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: --lowpass 16 and lame_dm_rev6 (Read 2950 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

--lowpass 16 and lame_dm_rev6

In Lame 3.90alpha7 encoding the same songs with:

(1) --dm-preset standard -Z
and
(2) --dm-preset standard --lowpass 16

resulted in a dramatic filesize decrease with --lowpass 16
e.g., 50-60 kbps...

But recently I've noticed that if you add --lowpass 16 to --alt-preset normal in lame_dm_rev6 the resluting change in filesize
is a much smaller decrease
e.g., 10-20 kbps...

I'm just curious as to why this happens, what tweaks have accomplished this?

Is it that now adding the --lowpass 16 does not decrease the resolution at which the frequencies below 16hertz are encoded at?  I remember Dibrom mentioned something about this in an earlier post:

Quote
Since we don't have this for sfb21 (long blocks) or sfb12 (short blocks) we have to rely on noise shaping in the last band to determine what to encode.  Then, the only way to encode this is to change the quantization resolution on a global scale.  Thus the frequencies over 16khz aren't what is causing the massive bitrate increases, it is all the other frequencies below 16khz which are now encoded with increased resolution (more bits) because we needed to encode above 16khz.


PS HAPPY THANKSGIVING!!

--lowpass 16 and lame_dm_rev6

Reply #1
Quote
Originally posted by RD
But recently I've noticed that if you add --lowpass 16 to --alt-preset normal in lame_dm_rev6 the resluting change in filesize 
is a much smaller decrease

e.g., 10-20 kbps...

I'm just curious as to why this happens, what tweaks have accomplished this?


The main reason for this is that --alt-preset normal does not encode as much "trivial" content above 16khz.  This is basically in line with the general philosophy behind the design of this preset, encoding less where it is inaudible, and more where it is audible.

In essence, what doesn't get encoded over 16khz should mostly be just noise that is so heavily masked by lower frequency content that it is normally inaudible or mostly inaudible.  This effect is emphasized by the fact that as volume decreases, more high frequency content is usually encoded, and also that all distinct transients (and a significant portion of their decays) should be encoded basically all the way up to the lowpass limit even if other surrounding content is not encoded as aggressively.

The effects of this can be seen visually in these pictures:

Normal:

http://ff123.net/export/ringing_normal.gif

r3mix:

http://ff123.net/export/ringing_r3mix.gif

As you can see, "normal" emphasizes the attacks and decays much more (you can see more encoded directly before and after the attacks), while r3mix is the opposite and pays more attention to the lower volume, but not nearly as audible regions.

Also, note that these shots are of the very first revision, and rev6 has been improved quite a bit since then, but for the most part the general behavior is similar.