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: An MP3 encoder without a cutoff point (Read 4868 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

An MP3 encoder without a cutoff point

Hello (sorry if my English is bad), there was a topic named "Why does LAME cut off frequencies differently for CBR and VBR ?" but my topic is slightly different from it: i want an MP3 encoder without a cutoff point in relatively high bitrates. Reason:  This frequency band is outside of audible range in V1 MP3 but inside of audible range in V2 and V2.5 MP3's (11025Hz and 5512.5Hz) so an MP3 without a cutoff point can sound different. Beyond that, i can want to use MP3 for scientific purposes so i want to have cutoffless MP3's for all versions (11025Hz, 22050Hz, and 44100Hz). Here's what i mean (bandwidth=22050Hz, MPEG version=2.5, linear):

X

I don't mean the right one has higher quality, this photo is just for clarifying. Both MP3's are 32kbps mono. Right one is ripped from a flash game by EightGames (Dream Magical World Escape) and the left one is exported by LAME. I want cutoffless MP3's especially for these bitrates:

11025Hz 24kbps mono
11025Hz 32kbps mono
11025Hz 40kbps stereo
11025Hz 48kbps stereo
11025Hz 56kbps stereo
11025Hz 64kbps stereo
22050Hz 48kbps mono
22050Hz 56kbps mono
22050Hz 64kbps mono
22050Hz 80kbps stereo
22050Hz 96kbps stereo

By the way, i want to say that i don't like LAME's "sound color" but i like Adobe Flash CS6's "sound color". Again, i don't mean Flash gives higher quality.

Re: An MP3 encoder without a cutoff point

Reply #1 –
--lowpass -1

Re: An MP3 encoder without a cutoff point

Reply #2 –
--lowpass -1 is for which terminal?

Re: An MP3 encoder without a cutoff point

Reply #3 –
I prefer to use LAME 3.97 for low bitrate audio. I use -V 9, which is resampled to 24khz. The lowpass and samplerate is higher than new versions of LAME.
I use it mainly for audiobooks, but even music sound fine. But as it is VBR it will use much more bits for music, than for speech.

Re: An MP3 encoder without a cutoff point

Reply #4 –
But i need 11025Hz, 22050Hz, and 44100Hz at CBR...

Re: An MP3 encoder without a cutoff point

Reply #5 –
Addition: I also need 11025Hz 32kbps stereo but it can have cutoff.

Re: An MP3 encoder without a cutoff point

Reply #6 –
@shadowking How can I find a working and compiled Lame? The GXLame that @Kraeved gave me cannot handle --lowpass -1 .

Re: An MP3 encoder without a cutoff point

Reply #7 –
If GXlame has the -f switch, Try it as it uses less psycho-acoustics and has a more natural spectogram.

Re: An MP3 encoder without a cutoff point

Reply #8 –
@shadowking Unfortunately, this does not change anything. The output is not sharply lowpassed by a frequency but the frequencies those near to the nyquist limit are quieter.

Re: An MP3 encoder without a cutoff point

Reply #10 –
Quote from: Klymins on
@shadowking Unfortunately, this does not change anything. The output is not sharply lowpassed by a frequency but the frequencies those near to the nyquist limit are quieter.
Technically it would be stupid to risk image frequencies only for the sake of some more Hz in the transmission band. Use a normal sample rate instead!

Re: An MP3 encoder without a cutoff point

Reply #12 –
@Sunhillow Normal sample rate?

Re: An MP3 encoder without a cutoff point

Reply #13 –
@saratoga This gives even worse results. The output is sharply lowpassed.

Re: An MP3 encoder without a cutoff point

Reply #14 –
@l4cache How can I download a compiled Shine?

Re: An MP3 encoder without a cutoff point

Reply #15 –
Quote from: Klymins on
@l4cache How can I download a compiled Shine?
Maybe from msys:
https://packages.msys2.org/package/mingw-w64-x86_64-shine?repo=mingw64
Or if you trust me.

Msys one needs dll with exe, attached file is single executable (well, you still need some Windows runtime dlls)

Sounds terrible at low bitrate though.

Re: An MP3 encoder without a cutoff point

Reply #16 –
@l4cache How can I set the sample rate? And, why it sounds terrible at mono even if the bitrate is high?

Re: An MP3 encoder without a cutoff point

Reply #17 –
I don't know. I never actually used it before.
I just use resampled temporary .wav file as input.

Re: An MP3 encoder without a cutoff point

Reply #18 –
Thanks.

Re: An MP3 encoder without a cutoff point

Reply #19 –
The music that I gave it's spectrogram was probably encoded by Lame according to mp3guessenc, but I could not get a good cutoffless MP3 with a new version of Lame. Which Lame version is good for me?

Re: An MP3 encoder without a cutoff point

Reply #20 –
Okay, it turned out that this was caused by the resampling process. The problem was solved when I use an external resampler.

Re: An MP3 encoder without a cutoff point

Reply #21 –
Quote from: Klymins on
Okay, it turned out that this was caused by the resampling process. The problem was solved when I use an external resampler.

... but apparently it didn't. I wrote this after trying just 1 bitrate-samplerate combination and now I found that I still can't get a fully cutoffless MP3 for some bitrate-samplerate combinations (for example 11025Hz 24kbps mono). The file was actually cutoffless but the upper frequencies were encoded without too much attention.

How can I get high-quality and completely cutoffless MP3's? Does anyone have an idea about that?

Re: An MP3 encoder without a cutoff point

Reply #22 –
It is not actually possible for frequencies to extend all the way to the cut off, but depending on the sampling rate, resampler and how you analyze the data it may look more or less like it does in a spectrogram, or you may see distortion and confuse that for audio frequencies.  How close does it look?

Re: An MP3 encoder without a cutoff point

Reply #23 –
Quote from: saratoga on
It is not actually possible for frequencies to extend all the way to the cut off, but depending on the sampling rate, resampler and how you analyze the data it may look more or less like it does in a spectrogram, or you may see distortion and confuse that for audio frequencies.  How close does it look?

Sorry, I couldn't quite understand that, can you explain a little more?

Re: An MP3 encoder without a cutoff point

Reply #24 –
For me lame seems to show a different behavior for 48000 and 44100 than for other sampling frequencies. I generated a stereo sweep from 0.9 bandwidth to max at sampling frequencies: 48000, 44100, 32000, 22050, 16000 and 11025. Then I encoded them with "lame -b 32 --lowpass -1". Only 48000 and 44100 preserved the signal at the top:
X