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: VBR vs. CBR in regards of quality fluctuations (Read 24698 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

VBR vs. CBR in regards of quality fluctuations

Reply #25 –
well

16 bit 2ch 44.1 kHz = 16*2*44.1 = 1411.2 kbps which is 4.4 times 320 kbps so not really about half

and yes, there is a reason why CBR 320 kpps is called "insane" in the presets
"We cannot win against obsession. They care, we don't. They win."

VBR vs. CBR in regards of quality fluctuations

Reply #26 –
Quote
Using CBR at 320kpbs, thats about half the bitrate of uncompressed 16bit 44.1kHz, wouldnt it rather then make more sense to simply use an compressed lossless format like FLAC (which IIRC also gives an average saving of 50%)?
[a href="index.php?act=findpost&pid=264623"][{POST_SNAPBACK}][/a]

320kbps is not the half...it's 4.4...705kbps would be the half...also, lossless doesn't give you 50% either...you rather safe 33%...
--alt-presets are there for a reason! These other switches DO NOT work better than it, trust me on this.
LAME + Joint Stereo doesn't destroy 'Stereo'

VBR vs. CBR in regards of quality fluctuations

Reply #27 –
Damned, I forgot to consider its stereo :-P My fault.
So, even if you cant hear any difference there is a quality fluctuations in CBR 320kbps? Or not....?

VBR vs. CBR in regards of quality fluctuations

Reply #28 –
No more than any of the VBR presets, which is usually none barring problem samples.  320 is using the maximum frame size for every frame.  256k CBR and lower may have these so called quality fluctuations
Nero AAC 1.5.1.0: -q0.45

VBR vs. CBR in regards of quality fluctuations

Reply #29 –
Quote
No more than any of the VBR presets, which is usually none barring problem samples.  320 is using the maximum frame size for every frame.  256k CBR and lower may have these so called quality fluctuations
[a href="index.php?act=findpost&pid=264650"][{POST_SNAPBACK}][/a]


And the reason why CBR does not is due to bit reservoir?

So, just to sum it up, DreamTactix you say quality does not fluctuate, and SirGrey says it does fluctuate... Who is right?

VBR vs. CBR in regards of quality fluctuations

Reply #30 –
The quality obviously fluctuate when using CBR. Only vbr can try to keep constant quality.

VBR vs. CBR in regards of quality fluctuations

Reply #31 –
Quote
The quality obviously fluctuate when using CBR. Only vbr can try to keep constant quality.
[a href="index.php?act=findpost&pid=264729"][{POST_SNAPBACK}][/a]

Thanks for answering my question! I think everything is clear now, you as a LAME dev should know (I hope :-) ).

VBR vs. CBR in regards of quality fluctuations

Reply #32 –
Concerning quality fluctations, would it be correct to think of a transparent VBR file as a boundary for perceivable fluctations (assuming it's the smallest possible, but still transparent, VBR file possible. Typically the sweet spot aps is aiming for)? So, whenever the bitrate exeeds this limit, it will have a higher guality in terms of (less) noise, but not in terms of perceived quaility. Making it a waste of space.

Edit: forgot the question mark...

VBR vs. CBR in regards of quality fluctuations

Reply #33 –
Quote
Higher than 320 kbps is not possible. So the VBR file is not VBR, cause bitrate is fixed at 320.

Since 320 kbps is the highest possible, offcourse CBR will bring the best results, because VBR will always use a lower average bitrate due to it's nature. Chances you're gonna hear it are quite low though. I suggest you do some ABX testing and start it real easy with 128 kbit mp3. (I bet it will be more difficult than you think)

<edit> CBR will probably always lose compared to VBR at the same filesize, so I don't see the logic of comparing CBR/VBR when using different filesizes.  </edit>
[a href="index.php?act=findpost&pid=264493"][{POST_SNAPBACK}][/a]



Is freeformat mp3 higher than 320 KBPS?

I was wondering, because allofmp3.com has freeformat encodes available
( or did)

VBR vs. CBR in regards of quality fluctuations

Reply #34 –
About quality fluctuations:

Okay, i need to explain this in two steps.....

step one(practice): we asume that 320kbit can <always> achieve unlimited *technical* quality on any weird sample

In this case: VBR = constant quality, CBR below 320kbit = fluctuating quality, CBR@320kbit = constant quality.


step two(theory): we asume that 320kbit can not achieve transparency in some ultra rare problem cases:

In this case VBR as well as CBR@320kbit does almost always achieve transparency, but as mentioned above there are some ultra-rare problem cases where a minor difference can be detected. This means logically that in these problem-cases *perceivable* quality fluctuates because everything else is encoded transparent.
_______

The bottom line is this: ignoring the MP3-format, VBR always achieves constant *technical* quality, while CBR *technically* does not. But the MP3-format has a limit of 320kbit - when this boundary is hit in VBR-mode, then the *technical* quality logically fluctuates even in VBR.

However, all this is purely theoretical when it comes to *perceivable* quality - you will probably never notice it. In reality, you can just asume that "step 1" is true.

Or you can just encode at --preset standard and stop worrying about all this confusing stuff while still sleeping peacefully at night.

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

VBR vs. CBR in regards of quality fluctuations

Reply #35 –
Quote
Does this mean with CBR the less complex frames will sound better than those complex frames as you have more then enough bits (enough bits=number of bits which VBR will spend) left? Or will this somehow and to which extend be compensated by CBRs bit reservoir technique?


The reverse is also true with CBR. Let's say a particular frame is transparent, ie. indiscernable from original source music, for the listener at 192kbps. If you choose to encode this at 160 CBR, then the quality of that frame will be worse than if it were encoded by a sufficiently high quality VBR encode.

The point of VBR is to maintain constant quality as chosen by the encoding individual. Assuming it is of sufficient quality (eg. Lame VBR preset standard), the quality of any given frame will probably be better than that of a CBR encoded file (obvious exception: 320 CBR).

VBR vs. CBR in regards of quality fluctuations

Reply #36 –
If the methods used to calculate necessary frame size in VBR were PERFECT (which they are not), then a VBR file of maximum quality would sound identical to a 320kbps CBR file, at hopefully smaller size. In reality, because the VBR method fails sometimes, a 320kbps CBR file will probably be safer...

having said that, 320kbps is overkill, and a VBR file around 200kbps (like --preset standard) will sound identical, 99% of the time.