Stereo vs. Joint Stereo - Page 2

Lite · 02-09-2009, 02:36 PM

I hope you will come to that yet. To hear a difference and feel it, I went to this a long time. But you must have reasonably good hearing i think. I wish he is still hear.

As for AAC, I meant a general concept. In AAC-Plus using of SBR greatly improves the quality in bitrates lower than 120, but also greatly cut the upper limit of frequency at about 16-17 kHz on best AAC+ encoder (Nero Digital Audio Encoder). But it by ear. No precise measurements were made.

So, how about a link where you get the "quote" about Joint Stereo?

t4e · 02-09-2009, 04:44 PM

The Ins and Outs of Audio Compression

Tarek · 02-09-2009, 06:05 PM

Quote:

Originally Posted by XDR

What it would do is create more space for relevant information in a similar-sized file.
An mp3 of the same size would sound better in Joint-Stereo than in Stereo because it contains more relevant data on the music.
I'm not sure if this is exactly how it works but this seems to me what he is implying.

Oh yeah..I forgot we were starting with CD quality..it makes sense..

dfx · 02-09-2009, 09:32 PM

another aspect of joint stereo is a simple mathematical trick.

instead of encoding the stereo audio data as two channels LEFT and RIGHT (which basically leaves half the bitrate for each channel), the audio data will be encoded as MID and SIDE. you still have two channels, just the meaning is different. the MID channel is basically the average (or you can also say the sum) between the LEFT and the RIGHT, while the SIDE channel is the difference between LEFT and RIGHT. the advantage of this is that the difference between LEFT and RIGHT is often very small, resulting in a SIDE channel comprised of very small values (more closely around the zero mark than any of the other channels). and small values can just be compressed better and easier (because in binary they're largely consisting of zeroes), leaving more bits for the MID channel and thus more bits for more details of the audio data.

an additional advantage is that in joint stereo mode, the mp3 encoder can decide on a frame-per-frame basis whether to use left/right coding or mid/side coding for each particular frame - depending on which mode can retain more audio details for the particular frame. so basically joint stereo is a win-win situation.

i try to use joint stereo whenever i can and whenever i have the choice. but i also wouldn't jump through hoops to use it, dual channel mode is quite fine too. the reason is that EDM is engineered very delicately for stereo seperation, so there is quite a lot of information left for the SIDE channel. i got curious to how much it really is, so i took a random track (in WAV format) and tried to simulate mid/side coding in audacity. see the screenshots below for the results. the top two tracks are the LEFT and RIGHT channels, the third track is what would be the MID channel and the last track is the SIDE channel.

the first three shots show the full track, while the last three shots show a zoomed-in part of the track. each three shots are waveform, linear spectogram and logarithmic spectrogram. in the latter you can nicely see how the bass and bass kick are missing from the SIDE channel, however it's usually more the high frequencies the are problematic in encoding and are more likely to be dropped due to insufficient bits left in the mp3 stream (the linear spectrogram shows these).

so my personal conclusion is that while joint stereo will always leave more bits for audio information in the stream and thus is always better to use than regular stereo (at least with a modern encoder), the audible difference at high bitrates will be minimal to nothing.

Lite · 03-09-2009, 01:16 AM

This is exactly the algorithm used by Lame. Called Safe Joint Stereo.

Quote:

Originally Posted by www.hydrogenaudio.org

The Joint-Stereo mode of lame (especially with VBR and above 160kps CBR) is Safe-Joint-Stereo mode. i.e. It will decide dynamically whether to use M/S (Mid Side or commonly known as Joint Stereo) or L/R (Left Right or commonly known as Stereo) frame. But most frame analyzers will show the whole file as Joint Stereo.

But the difference can be heard.
To hear the difference Monitor Headphones is needed. DJ headphones are not suitable.
At least not all of them is suitable. I use Philips SHP8500\00. They were enough to hear.
We must listen to the overall emphasis, volumetric of frequencies, sharp acid sounds. And of course original must be lossless and near to make a compare. Once heard and understood it more, you can feel the difference without any additional effort. Good to hear Psy Trance / Full On. For example on "Digicult - Out Of This World" album or "Ananda Shake - Inside the Sound" album.

t4e · 03-09-2009, 09:25 AM

Quote:

Originally Posted by Lite

But the difference can be heard.
To hear the difference Monitor Headphones is needed. DJ headphones are not suitable.
At least not all of them is suitable. I use Philips SHP8500\00. They were enough to hear.
We must listen to the overall emphasis, volumetric of frequencies, sharp acid sounds. And of course original must be lossless and near to make a compare. Once heard and understood it more, you can feel the difference without any additional effort. Good to hear Psy Trance / Full On. For example on "Digicult - Out Of This World" album or "Ananda Shake - Inside the Sound" album.

right there you are agreeing with my view that the sound quality is not noticeable unless you use specialized equipment

regular listeners will not listen on monitor headphones

Lite · 03-09-2009, 11:36 AM

But at bitrates lower than 320kbps heard, even on normal soap box, let alone monitor headphones. ;)

Lite · 03-09-2009, 01:49 PM

Ah, yes! I constantly forget, that not all people have a good hearing. Specific features have an effect. And if someone does not hear the difference, then he's hearing simply not developed enough. I always forget. Sorry.

t4e · 03-09-2009, 02:03 PM

has nothing to do with having perfect hearing, most people have normal hearing

you just made a statement saying that you must listen in monitor headphones to hear the difference...if that's the only way to hear the difference the experiment is useless

Lite · 03-09-2009, 03:01 PM

I knew you'd say that ©

Tarek · 03-09-2009, 05:13 PM

Quote:

Originally Posted by t4e

has nothing to do with having perfect hearing, most people have normal hearing

you just made a statement saying that you must listen in monitor headphones to hear the difference...if that's the only way to hear the difference the experiment is useless

t4e I looked up his headphones and their specs..they're not that out of the ordinary. The frequency ranges are a bit bigger than the standard 20-20k (his are 10–28,000 Hz) , but the sensitivity rests at 106db ...of course, his own hearing is a factor, and I cant really judge based on specs, but the capacity to notice the difference is higher (deeper basses / rounded highs)

My point is, more and more headphones will surpass the 20-20k mark once you hit over an average of 40 $ on headphones for personal listening. DJ headphones are even more likely to have a wider sound spectrum than general consumer headphones dubbed "hi-fi".

Also, at the risk of pushing a stereotype forward and potentially offending people, older people tend to care more about sound quality than younger people (I happen to be an exception, and man..am I glad to be one)..therefore the chances of people on this forum having decent headphones (judging by what I am assuming are the people's average age groups) is fairly good.

What I think would happen if we switched to Joint Stereo is that no one would notice..but if someone like Dan announces its upcoming, and maybe is able to give a demo on the air , many people will see the difference and appreciate it, even if it seems minimal a change..after all, this IS a free service...

Lite · 03-09-2009, 05:37 PM

Why I say that DJ headphones do not fit all, because they mainly focus on the beats and monitor headphones is another matter.
By the way I took these headphones on these tests:
Testing the monitor 23 Hi-Fi-headphones on hardnsoft
As the next model SHP8900/00 worse on the tests. Much better SHP9000/00. But they are much more expensive in cost.
And SHP8500/00 are much better than many of KOSS, AKG, Sennheiser etc... in this and even higher price range

Tarek · 03-09-2009, 05:59 PM

The DJs focus on the beats, but the headphones still do their job with the full spectrum..even though they may be EQ'd from the manufacturer, it does not shut out monitoring potential whatsoever.

The website you link to seems to compare headphones they pulled out of a hat, with no regard to age, price range, or intended consumer..

Lite · 03-09-2009, 06:15 PM

Price range can be seen at the top of page text of each model of headphones, a distinction of age, I have never seen anywhere else.
For example on that page

Quote:

Price range from 50 to 100 dollars

Here another specialized website http://doctorhead.ru/

An interesting article on another specialized site Headphones and human hearing - stereohead.ru

English such sites do not know.

t4e · 03-09-2009, 10:30 PM

Quote:

Originally Posted by Tarek

t4e I looked up his headphones and their specs..they're not that out of the ordinary. The frequency ranges are a bit bigger than the standard 20-20k (his are 10–28,000 Hz) , but the sensitivity rests at 106db ...of course, his own hearing is a factor, and I cant really judge based on specs, but the capacity to notice the difference is higher (deeper basses / rounded highs)

My point is, more and more headphones will surpass the 20-20k mark once you hit over an average of 40 $ on headphones for personal listening. DJ headphones are even more likely to have a wider sound spectrum than general consumer headphones dubbed "hi-fi".

Also, at the risk of pushing a stereotype forward and potentially offending people, older people tend to care more about sound quality than younger people (I happen to be an exception, and man..am I glad to be one)..therefore the chances of people on this forum having decent headphones (judging by what I am assuming are the people's average age groups) is fairly good.

What I think would happen if we switched to Joint Stereo is that no one would notice..but if someone like Dan announces its upcoming, and maybe is able to give a demo on the air , many people will see the difference and appreciate it, even if it seems minimal a change..after all, this IS a free service...

you missed the point i was making....regular listeners don't sit around with headphones, at least not all the time, let alone monitor headphones, they chillout some place in their home with music playing out of the speakers

i tell you what would happen if Dan announced he introduces Join Stereo, the same as it happened when some station switched from 192 to 256,
everyone claimed the sound was better....that's BS, human ear at normal listening can not detect the difference, however the human brain will try to "hear" a better sound because we've been told that higher bitrates have higher quality