VBR or CBR for convert in mp3?

[cam]

Hello guys i'm cam and this my question

sorry my english is very limited

[pandy]

VBR unless CBR is required (lot of CBR are VBR anyway only stuffing is added to fill payload up to CBR)

[jagabo]

VBR gives better sound at low bitrates (smaller file sizes). But CBR is compatible with more devices. I always use CBR at 192 kbps or higher since ultra small files is not a priority.

[jman98]

Use CBR at the highest bit rate you can. VBR exists only because MANY YEARS AGO the devices that could play it didn't have a lot of space so VBR is a way to get more quality at lower bit rates. Lower bit rates means smaller file sizes, which means you can have more files. But now storage is cheap and big so there is no reason at all to use VBR today. If you can use CBR and 320 mbps then that is the best quality you can possibly get and VBR cannot be better than that at any setting.

[PuzZLeR]

I'm similar to the last two posts in opinion.

VBR is a nice concept in theory, kind of like quality-based encoding is for video, but in practice unfortunately it still gives many audio devices headaches for some reason. As well, any real quality advantage of VBR would be in very small bitrates that would reveal alot of lossiness anyway.

Since MP3 is small, and storage is now bigger than ever, just use CBR. It's safer this way for universal compatibility, and you still get great quality.

You can get away with CBR 128mbps for most content if you want real small file size. Use CBR 192mbps for better quality. Use CBR 320mbps for ultimate quality retention.

[pandy]

I'm similar to the last two posts in opinion.

VBR is a nice concept in theory, kind of like quality-based encoding is for video, but in practice unfortunately it still gives many audio devices headaches for some reason. As well, any real quality advantage of VBR would be in very small bitrates that would reveal alot of lossiness anyway.

Since MP3 is small, and storage is now bigger than ever, just use CBR. It's safer this way for universal compatibility, and you still get great quality.

You can get away with CBR 128mbps for most content if you want real small file size. Use CBR 192mbps for better quality. Use CBR 320mbps for ultimate quality retention.

Hmmm... it is only half story - lot of so called CBR are VBR with added stuffing packets to create CBR. Any VBR mp3 can be "promoted" to CBR without recoding. http://wiki.hydrogenaud.io/index.php?title=MP3packer

For ultimate quality retention mp3 is inappropriate, i would recommend FLAC or MPC (if lossy coding is not a problem)

[El Heggunte]

At 256, 320 and 384 kbps, MP2 is better than MP3 ^_^

And to whom this may interest, DTS should be "perceptually-lossless" at 255kbps per channel (for 48kHz, bien entendu ---
--- lower sampling frequencies require less bitrate, of course).

I am not a fan of Musepack, since it cannot be stored in ACTUAL containers (MKV, MOV, AVI, MPG, whatever) -_-

[fritzi93]

I use EAC and automatic VBR settings. Average bitrate is usually a little over 192 kb/s. Not transparent, but pretty good.

I do the above because compatibility is not a particular concern. I only play my MP3s via the HTPC in the HT, or Plex at other locations in the house.

As mentioned, if compatibility is important, use CBR and at least 192 kb/s.

[Hoser Rob]

Use CBR at the highest bit rate you can. VBR exists only because MANY YEARS AGO the devices that could play it didn't have a lot of space so VBR is a way to get more quality at lower bit rates. Lower bit rates means smaller file sizes, which means you can have more files. But now storage is cheap and big so there is no reason at all to use VBR today. If you can use CBR and 320 mbps then that is the best quality you can possibly get and VBR cannot be better than that at any setting.

+1. VBR is evil. Like packed bitstream video.

[jman98]

+1. CBR is evil. Like packed bitstream video.

I think you made a typo and mean VBR.

As far as pandy continuing to rant about how "a lot of CBR is really VBR", well, if you encode it yourself you don't have to worry about that if your encoder is good. And until he provides some evidence, this is just a bunch of pulling stuff out of his butt as far as I am concerned. And I have no idea at all what he is talking about in suggesting MPC unless maybe it's a horrible typo for AAC.

[jagabo]

And I have no idea at all what he is talking about in suggesting MPC unless maybe it's a horrible typo for AAC.

I think he means Musepack: https://en.wikipedia.org/wiki/MPC_%28audio_compression_format%29

[PuzZLeR]

For ultimate quality retention mp3 is inappropriate, i would recommend FLAC or MPC (if lossy coding is not a problem)

As per my post saying "Use CBR 320mbps for ultimate quality retention", keep in mind "ultimate" is not synonymous with "complete". (Hopefully it doesn't mislead anybody now. )

I understand MP3 isn't the best for quality retention, but can still say with confidence that MP3 at 320mbps can sound like it indeed is complete retention, at least with most sources and to most ears.

[jman98]

And I have no idea at all what he is talking about in suggesting MPC unless maybe it's a horrible typo for AAC.

I think he means Musepack: https://en.wikipedia.org/wiki/MPC_%28audio_compression_format%29

Yes, that makes perfect sense after checking out that link. Never heard of it before.

[hello_hello]

Nobody seems to have mentioned that prior to the current version of the LAME encoder (3.99) it encoded VBR and CBR differently. I think VBR has always been tweaked and developed whereas CBR was left to go the way of the Dodo, but apparently with version 3.99 it was changed and CBR and VBR now use the same method.
http://wiki.hydrogenaud.io/index.php?title=LAME#cite_note-2

When it comes to MP3 (LAME encoder) it has variable bitrate presets (V2, V1, V0). They're considered "transparent". According to the golden eared folks over at hydrogenaudio, nobody has yet been able to demonstrate the highest possible constant bitrate (320kbps) sounds more transparent.

The LAME encoder uses a "bit reservoir" so CBR MP3s are a little VBR too.

If you encode with a CBR, you can still specify different qualities. I think the LAME default is q3. Most encoder GUIs seem to set it to q2. It's slower. There's also q1 and q0. They're painfully slow. Each quality can be specified for any given constant bitrate.

I've never had a problem playing MP3 audio because it's VBR. Never. Not playing songs with an MP3 player, not when it's the audio in AVI/MKV files. I don't understand why others find it problimatic, but I haven't. I assume the "don't use VBR" advice pretty much refers to MP3 inside AVI/MKV etc? I'd not consider the possibility of thinking about CBR MP3 when encoding tracks for listening with a portable MP3 player. They're often still somewhat limited in respect to storage space so VBR makes more sense to me.

[hello_hello]

+1. VBR is evil. Like packed bitstream video.

Could you explain the way in which packed bitstream video is evil to help me understand how it also applies to VBR audio?

I don't get it. Nobody here would use a CBR when encoding video if they didn't have to because it's wasteful. Why does the same not apply to audio? Fair enough if VBR audio causes playback problems.... then sticking with CBR would be a clever idea. I've not had such playback problems myself though. Maybe I've just been very lucky?

[pandy]

Guys, technically speaking all (almost) mp3 are VBR, only encoder to meet required 320kbps adding stuffing bytes to provide constant bitrate - this same for video and audio compression - there is no such thing as half of bit.

And yes, by MPC i've referred to Musepack which is based on MPEG 1 Layer II (mp2) codec but have added lot of new features to improve quality. Why less complex codec can be better than more complex in terms of bitrate/quality? because is less complex - most of complex codec have high spectral resolution at a cost of time resolution (they need more samples in one chunk to provide higher frequency resolution - there are some workarounds to deal with reduced time resolution such as window overlapping but generally mathematics limitations are unavoidable - for those interested in details i would recommend to read about frequency domain, spectral leakage, windowing - this is unavoidable).

Anyway mp3 as a format have some serious limitations that can be only partially overcome by coding tricks - someone that going for 320kbps for sure is able to change coding to more transparent or even lossless - for others mp3 wit 224 - 256kbps is overkill for regular everyday usage. And yes VBR is more efficient method to use encoder without wasting bits. For many sources 320kbps will be filled with stuffing - this can be easily verified by MP3Packer.

[jman98]

I don't get it. Nobody here would use a CBR when encoding video if they didn't have to because it's wasteful. Why does the same not apply to audio? Fair enough if VBR audio causes playback problems.... then sticking with CBR would be a clever idea. I've not had such playback problems myself though. Maybe I've just been very lucky?

I have indeed used CBR to encode a video. I've got a small number of laserdisc to DVD jobs I've done for myself and when the source is 1 hour or less, there is really no point in using VBR video because you can't even fill a single layer disc with CBR video. Granted this is not a normal circumstance, but your post is acting like there is not even one reason why someone might want to use CBR and there is.

Nobody has ever reported a problem with CBR MP3 audio playback. Problems were reported years ago with VBR, especially when used in AVI containers. Look, the whole VBR thing was invented to solve a problem that no longer exists - storage space. If people are still married to ancient audio players with space limitations, not much I can do about that, but for people who use modern methods like streaming, VBR doesn't do anything for us because storage space isn't a problem any more.

[fritzi93]

I assume the "don't use VBR" advice pretty much refers to MP3 inside AVI/MKV etc?

Oh yeah, VBR MP3 is evil when it's in an AVI. Progressive de-sync of the audio was (is?) common. Veteran members here will remember that quite well.

[Detmek]

It seams many people claim that many devices have problem with VBR MP3. I am using VBR MP3 (even in AVI) since 2008 and I didnt have any problem with devices I have/had, DVD player, MP3 player, mobile phones. It is hard to belive that, 6 years latter there are devices that have problem playing back VBR MP3. Unless those devices are over 10-12 years old.

About storage space. Just because we have it in a lot more quantities then before it doesn't mean we should waste it. Why should I buy another HDD to store my movies and music if I can use more efficient bit/quality compression and store same amount of music and movies on 1 less HDD? Or save bandwidth and space if I upload those file on some online service.

Back on topic. Question was about music. VBR is recommended because it is more efficient. You can get same quality with lower average bitrate and save space to store more files (on portable player) or you can use higher quality settings to get better quality at same average bitrate compared to CBR mode. This does not apply for 320kbs CBR files as that is maximum bitrate for standard MP3 and you can not get better quality using VBR. But, using VBR you can get same quality at lower average bitrate.

You can also consider AAC or Vorbis if your devices support those. AAC and Vorbis are more efficient then MP3 so you can get same quality at smaller bitrate or higher quality at same bitrate compared to MP3.

For lossy codecs maximum quality is transparency. Once you are not able to tell difference between original and encoded file you can not increase quality and notice that. Transparency is individual and no one can tell you what is your limit. Only you can.

[Cornucopia]

My 2¢:

Lame is NOT the only encoder used, and some non-Lame encoders still do CBR "the old way", where it actually IS CBR. Plus, believe it or not, for many reasons (some good, some not-so-good), people still use older versions of Lame (older hardware, install freeze-out policy).

VBR audio of just about ANY format gets flaky WRT syncing when used inside a general MM container (it is at its worst when used in AVI). This is present-day.

VBR audio is often not recognized correctly WRT running time, and will have seek difficulties, in some players. This is present day.

There are reasons for using all 3 types of encoding methodologies: CBR, VBR & CQ/CFR. Anyone who tells you differently is speaking falsely. Reasons for CBR: proscribed transmission channel or equipment that expects/requires CBR, speed (it's still the quickest method, because it doesn't have to "think" as much). Plus, CBR is not "WASTEFUL" when it is intended to be using the MAX rate available to the spec.

It is NOT hard to believe that people are still having problems, nor is it hard to believe that people are still using older equipment. People don't throw away their hard-earned stuff just because you think it's old hat and outré. They change when they can afford it and when they're ready to upgrade, or when something STOPS working and needs replacement. Something that had limitations was understandable when they bought it, and it just still has those same limitations and they live with it.

Why does someone use CBR mp3? Universal compatibility. The same cannot be said (STILL) of VBR mp3. Yes, things are MUCH better, but there's still a LONG way to go.
Why does someone use 320kbps CBR mp3? Same reason (wanting ~transparent compressed audio, but also wanting compatibility).

Scott

[AlanHK]

Hello guys i'm cam and this my question

sorry my english is very limited

Many different opinions here.
And that's because you haven't said why you want to do this and how you're going to use it.

Compatibility is the main issue.

Some videos with VBR sound have sync problems, so if that's your application, CBR is safer.

If stand-alone MP3 to listen to, then VBR gives higher quality than CBR in the same filesize. Or smaller file for same quality.
As long as the player (hard or software) can handle it, use VBR. But if you don't know what player will be used, then CBR is again safer.

[hello_hello]

I don't get it. Nobody here would use a CBR when encoding video if they didn't have to because it's wasteful. Why does the same not apply to audio? Fair enough if VBR audio causes playback problems.... then sticking with CBR would be a clever idea. I've not had such playback problems myself though. Maybe I've just been very lucky?

I have indeed used CBR to encode a video. I've got a small number of laserdisc to DVD jobs I've done for myself and when the source is 1 hour or less, there is really no point in using VBR video because you can't even fill a single layer disc with CBR video. Granted this is not a normal circumstance, but your post is acting like there is not even one reason why someone might want to use CBR and there is.

Allow me to qualify my previous post.....

I don't get it. Unless they were trying to maximise the file size to fill a disc or something similar, nobody here would use a CBR when encoding video if they didn't have to.....

Nobody has ever reported a problem with CBR MP3 audio playback. Problems were reported years ago with VBR, especially when used in AVI containers. Look, the whole VBR thing was invented to solve a problem that no longer exists - storage space. If people are still married to ancient audio players with space limitations, not much I can do about that, but for people who use modern methods like streaming, VBR doesn't do anything for us because storage space isn't a problem any more.

People who use modern methods like streaming and for whom storage space isn't a problem, wouldn't be re-encoding the original, CBR audio anyway, would they? They'd be keeping the original DTS or AC3 while marvelling at their abundant storage space, or ripping their CDs to FLAC while ignoring the fact FLAC is VBR .
Those of us who re-encode audio would mostly only do so for one reason. Smaller file sizes. And not just because of storage space. Sometimes streaming isn't practical. Sometimes you need to physically transfer a file to a device. Or upload it etc. So the question becomes, which encoding method offers the best bitrate/quality ratio?

If there's a portable MP3 player on the planet which has a problem with VBR I'd be astounded, and the reality of it is even portable players produced today may still have limited storage space (4GB wouldn't be uncommon), so for that sort of use VBR makes more sense.

I'll confess the media player in the second TV in this house is used to play the occasional AVI where the audio wanders out of sync. Usually pausing and resuming fixes it, but I'm fairly sure while it doesn't happen much, it can happen whether the MP3 audio is CBR or VBR. I've never bothered to try and work out why. 99% of the time it plays MP3 audio in AVIs fine, so if there is an audio sync issue, logically it's not because it's VBR as such. Maybe the way it's muxed, the type of MP3 header used..... I don't know. The same media player never has sync problems with other containers such as MKV or MP4 regardless of the audio type (I re-encode to VBR AAC almost exclusively these days). At least none I can recall.

[Cornucopia]

Yes, FLAC is VBR. So it DTS-MA and DD+/TrueHD. So are lots of other codecs. But if the stream is encoded in a way that makes it clear to the player how to parse it and correctly organize it's time slices, it doesn't have a problem. VBR MP3 has been a problem for so many specifically because, for the first 5-7 years of its existence (the '90s), there was NO good VBR mp3 encoder, and that packet timing wasn't made clear. So players have been built for over a decade, not knowing how to deal with it correctly. That legacy is waning, but continues and has a long tail.

VBR video isn't the same thing as VBR audio, at least in terms of expectations by the player. From the very beginning of compressed video, VBR has been one of the options (I remember TrueVision MJPEG as VBR in ~'93). So players & codecs have ALWAYS been built to correctly accommodate it in its various forms. VBR audio is a very recent addition in the big timeline of technology. And when a player/editor has to accommodate BOTH VBR video AND VBR audio, it gets tricky, both for bandwidth reasons and for packet timing reasons. That's why I keep going back to: compatibility.

Scott

[hello_hello]

I have a vague recollection of having to be careful when choosing one of the first generation of portable MP3 players, to make sure it supported VBR if you had VBR MP3s. But that was also 20 years ago (in the 90's). If there's been a portable player produced in the last 10 years which doesn't play VBR MP3s correctly I'd be astounded. Especially VBR MP3s encoded today. I'd imagine the OP doesn't intend using an MP3 encoder from the 90's.

As tricky as it might be, none of the hardware players I've owned, from old AVI capable DVD players onwards, seem to have had a problem with VBR audio, MP3 or otherwise.
I'll concede though, if you're re-encoding "soundtrack" audio as MP3, odds are 99.9% of the time it's to reduce the file size, and the ultimate in fidelity isn't the most important consideration, in which case something like 128kbps CBR MP3 would be a nice safe bet.

It's a pity the OP didn't specify why he's converting audio to MP3 and whether it'll be muxed with video etc.

[mike20021969]

But that was also 20 years ago (in the 90's).

Move over Carol Vorderman... we got ourselves a maths genius here

[hello_hello]

But that was also 20 years ago (in the 90's).

Move over Carol Vorderman... we got ourselves a maths genius here

What are you on about?

[hech54]

My rule (which I rarely actually follow because almost none of my video work uses MP3):
CBR for audio in video.
VBR for plain audio files(music).
Done.

[mike20021969]

What are you on about?

It's nice to know you worked out that 20 years ago it was the 90's.

What else could it have been 20 years ago... the 70's?

[hello_hello]

What are you on about?

It's nice to know you worked out that 20 years ago it was the 90's.

What else could it have been 20 years ago... the 70's?

Let me lay out the flow of conversation for you, given you seem to have missed it.

The post I replied to:

....for the first 5-7 years of its existence (the '90s).....

My reply:

But that was also 20 years ago (in the 90's).

Sorry if my reason for emphasising the time period in question confused you, but thanks for playing.