I was wondering if anyone out there who has ever worked with holographic subliminal messages in audio, or hemi-sync (where slightly different frequency tones are played in each ear through headphones) have experimented with, or at least read of a study done to see if lossy audio encoding has a destructive effect on these technologies.
I'm concerned that subliminal audio and hemi-sync might partly be based on sounds that the human brain doesn't readily perceive, and thus the encoding process might "throw away" sound that is required for subliminal audio to be effective.
I'm especially concerned about whether joint stereo will wreck the effect of these techniques, since they are based on stereo separation of sound. Would it? Should I force encoding to actual stereo, or is Joint Stereo okay for this stuff?
I try to use lossless for this stuff wherever possible, but obviously that's not always practical.
Any informed responses much appreciated!
MP3 compression works by removing high-frequency components from the sound. Binaural beats are based on two slightly different tones that get mixed inside your brain. The tones can be, for example, 400 Hz and 410 Hz. When these are mixed together, the result is a 10 Hz binaural beat. Thus, there are no high-frequency components present, and MP3 compression has no degrading effect on the output. As long as the compression is applied to both left and right channels independently (i.e. compression does not mix them together), MP3 compression can be used on binaural beats without problem.
You can check the results of MP3 compression very easily: if you hear a continuous tone when listening to left and right channel separately, and a pulsating sound (binaural beat) when listening to them with headphones, the output is correct.
http://www.bwgen.com/faq.htm#MP3COMPRESSION (http://www.bwgen.com/faq.htm#MP3COMPRESSION)
I read the article as saying that mp3's should be encoded in stereo mode. It is over a year old though, and might not currently be valid. Others with more knowledge of mp3's encoding processes will be able to chime in on this.
With Lame 3.96.1, on a wav file made of a -6 dB 400 Hz sine on the left, and a -6 dB 410 Hz sine on the right, at -V5, -V4, and -V3, joint stereo uses 100% mid/side frames.
At -V5 and -V4, the addition of -m s (100% of L/R frames) does not change quality for me. The bitrate is 32 kbps.
But at -V3, the bitrate jumps at 41.7 kbps (nearly the same with -m s).
There, joint stereo (100 % mid/side encoding) provides better quality for me.
ABX L/R versus M/S : 8/8
I really appreciate the replies, guys, but I'm not sure that I completely understand what Pio2001 is saying... I'm interpreting that as meaning that Joint Stereo does not affect the results in this case... especially at high levels of quality (i.e. preset standard, or preset extreme)
Is there some chance that the sound from each channel could occasionally get "mixed" together in joint stereo?
I think we already have a tutorial about "Joint Stereo"
Joint Stereo, in fact, improves quality in most cases as it allows for:
- Control of noise imaging
- Reduction of the bit rate demands when bands (channels in MP3 case) are correlated
So - Joint Stereo allows MP3 codec to allocate less noise per frame, than it would be able to do with simple Stereo. Also it eliminates some artifacts that would arise by independently coding channels that are supposed to be played in pair.
Of course, this holds if and only if a particular MP3 codec implementation has good MS/LR switching algorithm (which decides which frame should be coded as LR and which as MS) - and LAME is certainly such codec.
Is there some chance that the sound from each channel could occasionally get "mixed" together in joint stereo?
Chances are bigger that the LR Stereo Mode (so not Joint Stereo) would generate infamous noise unmasking due to lost correlation of the masking noise and hence much lower quality than Joint Stereo
Why cant LAME just rename joint-stereo to "smart stereo" and the current "stereo" to "dumb stereo" so that these kinds of misunderstandings finally stop? :)
- Lyx
It has been said on this board that lame had problems encoding dolby surround files, so perhaps that these strange sounds you refer to could be affected, I guess.
About joint stereo renaming, I think it was already discussed and it's a standard naming problem. Joint-stereo is in the mp3 standard, and it is not called smart stereo or whatever. Lame developpers could probably modify the displayed name but it would not affect any decoder or tag reader; in the end it would be useless.
I really appreciate the replies, guys, but I'm not sure that I completely understand what Pio2001 is saying... I'm interpreting that as meaning that Joint Stereo does not affect the results in this case... especially at high levels of quality (i.e. preset standard, or preset extreme)[a href="index.php?act=findpost&pid=349750"][{POST_SNAPBACK}][/a]
In -V3 (below preset standard), joint stereo is better than L/R stereo. ABX 8/8
In -V2 (preset standard), L/R stereo is better than joint stereo for this sample (-6 dB sines). ABX 16/16. (Joint stereo is 99.74 % M/S).
In -V0 (preset extreme), L/R stereo is better than joint stereo on this sample. ABX 16/16. Strangely enough, joint stereo is then 98.96 % L/R
In -b 320 (preset insane), the bitrate is 320 kbps and the sample sounds ok in Joint stereo (99.2 % L/R) as well as in L/R stereo.
In short, this is a killer sample. Artifacts are obvious with preset standard, and audible with preset extreme. The bitrate is 128 kbps in both cases.
Sometimes joint stereo is better, sometimes L/R stereo is better.
Use preset insane (alias -b 320) in order to encode it to MP3.
Sample :
-6 dB 400 Hz sine on the left channel
-6 dB 410 Hz sine on the right channel
Duration : 10 seconds.
Thanks so much Pio2001, for clarifying that a bit.
Obviously, with material such as this, I wouldn't normally want to use a low level of quality, since it could affect the performance of the technology, so I'd be using the high qualities... standard and up. In this case, it actually makes sense to force stereo, so that does answer my question the way I thought it might.
My interpretation of the short answer is that:
To encode binaural or holographic subliminal recordings:
-Use Insane (-b 320) in either stereo or joint stereo (doesn't matter)
or if size may be an issue, and you need slightly smaller encodes
-Force L/R Stereo and use a quality of -V2 (standard) or better.
I presume that at these settings, there should be virtually no compromise in the efficacy of the technology. Perhaps with other settings, the encoding is less than transparent, and may affect the effect.
Thanks so much guys!
No, this is not really the case. This kind of sound is a big problem for the presets, they believe that it is easy to encode, but its not. Thus they don't allocate enough bits.
What you need is CBR presets. Preset insane is a CBR one, that's why it works well. Preset standard and extreme are VBR, and don't work well. In order to get smaller files than with preset insane, use -b 256, or -b 192, or -b 160.
Preset standard and extreme end up at 128 kbps, so they will be roughly the same as -b 128.
For joint stereo vs L/R stereo, I can't tell. I saw that depending on the preset, one or the other could be better. There is one artifact in preset standard and extreme with joint stereo that goes away in L/R stereo. It must be the Mid/Side frame that causes this artifact (there is only one M/S frame for one of the presets, and three ones for the other). But I don't know why and how.
Maybe if the frequences are different, the encoding will behave differently.
What exactly is "holographic subliminal messages in audio"? :)
What exactly is "holographic subliminal messages in audio"? :)
[a href="index.php?act=findpost&pid=349962"][{POST_SNAPBACK}][/a]
And the even more interesting question - can they actually be ABXed in *normal music*? In other words: would it make any difference if lame removed them or not?
- Lyx
What exactly is "holographic subliminal messages in audio"?
[a href="index.php?act=findpost&pid=349962"][{POST_SNAPBACK}][/a]
I believe he's referring to what in scientific literature are labelled as "binaural beats".
They can induce interesting psychological and consciousness effects in the listener and have been studied fairly extensively.
I see no problems in mp3/lame handling binaural beats to sufficient quality. The frequencies, amplitude envelopes and stereo separation issues should be fairly easy for lame.
Actually, "holographic subliminal messages" and binaural beats, a.k.a. "Brain Wave Generators" or called by the Monroe Institute, "Hemi-Sync" are different things.
I don't really understand the subliminal technology so much, but perhaps I can make some samples available later on. The idea is that powerful messages are spoken with emotion by someone, like, "You are a passionate, driven person", and then they are somehow encoded in audio using stereo effects in such a way that the conscious mind cannot hear them, but apparently, they will be picked up by your subconscious.
Binaural beats, on the other hand, have been found useful in producing syncronized brainwave patterns between the brain hemispheres... where you have a slightly different tone played in each ear, and then the beat frequency (heard in your mind, but not in reality) resulting from the difference of the frequencies actually can entrain your brainwaves to reflect that frequency. Some people use this type of brainwave generation, Hemi-Sync, et al for meditation and experimentation with out of body experiences and other interesting things.
The common ground here is that I was wondering if MP3 (or other forms of audio compression) would wreck the stereo separation or introduce artifacts that could impair the functionality of either technology.
While on the subject, although this is the wrong area, I wonder if a hybrid codec like wavpack lossy would be well suited to this kind of task?
It's mostly not relevant anyways... if you are in a scenario where you can use wavpack lossy, you could probably just as well use any lossless format, but I was wondering how one could pack their subliminal tapes with them on a DAP or portable device easily.
The idea is that powerful messages are spoken with emotion by someone, like, "You are a passionate, driven person", and then they are somehow encoded in audio using stereo effects in such a way that the conscious mind cannot hear them, but apparently, they will be picked up by your subconscious.
Kinda reminds me of those stereoscopic noise images (which I could never see).
The basic idea, being influenced somehow by hearing mantras while asleep (or whatever), is odd enough, but with THAT added to the mix it's even odder. How's this extra obfuscation supposed to help anyway?
And how does the word "holographic" fit in? Holography is visual, unless it's about documents written and signed by the same person, and that's even less related. ;)
I confirm that in CBR, the 10 Hz beating between 400 and 410 Hz is perfectly encoded.
Pio2001, do you mean in any bitrate, or at 320 kbit/s?
As for choosing the stereo mode, you can't seem to tell a difference between -m s (L/R stereo) and -m j (Joint Stereo), right? Implying that it would make more sense to use Joint Stereo, which normally gives higher levels of quality, especially at low bitrates. Perhaps for my own purposes I'll use L/R stereo for anything 256 K/bit and over, since the quality shouldn't be impacted too noticably at that high of a bitrate.
Pio2001, do you mean in any bitrate, or at 320 kbit/s?[a href="index.php?act=findpost&pid=350066"][{POST_SNAPBACK}][/a]
The last one I tried was the 320 kbps one.
So when a company releasing this type of audio states that only uncompressed files will work (http://www.centerpointe.com/about/faq.php#4.5), what is its motivation/cause to do so? Cursory testing of lossy/lossless codecs? Maintaining a simple, easily-enforcable product support policy?
They hope it will lead to less theft.
Not having a clue; using lousy codecs; using low bitrates. All sorts of things.
So when a company releasing this type of audio states that only uncompressed files will work (http://www.centerpointe.com/about/faq.php#4.5), what is its motivation/cause to do so?
You have to tweak the command line options in order to get a proper lossy encoding (enforcing CBR in mp3). The standard setup of a lossy encoder produces a poor result.
Why not try other codecs ?
So when a company releasing this type of audio states that only uncompressed files will work (http://www.centerpointe.com/about/faq.php#4.5), what is its motivation/cause to do so? Cursory testing of lossy/lossless codecs? Maintaining a simple, easily-enforcable product support policy?
CONVERTING TO THE FOLLOWING AUDIO FORMATS WILL RENDER YOUR HOLOSYNC SOUNDTRACKS USELESS: MP3, WMA lossless and lossy, Apple lossless, and AAC.
So when a company releasing this type of audio states that only uncompressed files will work (http://www.centerpointe.com/about/faq.php#4.5), what is its motivation/cause to do so? Cursory testing of lossy/lossless codecs? Maintaining a simple, easily-enforcable product support policy?
CONVERTING TO THE FOLLOWING AUDIO FORMATS WILL RENDER YOUR HOLOSYNC SOUNDTRACKS USELESS: MP3, WMA lossless and lossy, Apple lossless, and AAC.
so ogg vorbis is alright then.
...and of course, lossless compression destroys the sound
So I take back what I said earlier. They just have no motivation for writing this. This is pure nonsense.
Lossless codecs, like, ruin the whole VIBE, man!
Clearly, the extra processing that lossless requires alters the sound, thus destroying the effect.
My friend has a ton of his CD's encoded in apple lossless. Should I inform him that they have been rendered useless?
Jokes aside, even if something were, in theory, completely devoid of any redundancies anywhere that a lossless codec could use to save space, would that not simply result in the losslessly-compressed file being equal in size to the original ?! (or even a tiny bit larger?)
Yes. Assuming there isn't some sort of bug with the codec, the file should still be lossless, even if larger.
Jokes aside, even if something were, in theory, completely devoid of any redundancies anywhere that a lossless codec could use to save space, would that not simply result in the losslessly-compressed file being equal in size to the original ?! (or even a tiny bit larger?)
If you put all possible bitstreams of a particular size through a lossless codec, the sum of all their sizes will remain constant (actually they will get bigger because of headers and things). Of course, with sensible bitstreams the picture is different.
Back to the original topic: if the effect that these recordings have is noticeable in any way, you could answer your question with an ABX test. Of course, lossy codecs are designed to be perceptually transparent - they are not designed to retain other properties of the signal, including strange "subconscious" effects. I think you are better off sticking with lossless.
I think most people here know (but it's worth repeating...) that if you want to compress really simple signals which, for whatever reason, lossy codecs might choke on (e.g. sines with odd phase/frequency relationships, impulses etc) , lossless audio encoders are one option, but good old zip (or rar, or...) can be even more efficient!
That's not much use for the subliminal tracks, but perfect for binaural beats.
Cheers,
David.
I think most people here know (but it's worth repeating...) that if you want to compress really simple signals which, for whatever reason, lossy codecs might choke on (e.g. sines with odd phase/frequency relationships, impulses etc) , lossless audio encoders are one option, but good old zip (or rar, or...) can be even more efficient!
That's not much use for the subliminal tracks, but perfect for binaural beats.
An excellent point. The main issue with archive filetypes, though, is playability on a digital jukebox. I assumed that Apple lossless on the iPod wouldn't be a problem. I then saw that in their FAQ and went 'huh?' To be fair, the lossless files
felt like they were working when I tried them, but that could easily be placebo or the effect of just meditating regardless of auditory input. I have would zero confidence in passing a 14-for-16 ABX test on it.