What's everybody's thing with VBR here?
 

I'd just like to know why the people here that call themselves audiophiles prefer VBR over 320 CBR. Even gamemp3s has some weird obsession with it.

I believe "audiophiles" would prefer lossless encodings.

However, as far as lossy encodings go, VBR has many advantages over CBR. By using a variable bitrate you eliminate the problem of certain frames using a higher bitrate than what is necessary. See, even if a track has certain frames, which when encoded in API (CBR 320) displays a noticable incease in quality over a say, CBR 160 counterpart, you can be sure that far from all the frames on the given track would display the same inceased quality. So basicly, if you use API over a VBR encoding string that maxes out at 320 you're simply wasting your harddrive space.

Hell, I doubt people would even be able to tell the differance between properly encoded CBR 128 and API most of the time.

I wasn't talking about lossless formats as the people that share music around here deal in MP3, and they're the same people that claim to be "audiophiles".

What you're talking about is basically saving hard drive space. But quality is a technical thing, isn't it? I find it hard to honestly say that a song that plays at 320 kbps sometimes is better than a song that plays at 320 kbps all the time. That's like saying 100% sometimes is better than 100% all the time: it doesn't make any logical sense. That whole thing about frames was really just extra info, at the end of the day, 100% all the time is better than 100% sometimes, it's logic. Whether you can hear it or not doesn't matter, technically it is better.

But you answered my question, it saves space. I just thought the "audiophiles" around here would go for the best. Thanks.

And odd thing about variable bit rates is that it only seems to be so "variable". I encoded the same song twice at the same variable bitrate and the bitrates repeated the same pattern both times. Interesting.

No, because there are portions of the waveform that do not require 320kbps. When your song contains a few moments of specific sound that would only need say 96kbps to faithfully replicate, demanding 320CBR is ridiculous.

'Variable' in this context doesn't mean 'random;' it indicates that the bitrate shifts to accommodate the file's waveform. Since you're sampling the same song twice, it should be obvious that you're going to get the precise VBR behavior both times.

yeah it's only variable between the frames, not between two songs. logically it would have been so random that you could miss a lot of stuffs if you're unlucky by getting almost 32kbps. how sad.

Actually, "logic" as you're using it doesn't hold up here. Why? Because you have incosistent proofs and are basing it on false premises. Thus, your conclusion doesn't hold up. That's logic.

100% in relation to what?

No. Just like everybody has said, at every moment in a tracks duration, VBR matches the highest bit rate needed at that moment. If there's a portion that needs 96kbps and no more, VBR gives you 96kbps; you are getting 100%. If there's a section that needs 192kbps and no more, VBR gives you 192kbps; you are getting 100%. You are getting 100% all the time.

Trench: If you can recognize that there are certain parts of a given track that would require more, power, a higher bitrate than other parts, isn't it then only logical to encode the track in a way that is in accordance to that? Obvious parts of intrest being the few seconds (or whatever it may be) of silence at the beginning and end of the track. And I mean, even if you were to compare a lossless version to a (properly encoded) VBR copy I'm sure there would be many parts were any audiable differance would be impossible to distinguish.

What I find pretty intresting is how this debate always seems to be limited to the question of bitrate vs. bitrate, when the question of what encoder you use is so much more important to the quality of the track. You will see people passing of a FhG CBR 320 encoding as highest possible quality merely because of the CBR 320 part when in actuality it will be fundamentally flawed since FhG is such a bad encoder.

Let's face it, unless you have superhuman hearing you will not hear sufficient (if any) differance between different VBR encodings or between VBR and CBR 320 (if they've all been encoded using the same version of Lame).

This sounds like something some dumb "audiophile" would say. Whether it is technically better is completely irrelevant unless you can actually notice any audiable differance, right?

Using CBR is a bit like saving a black and white GIF image with a 256 colour palette. It's pretty pointless because you're using a larger file than necessary to represent the original faithfully. It's really as simple as that.

I like that you're turning this free music thing into an insult against your character or some shit, though. It's precious. It's like the Hamptons have come to Gamingforce.

The real justification for VBR doesn't come at the 320kbps vs -V 0 debate, but rather the lower analogies.

For example: 128kbps vs -V 5 (which averages 125-140kbps); if a song contains a lot of silence, VBR will encode that as 32kbps and save space, then ramp up all the way to 320kbps for the most complex audio while maintaining a bitrate in the range laid out above. CBR will just slather 128kbps over everything, treating silence and the most complex audio as the same thing, while ensuring a consistent size (which ABR does better for consistent size/quality ratio BTW) makes the song overall lower quality. I dare anyone to test a CD encode of 64kbps and -V 9 and honestly tell me that 64kbps CBR is a better choice for quality. The only two uses I can ever see for CBR anymore is legacy support for ancient mp3 devices and web streaming, other than that CBR is worthless.

On this note you'll notice that lossless audio always codes as VBR, why? Simply because lossless compression cannot treat all data the same and ensure prefect quality, the only way to have lossless data as CBR is not to compress it at all. Yes I do realize 320kbps still compresses, but it introduces so tiny of an improvement in quality compared to -V 0 that I consider the size vs quality difference (being 64-96kbps) a waste of spave

I used to use APX all the time, but now I tend to use API. I just don't see how any frame of an MP3 can suffice with anything less than 320k and still make a perfect sound. And who decides what bitrate is needed during VBR encoding? Is it programmed or all automated? Surely you need 320k for every frame to ensure perfect sound?

I did my own encoding tests using the same source Wav and encoding to APX and API. The API yeilded a slightly higher frequency response.

I also have two huge HDDs so space isnt a problem for me.

What made you think that the 'perfect' value is 320 kbps? Technically speaking, MP3s can never be perfect since data is removed from the source regardless of the compression level used. WAV files taken from CDs are something like 1400 kbps.

I think when you say 'perfect', what you actually mean is 'transparent'--meaning no audible difference between the input source and output MP3. The point of VBR is realizing that since the format can't be perfect due to it being lossy, a set of algorithms to determine the complexity of each frame and allocating a number of bits to accomodate it reduces the filesize while still resulting in an output file that is transparent to most everyone.

So, to answer your questions: first, it is both programmed and automated since the algorithms were programmed and they automatically determine the bitrate for each frame. VBR is an attempt at compromise between reducing filesize (which, of course, is the whole point of the MP3 format) while sacrificing only audio data that does not affect the transparency of the output file.

I would be interested to learn the procedure of your encoding tests, since it has been shown via double blind tests on high-end equipment that the difference between high-end LAME encodes is pretty much undetectable. When you say 'a slightly higher frequency response', that sounds to me like you used some software application to determine the frequency range of the file. I'm sure you know this already, but the human ear is not suited to detect frequencies at the high and low extremes while software can.

Have you tried doing your own double blind test with the two files? In case you didn't know, a double blind test in this case is when you listen to two files and eliminate any placebo effect by not knowing which is which. Foobar2000 has this functionality built in, so you might want to try that. All you do is pick the two tracks in your playlist, right click them and go to Utils, then 'ABX Two Tracks' and follow the instructions. Basically it gives you the option to play A, B, X, and Y. A and B are the same as X and Y, and you have to match them up based on how they sound.

Finally, space isn't the only issue of course--there's also bandwidth since we're transferring these files to one another. And even though I have a terabyte of storage space between my 6 hard drives, space is always an issue.

You don't get perfect sound, even with 320k CBR. The purpose of VBR is to get relatively consistent quality.

You could say it's "automated". And automated very cleverly at that. And as I said, you don't get "perfect sound" with MP3 even at 320kbps. You don't need 320kbps for every frame for transparent sound because different passages require different bitrates.

Probably because it uses a slightly higher lowpass (or none) as far as I know. Frequency response is not the be-all and end-all of lossy encoding.

Ok, it was late and I wasnt thinking. MP3 is by no means perfect, that was my mistake lol. What I mean is, I fail to see how ANY frame of music needs less than 320k to make the best sound possible. If you give every frame the maximum amount of bitrate, you are making a good a sounding file as possible.

Anything I rip is primarily for my use. I dont stop to consider "I may end up sharing this at some point". Besides, many people on the FTPs have CBR files, so I dont think its that big a deal.

You get consistent quality with 320k as well, then. Every frame is encoded at the highest possible quality.

Even with this mindset, it still depends on the encoder you use, as every encoder will decide differently about which 320k of each frame is worth hanging onto.

But the point that people here are trying to make is that the difference between high-quality VBR and 320 CBR is so incredibly minute, that it's not worth the large jump in file size. CBR is inefficient, and that's the bottom line.

If you are so concerned about quality, just go lossless. Seriously. If you're such an audiophile, then why settle for anything less?

Because like 320CBR, quality-size ratios aren't worth it.

This brings up a question I've been wanting to ask for a long time.

Why does FLAC even exist in the first place? And I mean, aside from getting 1:1 copies of CDs?

If you wanted -real- 1:1 copies, you could just go out and buy the CD yourself. Of course, price becomes an issue, and I can see why people might download a FLAC rip if it means you save money compared to buying a CD. But since VBR exists, there's another problem.

I'm an audiophile myself; and I can't tell the difference between a properly-encoded VBR mp3 and a FLAC file. Not unless I had super-ultra-mega-extra-volcanic-krakatoa-hurricane-katrina-headphones, which I'm sure don't exist. But - if anyone could tell any difference - it would be so small anyhow. And frankly, I'd rather take a 100MB VBR rip instead of a 600MB FLAC rip, if the difference is that small. Hard drive space becomes a concern at some point and I know most people wouldn't want to waste it.

If you have been keeping up with the CD rips I've done, what I'm saying may seem contradictory since I've ripped so many albums in FLAC. But I'll admit something now - the only reason I may rip stuff in FLAC is because other people want them. Hell, I don't know why... but if you prefer having the Final Fantasy XII soundtrack in FLAC at 1.20GB, instead of having the same album in VBR at 400MB with very little to no audible quality difference, be my guest.

I'm not asking anyone to get rid of lossless formats entirely. I'm just wondering why some people go after and rip their stuff in lossless instead of in VBR.

Only reason I don't do or collect any rips in FLAC is because of hard drive space: plain and simple. I start hearing minute differences just below VRR. 192 is not a big deal to me, but 160 and below is usually where I tend to tell differences. VBR is just fine for me for now. Always has been

Some frames need it, some don't. So you're not getting consistent quality.

I like to download albums in lossless whenever possible because that gives me the best possible source for making my own MP3s, Vorbises, etc. from the content. Sure, I could download WAV instead, but when you're talking about full albums, the difference between WAV and FLAC is a couple hundred megabytes.
When I have my nice Vorbises and MP3s, then I'll burn the FLACs for safe keeping (in case I want to do something else with them later) and delete them off of the hard drive.

I like to rip my own CD's to FLAC because then I have perfect backups in case something ever happens to the CD. It's also convenient in case in the future I want to convert a large chunk of my collection to another format - rather than transcoding from MP3 and taking a quality hit, or wasting time re-ripping the CD's, I can just pop in my DVDs full of FLACS and transcode away.

Yeah, and like I said I dont believe some frames dont need anything less than 320k. A lossless redbook CD is 1411kbps, so Im always going to give every frame the highest quality I can to ensure a very good sound production, i.e. 320kbps.

You'll have to pardon me, but that's simply ridiculous. The way your logic works, 5 seconds of pure dead silence requires the same amount of space as 5 seconds of distorted guitar.

Thats not what I said at all. I didnt mention anything like that.

I will say, though, that silence of course does not of course require the same amount of space as instrumentation. Im simply saying from any piece of music (silence excluded), I dont believe in giving every frame anything less than the maximum for the best sound reproduction.

The only time I really use VBR is -- for example -- if I have a piece of unreleased music from a digital TV music channel that I have recorded onto my computer that is already from a lossy source (some of these recordings from one channel dont have anything above 14khz or so). 320k would be a waste of time in this case, so I encode in APX.

Except many frames don't need 320kbps in Lame...

Also, I wish people would stop using APM, APS, APX, and API. Lame 3.97 is the officially recommended version of Lame, and it's based on a -V x --vbr-new system for a reason.

All Im saying is, I dont believe any material (unless its something like silence or really basic material like a drone) can still sound as good as the same material with the maximum amount of bitrate given to it.

With all due respect, I think you need to understand how the encoding process works.

Iceboy, I don't see a problem at all with using -b 320 in LAME if it's for your own personal use, is never going over a network, and you don't care about the extra space it takes up. All of which you've stated. LAME's VBR is meant to be as efficient as possible in the way it provides high quality at compressed file sizes.

Although, to be honest, your argument of "I dont believe in giving every frame anything less than the maximum for the best sound reproduction" could very well be used to justify keeping uncompressed WAVs around (or encoding into the lossless FLAC format). After all, MP3 is lossy. Don't you want the maximum quality for every moment of your audio? And there's certainly nothing more maximum than the exact original file! (Or one that outputs to exactly the same, in the case of FLAC.)

The point is that MP3, at any bitrate, is meant to reduce the size of audio files. The goal of high quality VBR modes, notably the old --alt-preset standard and newer -V 2 --vbr-new, is to use just enough bits to attain audio that is audibly no different from the original. Admittedly, this is only a goal, and the implementation is based on the very complex human hearing model and can never be quite perfect. However, with that disclaimer out of the way, a huge majority of folks will never hear any difference between such VBR modes and 320Kbps CBR. Many people actually have a very difficult time hearing any difference between files just over 128Kbps and the original audio, when challenged to a scientific double-blind listening test. Seeing the 320, much larger than 128 or 160 or 192, has a strong psychological effect that can cause you to think it sounds better when that is not actually the case.

But, as I said, it's your choice to make. Encoding in 320Kbps isn't a sin, maybe just a bit wasteful, but that's all.

Bandwidth and disk space are preciously limited resources, guys!

[/1999]

[/ozzie internets]

Until a few months ago, I was using lossless encoding for my CD singles. Then I realised how much space it was eating up, and re-encoded them as 320k MP3s. I dont mind using some hard disc space, but there is a line I draw!I so no point keeping lossless ones, because comparinf the 320k and lossless files it was hard to find any real differences in quality.

Thank you. That just goes to show why I tend to (but not always) opt for 320k.

As for psychological effects, etc. I know all about that. But I have fairly good hearing and I can definitely hear differences at 128k from the original WAV.

And thanks for being so fair! Until now, everyones slagged me off for using 320k. You're the first person to say "do what you want." So thanks man.

Glib Argument GET!

Am I missing something?:confused:

Cal gets that a lot, don't worry about it.

Bandwidth -- not so much, but hard drives aren't yet cheap enough that everyone could afford a 500 GB HDD just for complete discographies of ten bands. Maybe in a few years.