Help - Search - Members - Calendar
Full Version: Question about switches -q and -k
Hydrogenaudio Forums > Lossy Audio Compression > MP3 > MP3 - Tech
Sylph
I know these have been discussed many times before and that the recommended settings say that, generally, less switches is better. That, for example, doesn't have to include cases like when you know you can't hear frequencies above a certain threshold etc.

Anyway, since LAME 3.98.2 ignores the -k switch, can I keep all frequencies by specifying --lowpass 20.5 -- highpass 0? blink.gif It is a purely theoretical question.

And since that same version of LAME apparently makes no distinction between -q from 0 to 3, why are those still present as possible values? I.e. why isn't it all one setting?
pdq
Keeping them as separate values that give the same result provides backward compatibility.
[JAZ]
LAME by default does not use a highpass filter, so the --highpass switch is not needed in this case.

Also, -q 0, 1, 2, 3 are the same, just with VBR (and vbr-new). It does have an effect with CBR (because the difference in the algorithms).

Sylph
I've just tried this and, obviously, --lowpass 20.5 --highpass 0 returns an error.

No way to include all the frequencies and go around -k (which doesn't work)?
pdq
Since there is by default NO highpass, --highpass 0 was redundant and probably caused the error.
WonderSlug
According to LAME 3.98.2 command line, "highpass" must be at least 0.001

So, the "--highpass 0" specification did cause the error.

Use "--highpass 0.001" if you truly want it. That specifies one hertz.




EDIT: changed "must be greater than" to "must be at least"
pdq
As I recall, the real lowest highpass is several hundred Hz. Anything less than that is still no highpass.
Sylph
Thanks.

But when I used the --lowpass 20.5, LAME said it used polyphase lowpass and then listed a range from which it filtered out the frequencies.

So perhaps, even that didn't serve much purpose.
[JAZ]
QUOTE (Sylph @ Jun 1 2009, 20:16) *
But when I used the --lowpass 20.5, LAME said it used polyphase lowpass and then listed a range from which it filtered out the frequencies.


Please, if you don't know what you are doing or what you are talking about, refrain from doing it.


A filter never filters at an exact Hertz. It has a range of decay, and that's what LAME is telling you when it is used.
I assume you're not as stupid to think that it uses a "passband between 18Khz and 19Khz", do you?

Gabriel
QUOTE (Sylph @ Jun 1 2009, 19:16) *
But when I used the --lowpass 20.5, LAME said it used polyphase lowpass and then listed a range from which it filtered out the frequencies.

No. It actually told you about the transition band it's using. (and if I remember well, "transition band" is explicitely mentionned)

You might find the transition band to be a bit wide. This is because Lame is directly using the bands from the mp3's polyphase filter to do the filtering. (and anyhow, brickwall filters are not really a good idea for a lowpass)
Sylph
QUOTE ([JAZ] @ Jun 2 2009, 00:17) *
Please, if you don't know what you are doing or what you are talking about, refrain from doing it.


A filter never filters at an exact Hertz. It has a range of decay, and that's what LAME is telling you when it is used.
I assume you're not as stupid to think that it uses a "passband between 18Khz and 19Khz", do you?


Wait. How did this conversation plummet into offensive statements? Now I'm stupid? blink.gif

Anyway... I know that presets are the golden standard. I've said it many times. I meant to use these MP3s only in this experiment.

If you knew my command options wouldn't work, why didn't you say so? The basic question was: how do you go around the fact that -k isn't an option anymore? I obviously didn't find the minimum highpass and maximum lowpass and that's the whole point - so I asked.

So yes, I don't know what I'm doing, which is why I'm asking you all here.
pdq
Couldn't you just use an earlier version of Lame in which the -k switch works?
2Bdecided
QUOTE (pdq @ Jun 2 2009, 15:52) *
Couldn't you just use an earlier version of Lame in which the -k switch works?
Like the equivalent alpha version?
lvqcl
QUOTE (Sylph)
The basic question was: how do you go around the fact that -k isn't an option anymore? I obviously didn't find the minimum highpass and maximum lowpass and that's the whole point - so I asked.


It depends on what do you want to do.

If you want to "keep all frequencies by specifying --lowpass X -- highpass Y" - then no, you cannot keep all frequencies. LAME 3.98.2 cuts all frequencies above ~20.5 kHz, and you cannot cancel this.

QUOTE (Sylph)
But when I used the --lowpass 20.5, LAME said it used polyphase lowpass

Surely.
robert
I didn't tell you that, but try "--lowpass -1"
pdq
QUOTE (robert @ Jun 2 2009, 11:15) *
I didn't tell you that, but try "--lowpass -1"

Are you sure you want to encourage this silliness? rolleyes.gif
lvqcl
QUOTE
try "--lowpass -1"

Wow.
[JAZ]
QUOTE (Sylph @ Jun 2 2009, 15:09) *
Wait. How did this conversation plummet into offensive statements? Now I'm stupid? blink.gif


I wasn't polite, but i didn't really tell you that.
The point was that you weren't clear on your questions, and said some things that didn't make sense.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2009 Invision Power Services, Inc.