for SFB21) must be quite low. That means that there can't be larger stepsize values (lower resolution) for all the other under SFB21 bands, only smaller stepsize values. That also means that higher resolution than maybe necessary or wanted must be used for scalefactor bands 0-20 (under 16kHz).

However, there's one practical way to decrease this effect and still maintain somewhat decent high frequency production. You use --ns-sfb21 switch to raise the masking threshold for SFB21, so that it wouldn't need so high resolution. Of course high frequency production/accuracy is not as good anymore then.

Juha Laaksonheimo
Go to the top of the page
+Quote Post

Posts in this topic
- amppa   how to get the lower bitrates?   Jan 12 2002, 19:46
- - tangent   I would recommend only using --lowpass (try 17.5) ...   Jan 12 2002, 20:16
- - ami   hi, hmm since ive read, that dibrom listens and en...   Jan 12 2002, 20:25
- - amppa   So, if I got it right, lame --alt-preset [rate] sh...   Jan 12 2002, 21:38
- - NeoRenegade   You got it dead on   Jan 13 2002, 23:46
- - TheBashar   I don't recall ever reading Dibrom say that with h...   Jan 14 2002, 00:56
- - Weiman   How about --alt-preset standard -V 3 -B 256? It gi...   Mar 7 2002, 21:29
- - Dibrom   My stance on this: Lowering the lowpass is by far...   Mar 7 2002, 21:57
- - mp3fan   QUOTE like employing an adaptive lowpass, with a l...   Mar 8 2002, 03:59
- - kxy   Newbie questions... when I use -alt-preset stand...   Mar 8 2002, 08:22
- - Sachankara   I still donīt see why you donīt use --alt-preset 1...   Mar 8 2002, 22:55
- - Dibrom   QUOTE Originally posted by mp3fan Currently [b]--...   Mar 8 2002, 23:42
- - kxy   Thx Dibrom. So despite the sfb21 warning, if some...   Mar 9 2002, 15:32
- - Dibrom   You don't need to use a lowpass when you use -Y. ...   Mar 9 2002, 19:10
- - kxy   Oh yea, of course I don't need to use -y and --low...   Mar 9 2002, 20:50
- - Dibrom   QUOTE Originally posted by kxy To recap, the -Y o...   Mar 9 2002, 21:03
- - mp3fan   Dibrom, QUOTE By using -Y, and disabling noise sh...   Mar 10 2002, 03:10
- - jkeating   You are forgetting about stereo settings! Right no...   Mar 12 2002, 00:04
- - Dibrom   QUOTE Originally posted by mp3fan So by using a -...   Mar 12 2002, 00:21
- - Dibrom   QUOTE Originally posted by jkeating You are forge...   Mar 12 2002, 00:24
- - ErikS   QUOTE Originally posted by Dibrom The reason fo...   Mar 12 2002, 02:28
- - JohnV   QUOTE Originally posted by ErikS No. Sorry for be...   Mar 12 2002, 03:30
- - Delirium   JohnV and Dibrom -- thanks for the explanations, I...   Mar 12 2002, 09:40
- - mithrandir   Considering that MP3s primary function is to reduc...   Mar 12 2002, 17:35
- - mp3fan   QUOTE Also, when I use an --athlower number up to ...   Mar 16 2002, 20:06
- - Dalkus   Reading through the FAQs and other threads I can...   May 4 2003, 16:30
- - SometimesWarrior   Using the -Y switch with --ap extreme carries the ...   May 5 2003, 06:36
- - Dalkus   Thanx for the reply. I'm just thinking that if...   May 5 2003, 13:30

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: 27th November 2015 - 18:12