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: my VBR analogy (Read 6363 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

my VBR analogy

so perhaps it's worth writing down. This is how I've explained to people why VBR is more efficient, in principle than CBR encodes.
Imagine a running race over a hilly course. The top two competitors are of approximately equal talent and fitness, but have different racing tactics. Runner A keeps a constant pace, running as fast up the hills as he does on flats. Runner B keeps a constant effort level, so he runs slower uphill than he does on flats (assuming both runners have little biomechanical choice but to go fast on downhills). Assuming equal expenditure of effort, Runner B will finish ahead of Runner A. Or, if they finish together, Runner B will have used less energy than Runner A, because Runner B was more efficient with his use of energy. Here, a constant effort is parallel to a constant audio quality.
(Maybe this analogy only won't be helpful to nonrunners - it seemed pretty clear to me, but then I once won a race where I actually walked up the longest hill.)
God kills a kitten every time you encode with CBR 320

my VBR analogy

Reply #1 –
Quote
so perhaps it's worth writing down. This is how I've explained to people why VBR is more efficient, in principle than CBR encodes.
Imagine a running race over a hilly course. The top two competitors are of approximately equal talent and fitness, but have different racing tactics. Runner A keeps a constant pace, running as fast up the hills as he does on flats. Runner B keeps a constant effort level, so he runs slower uphill than he does on flats (assuming both runners have little biomechanical choice but to go fast on downhills). Assuming equal expenditure of effort, Runner B will finish ahead of Runner A. Or, if they finish together, Runner B will have used less energy than Runner A, because Runner B was more efficient with his use of energy. Here, a constant effort is parallel to a constant audio quality.
(Maybe this analogy only won't be helpful to nonrunners - it seemed pretty clear to me, but then I once won a race where I actually walked up the longest hill.)
[a href="index.php?act=findpost&pid=319132"][{POST_SNAPBACK}][/a]


Okay, what was the purpose of this analogy??? 
You messed up, now I gotta mess you up. It's the law!

my VBR analogy

Reply #2 –
Quote
Okay, what was the purpose of this analogy???  <_<
[a href="index.php?act=findpost&pid=319166"][{POST_SNAPBACK}][/a]

As you may or not know, explaining matters via visual analogies, is one of the best ways to explain complex technical stuff to "outsiders". This is because it is easier to apply a known scheme to a new object, instead of reinventing the wheel(starting from scratch).
I am arrogant and I can afford it because I deliver.

my VBR analogy

Reply #3 –
How about the old standby?

CBR = constant bitrate, variable quality
VBR = constant quality, variable bitrate
Was that a 1 or a 0?

my VBR analogy

Reply #4 –
Quote
As you may or not know, explaining matters via visual analogies, is one of the best ways to explain complex technical stuff to "outsiders". This is because it is easier to apply a known scheme to a new object, instead of reinventing the wheel(starting from scratch).
[a href="index.php?act=findpost&pid=319167"][{POST_SNAPBACK}][/a]

Quote
How about the old standby:
CBR = constant bitrate, variable quality
VBR = constant quality, variable bitrate [a href="index.php?act=findpost&pid=319183"][{POST_SNAPBACK}][/a]

Yeah, maybe it wasn't worth putting in here at HA - I'm trying to get across not only the definition (what DigitalMan wrote) but also why VBR is advantageous. The analogy has helped me explain this to a couple of people, but then, I've got a good % of my friends (from college, at least) through the cross-country and track teams.
God kills a kitten every time you encode with CBR 320

my VBR analogy

Reply #5 –
Here's mine:

Imagine if you will, a microwave oven, it just heats stuff up at the level wich you select, sometimes you hit the nail on the head and get nice hot <microwaveable product> with the settings for power and time you used, but sometimes your product is still frozen, or explodes inside the oven, this is like CBR.

Now a magic microwave oven appears wich looks at your <microwaveable product> and adjusts it's temperature and cooking time itself to give you great results, without having to guess anything yourself. This is like VBR.

How's that?
we was young an' full of beans

my VBR analogy

Reply #6 –
Quote
Here's mine:

Imagine if you will, a microwave oven, it just heats stuff up at the level wich you select, sometimes you hit the nail on the head and get nice hot <microwaveable product> with the settings for power and time you used, but sometimes your product is still frozen, or explodes inside the oven, this is like CBR.

Now a magic microwave oven appears wich looks at your <microwaveable product> and adjusts it's temperature and cooking time itself to give you great results, without having to guess anything yourself. This is like VBR.

How's that?
[a href="index.php?act=findpost&pid=319200"][{POST_SNAPBACK}][/a]

Interesting. Just one minor flaw: In CBR, you can cause no damage by choosing a too high bitrate(you will just waste space). Your food in the microwave on the other hand would make fore a nice splatter-movie ;-)
I am arrogant and I can afford it because I deliver.

my VBR analogy

Reply #7 –
LOL!

good point.
we was young an' full of beans

my VBR analogy

Reply #8 –
lets say i have 1 foot by 1 foot perfect square diamond and i want to make a ring for my g/f. if i reduce it down to a 1 inch by 1 inch diamond its still pretty ugly because its a square. now lets say i take a little bit off in someplaces and lot of bit off in others, then i have nice looking diamond.
[span style=\'font-size:8pt;line-height:100%\']"We will restore chaos"-Bush on Iraq[/span]

my VBR analogy

Reply #9 –
I think tim (the threadstarter) was on the right track. VBR is just a more efficient way to manage a given "resource". Thus, i guess an analogy which is about some kind of resource and two approaches to "spent" this resource would fit. The hill-example may be a bit more complex than necessary, but the basic idea behind it seems fitting to me.
I am arrogant and I can afford it because I deliver.

my VBR analogy

Reply #10 –
You can also make an example with vessels and then filling them with water. Different vessel sizes reflect the difficulty to compress the audio (e.g. small jar is silence and a big bucket is noise), CBR would then be trying to fill all the vessels with the same amount of water and VBR would be filling the vessels according to the vessel size. The amount of water (tried) to get in each vessel represents the amounts of bits used

With CBR obviously water gets spilled on the small vessels and the large vessels aren't totally filled up (unless large amounts of water are used).
"We cannot win against obsession. They care, we don't. They win."

my VBR analogy

Reply #11 –
Let me confuse you:

Even a CBR mp3 file contains frames using more bits and frames using less bits. CBR mode is about allocating always a fixed sized frame. Due to  the short time buffer--also known as bitreservoir--you can hold back some bits on easier parts and use them on not so easy to encode parts. The short time buffer is limited in size (well it's a short time buffer). CBR is not so flexible as VBR, iff the encoder can't make use of whole frames, the encoder wastes bits (ancillary data) or encodes at higher quality as the psychoacoustic model says is necessary. Iff the encoder would need more bits than assigned frame sizes provide, encoder can't get the quality the psymodel would say is necessary.

my VBR analogy

Reply #12 –
I find making the so called "outsiders" believe highly optimized encoders like LAME actually know when to save bandwidth harder than getting them to grasp what VBR is about...

EDIT: grammar... as usual...
WavPack 5.8.1 -b384hx6cmv / qaac64 2.84 -V 100

my VBR analogy

Reply #13 –
The encoder would probably waste most bits on higher precision, instead of trying to achieve constant quality. And bit reservoir is 511 bits maximum? Just a few drops!
"We cannot win against obsession. They care, we don't. They win."

my VBR analogy

Reply #14 –
Quote
I find making the so called "outsiders" believe highly optimized encoders like LAME actually know when to save bandwidth harder than getting them to grasp what VBR is about...
[a href="index.php?act=findpost&pid=319630"][{POST_SNAPBACK}][/a]

That is because of “save bandwidth”, which almost automatically leads to “OMG, it shaves off quality as a tradeoff, i'd better be using CBR/192/stereo”…
Infrasonic Quartet + Sennheiser HD650 + Microlab Solo 2 mk3. 

my VBR analogy

Reply #15 –
Quote
That is because of “save bandwidth”, which almost automatically leads to “OMG, it shaves off quality as a tradeoff, i'd better be using CBR/192/stereo”…

Simple response: unless you encode at 320kbit CBR, the encoder will *ALWAYS* shave off quality - the difference is just that with 192kbit CBR MS it will blindly shave off quality, while with VBR JS it will intelligently shave off quality where you won't notice it. So you loose quality anyways - the choice is just: should it be more or less noticable?

(Note: for simplicity, i ignored the bit-reservoir)

- Lyx
I am arrogant and I can afford it because I deliver.