I've read some users prefer to use the -Y switch with -aps in LAME. What does this do? Why might this be preferable to the regular -aps only switch?

It's not preferable.
I wrote this yesterday...
http://www.hydrogenaudio.org/forums/index....4552#entry46392

Thanks. I tried searching for -Y but the character limit of the search tool is too high.

I think if someone were to properly tune -Y along with some ABX listening tests, we could lower the bitrate substantially and get just as good a quality as the standard preset.

mp3

Hi JohnV,

from what i read in that post and what i know about -Y, it *is* preferable over just aps when you want as much quality as possible ...

Most people cannot hear beyond 16kHz and -Y just says, that those frequences will not be encoded when that would need too much bits. It *will* be encoded, when there is there is still free room ...

So, did in understand something wrong?!?!!?

Confused ... again ... :-(

It's not when there's "free room" that the higher frequencie will be included. Since --aps is vbr it has most of the time (when it doesn't hit the 320kbps limit) free room. Only some of the very high energy high frequencies get encoded with -Y.

--aps -Y is not quality wise preferable over only --aps. When you add -Y, the over 16khz content goes pretty much to minimum. -Y won't increase the quality in anyway (there's no extra bits allocated to lower frequencies), so it is only good if you need to save in file size.

@JohnV:

Would adding lowpass allocate bits to lower frequencies or will it just cut abruptly at the given value? I'm talking about the VBR presets now, I know it does it (allocate) with ABR and CBR. I just haven't been able to find it in the FAQ.

It is to old guys like me who can't hear anything over 16K.

... or anyone playing MP3's through WinAmp with the Equaliser on... c'mon - own up!

If the statement above is true, I think this is THE WORST scenario. I mean: If most of the people don't hear beyond 16kHz, why encode those frequencies at all, even when there is enough free room? And for the people, who *CAN* hear beyond 16 kHz it will be awful to have sometimes those frequencies, and not to have them at other times.
If you cannot hear above 16 kHz - you are just wasting space (bitrate, bandwidth) encoding them.
If you can - you sometimes have above 16 kHz, sometimes not, which leads to awful artifacts, I think.

The rank (according to my opinion) is:

1. To have all frequencies above 16 kHz (Best)
2. Not to have frequenices above 16 kHz (lowpass filter) (could be acceptable in certain circumstances)
3. In one song constantly changing availability of frequencies above 16 kHz (Not acceptable for me)

I would prefer a switch, where you can select a threshold for encoding frequencies above 16 kHz (or any selected range) in db.
For example - if there is presence of frequencies above 16 kHz whose level is (let's say) above -16 db they should be encoded. If their level is lower, they should not, anyway their level is too low to be heard...
This is only my point ov view.



P.S. Excuse me for my really bad english.

If you use lowpass with VBR it does not (at least in principle) allocate "extra" bits to lower frequencies.

Then the only reason to use lowpass is to just *save* bits for a small filesize ...

And to remove some artifacts or echos or whatever from frequences that were above 20.5 kHz and mirrored down below 20.5kHz, if i understood correctly. Such things can then even be heared.

But we never realy know - beside looking at every wav-File - if there was content above 20.5kHz and therefor we don´t know, if we need to cut some parts below 20.5kHz because of that.
I have the feeling that --alt-presets can work with that

So, it seems to be needed to go down as much as possible.
But if i read correctly, there were much debate about what frequences people can hear and if there is some additional effect, that you can only feel instead of hear ...

On a CD should be nothing higher than 20.5kHz, right?
But there could be stuff up to 22.5 kHz.
Do get away all possible artifacts or echos (don´t know that things), you would need to go as much down as 18.5 kHz ...
At least, that should be save, yes?
Being a bit more aggressive to that and we are at the standard setting of aps (19 kHz) ...
With my extremist extreme setting i wanted to get away all not unusable frequences, but i did not know nothing about that mirroring from above 20.5 kHz ... again more to read

Maybe i need to change my lowpass setting.

Just trying to get things together

This is only remotely about -Y, but about high frequences ...

I encoder applaus.wav with ape, ape -Z, ape -Z -V0 and ape -Z -V0 -lowpass 20.5 / 19.5 / 18.5 / 17.5 whereas ape by itself already has a builtin 19.5 version ...
Then i put those MP3´s into Nero Burning ROMs Audio Editor and looked at the spectral analysis:

All of that files still have content *way* beyond 20.5 kHz! Even in the range above -30 dB ...
That was not what i wanted to see ...
I am surprised and buffled. Those filter do not work as i though.
What´s up?!?!!?!

-Y is used because it can produce a significant reduction in bitrate without reducing subjective quality (in a variety of material, but not all). The reason why -Y can give much bang-for-the-buck is that it acts like a workaround for a deficiency in the MP3 spec (namely, no scalefactor for sfb21).

I normally use lame3.90.2 aps -Y. I thought I'd give 3.93.1 a try since it's got an extra switch "medium" added. And yes, the file size dropped a bit and I thought I'd add the -Y switch to see if I could reduce the size even more. But no, the -Y switch doesn't work with the "medium" preset! (at least I coudn't get it to work) Why is that? My conclusion is that the -Y switch is already there. Right? Will the new "medium" preset in 3.94 work in similar ways?

Hi,
about hearing beyond 16KHz. with a sinus beyond 19-20KHz, I for one am not sure annymore if i hear anything or not..
but probably i have not so good equipment to test this.
What i DO know from experience is that, maybe you cannot hear past
a certain frequency, but i always felt i could "feel it" somehow.
Somehow there is something there alright, i only know it doesnt sound
pleasant!
ie. I dunno how (for instance) a 20KHz sinus should sound like (be percieved
by the human ear)
Should it noticably sound like a disguisting eeeerry hard "peeeep" sound like a 16KHz sinus?
Or should a sinus-peeep indeed sound 'sharper but less loud' as the
frequency in KHz. gets higher?
I cannot hear a clear hard peep at 20KHz, but i can certainly feel something weird! -if you know what i'm saying..

anyone care to explain?,
maybe i'm just talking nonesense.

morelli

That is correct. Medium incorporates -Y.