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: Informed rebuttal of VBR disadvantages appreciated (Read 42490 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Informed rebuttal of VBR disadvantages appreciated

Reply #50
I actually know Pulse personally and he's not nearly as much of an ass as he seems to be on the forums.  Unfortunately for us, we get a lot of people who don't read the rules (it's Pioneer's forum and it's their rules, some of which people don't like), so Pulse deals with a bit of a heavy hand. 

Anyway, back on topic here...

Quote
About 80% are my own CD collection. The rest is MP3 mixes by different DJs, friends, my own.
MP3gain can tell you if the file has an issue (it won't be able to process it).


Excellent.  Thanks!  I've heard of MP3gain but never used it; I'll check it out.  Obviously since you did your own encoding, you wouldn't have come across the types of files I'm referring to (which I've been told via a nice PM that I can't do; sorry for the violation and I hope this isn't also crossing the boundary) which tend to be broken.  But trust me when I say, they're abundant.

@xmixahlx - excellent info, thanks for that.  Nice to see some concrete numbers.

Quote
You're mixing things in a way that it piles up according to your expectations and without any kind of sense. "Eyes blur, the ears (...) less" has nothing to do with VBR. Why do you think that?


You are correct sir.  I was trying to justify it and ended up deleting it because I clearly had not thought it all the way through.  I gotta stop doing that.

Quote
I was (possibly falsely) making the assumption that an audio CD would already be at least partially buffered to memory by such a player to allow tricks in real-time (you can't physically play a CD backwards... can you?) and that the MP3 data would be using the same buffer in the same way once decoded to an equivalent linear digital format and still using the same "trick" hardware/DSPs.


It's no secret that the CDJ's use a buffer to be able to manipulate the audio.  I actually tried one time to make a CD play backwards in a player (I turned the motor around - and it WASN'T a CDJ that I paid $1200 bucks for) - it didn't work.  LOL.  As to what digital magic is happening internal to the player, that's a trade secret, I'd assume.  There may be a valid reason for why VBR is not as supported as CBR - but I'm not privy to it.

Pioneer was one of the last hardware DJ companies to support MP3.  It was a major point of contention for a lot of people for a long time.  We never found out why it took them so long, but eventually support was released.  Perhaps the reason they don't support VBR "properly" is similar to the argument around Wavelab - Pio's products are aimed at "professionals" and for a long time it was assumed that they wouldn't use MP3's.  I'm not sure I agree with that, and it's speculation anyway, but it's an easy speculation to have I'd say.  The other possible reason it's not supported properly is because their support for it is so new, it's suffering from the same problems reported by many VBR implementations - lack of seek, difficulty reading it correctly, etc. etc.  Again, just speculation, and IMHO if VBR is as good as you're claiming (which I'm starting to believe), then there's no excuse for the lack of support.  Then again, the same could be said for their initial lack of support for MP3's at all.

Quote
The encoder is "starving" bits in complex parts and "wastes" bits in easy to encode parts. VBR simply aims to maintain a constant level of quality throughout the encoding.


I'd say that's true of VBR encoder settings that use the full range vs. CBR settings below 320kbps; it's NOT true of 320kbps, which is the max anyway and is what I use.  While at 320 it might be "wasting bits" on the uncomplex parts, the same could be argued for an uncompressed file - do we REALLY need all those bits to describe some minutia of the sound that we probably can't distinguish anyway?  Isn't the sonic quality presented by MP3 "good enough" for "most people"?  I'm sure you'd say no, but then we're right back to the same argument.

Quote
PS Apologies to DJRyanJ for the dig earlier. I got the impression that you weren't prepared to listen to logic or partake in reasoned debate. How wrong I was


Don't worry about it.  I can see how easy it is to make these kinds of assumptions; thanks for the warm welcome.

-r-

Informed rebuttal of VBR disadvantages appreciated

Reply #51
I'd say that's true of VBR encoder settings that use the full range vs. CBR settings below 320kbps; it's NOT true of 320kbps, which is the max anyway and is what I use.  While at 320 it might be "wasting bits" on the uncomplex parts, the same could be argued for an uncompressed file - do we REALLY need all those bits to describe some minutia of the sound that we probably can't distinguish anyway?  Isn't the sonic quality presented by MP3 "good enough" for "most people"?  I'm sure you'd say no, but then we're right back to the same argument.


You are right that a VBR mp3 file, when compared to a 320kbps CBR mp3 file, is not true 320kbps.  That doesn't mean anything though as you are kind of arguing against yourself here.  320kbps is wasting bits on most samples for most people and you are somewhat right about uncompressed audio.  There aren't many people out there that benefit from uncompressed audio.  That is why we have lossless though.  Lossless formats such as FLAC and Apple Lossless are able to shave off 10MB or more from the file size (when compared to uncompressed WAV) while retaining track tag information and still being bit-for-bit identical.  The main reason behind having lossless/uncompressed audio is so that someone can have a perfect backup of their music.  They can then use only lossless files or use a lossy format.  You are right that mp3 is normally good enough for most people.  However, mp3 is normally good enough for most people at much lower bitrates.  320kbps is a "safe bet" but you don't need that high of a bitrate when encoding audio even for DJ purposes.  The environments where DJs often play are not really optimal for music.  The rooms are often not acoustically tuned, the positioning of the speakers are not optimal (studies have shown that even with small speakers, moving them by 1 cm can change the perceived quality), and people absorb and reflect the sound waves while adding to additional noise influences by talking and whatnot.  That is why one doesn't really need 320kbps.

So go ahead and argue that 320kbps is a safe bet (or that you need it as your ABX tests have shown that).  However, I wouldn't argue that that uncompressed audio uses more bits that needed and make that look bad and then come back saying that 320kbps uses more bits than needed and that is alright.

Edit:  It is nice to see someone who will listen to logic and not come off as a butt hole.  I am sure that Pulse isn't like that in real life and I know that maintaining a forum can add to daily headaches.  However, he was coming off as someone who refused to listen to logic and would just ban people who disagreed with them.

Informed rebuttal of VBR disadvantages appreciated

Reply #52
On the switch subject, I just would like to add that the numbers should be inverted. That is, V1 to V9 (max. quality). I never understood why in the world "zero" was the best quality. The way it is it's simple, but quite confusing for newbies of course. Apologies that I went offtopic with others, but I believe that was a valid claim.

Informed rebuttal of VBR disadvantages appreciated

Reply #53
320kbps is a "safe bet" but you don't need that high of a bitrate when encoding audio even for DJ purposes.  The environments where DJs often play are not really optimal for music.


Generally, you'll get no argument from me and I said as much in my first post.  What I DO argue however is that while DJ environments are never (or rarely optimal), I don't want to be the limiting factor, so I play the highest-quality music I can given the tools I use.

Quote
So go ahead and argue that 320kbps is a safe bet (or that you need it as your ABX tests have shown that).  However, I wouldn't argue that that uncompressed audio uses more bits that needed and make that look bad and then come back saying that 320kbps uses more bits than needed and that is alright.


I would never say such a thing!  LOL and if I did, then that's not what I meant.  I was arguing that SOME PEOPLE might say that uncompressed is pointless as it uses more bits for the same thing that people can't hear anyway.  I would rather use uncompressed; but the realities of what I do make that impossible (for now).

-r-


Informed rebuttal of VBR disadvantages appreciated

Reply #55
What I DO argue however is that while DJ environments are never (or rarely optimal), I don't want to be the limiting factor, so I play the highest-quality music I can given the tools I use.
On that basis, if you found out that that you couldn't tell the difference between a CD quality source and, say, the ~128Kbps LAME VBR in our listening tests, would you then reconsider your need for 320Kbps CBR?

It's difficult to say exactly what average bitrate is required in VBR to outstrip the acoustic requirements of the venues you play in and the speakers you use, but I think it highly unlikely that you need a constant 320Kbps in reality when VBR at around 175Kbps suffices on high-end home equipment through high-end headphones for many of us. Yes, even the fussy ones. LOL

As you say though, whether or not you can use such files with any degree of reliability will be determined by the player you use. Hopefully either yourself or Pulse will try some LAME generated VBR files on your Pioneer players to find out if the VBR problem really is a global one or if it's just limited to badly written or broken VBR files.

Cheers, Slipstreem. 

Informed rebuttal of VBR disadvantages appreciated

Reply #56
Generally, you'll get no argument from me and I said as much in my first post.  What I DO argue however is that while DJ environments are never (or rarely optimal), I don't want to be the limiting factor, so I play the highest-quality music I can given the tools I use.


Even if you use lower bitrate music (128kbps-160kbps), you won't be the limiting factor.  There are far too many factors that degrade the perceived sound quality of music let alone in a club/dance hall/whatever.  Also don't forget that you are playing music for an audience where the majority of people are fine with 128kbps WMA/AAC music from legal online stores.  Just know that the quality of music that the DJ uses if often the very last limiting factor amongst thousands of other factors.

I would never say such a thing!  LOL and if I did, then that's not what I meant.  I was arguing that SOME PEOPLE might say that uncompressed is pointless as it uses more bits for the same thing that people can't hear anyway.  I would rather use uncompressed; but the realities of what I do make that impossible (for now).

-r-


I must have misinterpreted your post.  I took the comment of "While at 320 it might be "wasting bits" on the uncomplex parts, the same could be argued for an uncompressed file - do we REALLY need all those bits to describe some minutia of the sound that we probably can't distinguish anyway?" the wrong way.  I took it as you saying that we have all this uncompressed music that we don't really need and that is bad (it wastes bits) but you will go ahead and use 320kbps which wastes buts but that is alright.  The use of uncompressed audio simply is not needed in a day where we have lossless codecs and high performing lossy codecs that can achieve transparency at such low bitrates (128kbps).  I suggest that you conduct a few blind ABX tests with your best set of headphones and Lame 3.98.2.  I think that will help you further understand what we are trying to say.

As Slipstreem said, most people are fine with Lame's -V 3 setting and this includes the "picky" people.  Lame has changed a lot over the years and the perception of "320kbps or bust" is no longer needed.  I had a roommate that continued to use MusicMatch for all their ripping needs and they would use 320kbps (MusicMatch uses the FhG mp3 encoder).  They said that anything below 320kbps CBR sounded like "crap" especially when using their stereo system.  I sat them down and had them conduct a blind ABX test using their headphones, using my theater system (which is much better than theirs), and then using their stereo system.  They failed to properly distinguish between a -V 5 encoded file and the source lossless file every single time.  Even after that, they refused to believe that I was using -V 5 files and still continued to rip with MusicMatch at 320kbps.

[off-topic]Happy 1000th post! [/off-topic]


I didn't even notice that.  Thank you!

Informed rebuttal of VBR disadvantages appreciated

Reply #57
On that basis, if you found out that that you couldn't tell the difference between a CD quality source and, say, the ~128Kbps LAME VBR in our listening tests, would you then reconsider your need for 320Kbps CBR?


I'm not ready to do that just yet, for a couple of reasons:

1) I occasionally remix the tracks I use, and while I'm no pro remixer, I try to maintain quality where I can.  When an uncompressed version is not available, I must use my MP3 copy, and in that respect having the highest possible original copy is best.

2) I don't have a problem with the amount of space that a 320kbps CBR takes vs. a VBR.  I've accepted the amount of space that a 320kbps CBR takes and I'm OK with it; if we can all agree that the point of MP3 is to save space, then likely everyone has a different level that they're willing to accept in terms of saved space.  Mine is currently at 320kbps.

3) I don't totally trust my ears.  That sounds ridiculous given the topic at hand, but just because I can't tell the difference doesn't mean someone else can't (or won't be able to in the future - my music collection is also designed to last a long time).

Given those reasons, if someone could conclusively prove that VBR not only saved space over a 320kbps CBR but ALSO provided CONSISTENTLY better sounding files, then I'd be the first to switch over, as long as the program I use to play back with supported them.

Even if you use lower bitrate music (128kbps-160kbps), you won't be the limiting factor.  There are far too many factors that degrade the perceived sound quality of music let alone in a club/dance hall/whatever.  Also don't forget that you are playing music for an audience where the majority of people are fine with 128kbps WMA/AAC music from legal online stores.  Just know that the quality of music that the DJ uses if often the very last limiting factor amongst thousands of other factors.


I know this is true; and you forgot something: the sheer drunkenness of the crowd  .  But just because the general public accepts that level of quality doesn't mean that I should.  In my opinion this is one of the fundamental wrongs with the online music idea (and, in some ways, society in general - but this is hardly the forum for that  ); few seem to care about quality and are willing to accept things as being "good enough".

The use of uncompressed audio simply is not needed in a day where we have lossless codecs and high performing lossy codecs that can achieve transparency at such low bitrates (128kbps). ... I think that will help you further understand what we are trying to say.


This may be true.  And I do understand what you're trying to say; but until the day when we see wide-spread adoption of these lossless codecs as well as a wide-spread knowledge of the way to PROPERLY encode things with those high performance lossy codecs, then I think we may be stuck with the "evils" of uncompressed audio 

As for me, as mentioned above, while I may not be able to tell, and you may not be able to tell, and VBR may be good enough for me, I'm just not quite ready to accept it.  However, that in mind, I will continue to experiment, monitor, and otherwise expand my mind and opinions.

I will not, however, be spreading lies about VBR any longer, and will in fact promote them to appropriate people.

-r-

Informed rebuttal of VBR disadvantages appreciated

Reply #58
I'm not ready to do that just yet, for a couple of reasons:

1) I occasionally remix the tracks I use, and while I'm no pro remixer, I try to maintain quality where I can.  When an uncompressed version is not available, I must use my MP3 copy, and in that respect having the highest possible original copy is best.

2) I don't have a problem with the amount of space that a 320kbps CBR takes vs. a VBR.  I've accepted the amount of space that a 320kbps CBR takes and I'm OK with it; if we can all agree that the point of MP3 is to save space, then likely everyone has a different level that they're willing to accept in terms of saved space.  Mine is currently at 320kbps.

3) I don't totally trust my ears.  That sounds ridiculous given the topic at hand, but just because I can't tell the difference doesn't mean someone else can't (or won't be able to in the future - my music collection is also designed to last a long time).
Some extremely valid points.

This topic seems to have meandered quite a bit, and it's been interesting in part.

I would like to point out that personally I have no issue with someone choosing to use 320kbps CBR.  I do have an issue with people using FUD to stop people using VBR, or lower bitrate files, when they may actually be completely suitable for them, and more ideal than 320kbps CBR.

In essence, the software cannot handle VBR, and that is a failing of the software.  Trying to negate that failing by saying that people who use VBR or sub-320kbps CBR files are fools is just really, really, wrong.  Pulse should stick to what he knows.
I'm on a horse.

Informed rebuttal of VBR disadvantages appreciated

Reply #59
Well, I'm sorry but after finding out about Pioneer's VBR limitation my frustration with apparently flawed development processes of huge companies got the best of me (again). People on my blog keep reminding me of workarounds (most of which I'm already aware of) but it's the fact that I need a workaround (like with my Sony) that's ticking me off in the first place.

So when I saw the way the argument goes on the Pioneer forums I wanted someone with knowlegde about the technical issues set the record straight. My opinion is this:

a) a full-fledged MP3 player has got to be able to work with VBR or be renamed
b) properly encoded VBRs are in no way inferior to CBR

Since Pulse as a regular seemed to be leading the discussion on VBR over there with opinions rather than facts I was afraid some Pioneer developers might take this for granted and shun the idea of properly implementing VBR support in future hardware.

It appears to me that Pioneer has gotten it all wrong. It's similar to what's going on at the moment with DVD standlones and USB support. Since USB-HDDs use the same protocol as USB sticks those players can playback media from HDD as well. But some 2.5" HDDs draw more power from the slot than the 500 mA specified so they can't gurantee them to work which simply makes them say it's not possible. Pioneer's hardware does playback VBR but it's officially not supported, you cannot seek in those files and you don't get the remaining time (both vital for a DJ).

So I would expect them to properly implement support for VBR while at the same time making it clear in the manual that this applies only to properly encoded files (proper seek index and stuff). Of course one would have to check playback of each file prior to public performance which I would do even for my own encodes.

Informed rebuttal of VBR disadvantages appreciated

Reply #60
With reference to your last paragraph, have you tried playing a file generated with LAME in VBR with just the -V switch? To the best of my knowledge, this should produce an ISO/IEC compliant file. Maybe a LAME developer could verify that for me?

Cheers, Slipstreem. 

Informed rebuttal of VBR disadvantages appreciated

Reply #61
I'm gonna try em tonight. I've also converted some files I'd encoded from my own CDs on --preset standard with mp3packer on AUTO bitrate resulting in all files blown up to 320 kbit/s CBR. Is there a way to encode to the max frame used in the original VBR?

Informed rebuttal of VBR disadvantages appreciated

Reply #62
Well, I'm sorry but ...
I can't work out whether this is in response to my post.

As I say, I totally agree that the Pioneer should be fixed, so that it can use VBR properly.

Unfortunately, with mods like Pulse in that forum, I don't hold out much hope for you changing anyone's opinion, as he is likely to just respond to any post calling it into question with some crap that looks authoritative to other less-knowledgeable members.

Edit: Oh, and I thought mp3packer did just use the highest bitrate in the VBR file, not necessarily 320kbps.  I've only played with it though.
I'm on a horse.

Informed rebuttal of VBR disadvantages appreciated

Reply #63
If even a single frame uses 320 kbps then the entire file must, after repacking. The only exception would be if more efficient use of the bit reservoir allowed a 256 kbps frame to be used.

Informed rebuttal of VBR disadvantages appreciated

Reply #64
I can't work out whether this is in response to my post.

Nope, I wasn't addressing anyone in particular, just felt the need because I started all this fuss

If even a single frame uses 320 kbps then the entire file must, after repacking.

D'oh, of course 

Still, some tracks increased by 53% which is a lot. I've also tried LAME with the -V option and it doesn't make a difference to Pioneer: you can't seek and don't have the remaining time.

BTW, this manual entry could easily be adjusted to a VBR disclaimer for future models with proper VBR support:
Quote
MULTI READ
Supports playback of CD-R and CD-RW discs. (Some discs may not replay properly, however, due to certain special characteristics of some discs and recorders, as well as due to dirty or damaged discs.)

Informed rebuttal of VBR disadvantages appreciated

Reply #65
a) a full-fledged MP3 player has got to be able to work with VBR or be renamed


I don't entirely agree, at least in this context, because:

The Pioneer CDJ decks are NOT MP3 players anyway - they just happen to play MP3's AS WELL AS normal audio CD's.  That's not what they were truly designed for, it's just an added feature.

Since Pulse as a regular seemed to be leading the discussion on VBR over there with opinions rather than facts...


I think that Pulse's opinions were valid for what he knew.  While I agree with:
Quote
Pulse should stick to what he knows.

He did know what he was talking about, even if it was outdated.  I'm sure we can all agree that realistically, one bad experience can be enough to turn us off of something for a long while to come.  He's not perfect, just like any of us, so his "opinions" were "facts" - in my opinion (lol); they were just outdated.

Unfortunately, with mods like Pulse in that forum, I don't hold out much hope for you changing anyone's opinion...


I suggest you go and take a look at the current discussion.  Pulse has admitted that he has much to learn and is willing.

-r-

Informed rebuttal of VBR disadvantages appreciated

Reply #66
I suggest you go and take a look at the current discussion.  Pulse has admitted that he has much to learn and is willing.
All he needs to do now is actually act on it.  In his last post he still shows complete contempt for 128kbps MP3 files, and the MP3 format in general.

I think the roles played by yourself and the other moderator may have had a large influence in the "humility" of his later response.  We'll see.

At least djjay has suggested that the question will be raised with Pioneer; so I was, gladly, wrong on that count.
I'm on a horse.

Informed rebuttal of VBR disadvantages appreciated

Reply #67
Lossless formats such as FLAC and Apple Lossless are able to shave off 10MB or more from the file size (when compared to uncompressed WAV) while retaining track tag information and still being bit-for-bit identical.

Lossless compress to about 60% of a WAV size on average (iirc). Some things compress to 50% (half the size of the original WAV).

Informed rebuttal of VBR disadvantages appreciated

Reply #68

a) a full-fledged MP3 player has got to be able to work with VBR or be renamed

I don't entirely agree, at least in this context, because:

The Pioneer CDJ decks are NOT MP3 players anyway - they just happen to play MP3's AS WELL AS normal audio CD's.  That's not what they were truly designed for, it's just an added feature.

That's not how they're advertized (in Germany anyway) which is why I stand by my opinion. Roughly translated:

Quote
The ProDJ CD-Deck combines key elements of the Pioneer CDJ-Players with a host of brandnew functions like a full-fledged MP3-Player, Hot Loop, Beat Loop und our unique Loop Cutter...


The english page states:
Quote
Plays MP3s
* MPEG-1: Supports Audio Layer-3 sampling frequency 32kHz, 44.1 kHz, 48 kHz, Bitrate 32 Kbps ~ 320 Kbps.

Informed rebuttal of VBR disadvantages appreciated

Reply #69
Dj's of the world please stop bothering us with FUD because your decoders are BROKEN .

Informed rebuttal of VBR disadvantages appreciated

Reply #70
I'm trying to do some comparisons, and I'm looking for castanets.wav as it seems to be the one to test with.

I've searched this forum and come up with many links, unfortunately most of them are either down or have dead files.

Can anyone provide me with a working link to castanets.wav?  And if you're being generous, any other sound samples I should test things out with?

Thanks.

-r-

Informed rebuttal of VBR disadvantages appreciated

Reply #71
http://ff123.net/samples.html is a good place to start. It refers to the LAME problem samples page for castanets.wav, which no longer exists. However, castanets.wav exists at http://ccrma-www.stanford.edu/~bosse/ apparently.

Castanets is a good sample because it provides a clear example of pre-echo. Other killer samples for pre-echo which may be familiar to you as a DJ are Kalifornia by Fatboy Slim and Four Ton Mantis by Amon Tobin. Specifically, listen to the percussion in Four Ton Mantis and the distorted vocal sample in Kalifornia.

Informed rebuttal of VBR disadvantages appreciated

Reply #72
On problemSamples folder you'll find those samples I care about, including castanets.
lame3995o -Q1.7 --lowpass 17

Informed rebuttal of VBR disadvantages appreciated

Reply #73
However, castanets.wav exists at http://ccrma-www.stanford.edu/~bosse/ apparently.


Actually, castanets.wav is listed there but when I try to download it I get a 345 byte file.  Definitely NOT the proper file .

@halb27: Thanks for the link.

I'll search out those other tracks; I haven't heard them personally but that's alright.  I'm going to search my collection for some other "problem tracks" as well and see what I can come up with.

I'm planning a rather epic post over at the Pioneer forums with my own sets of data, samples, and trials to try to clear up any FUD that has been spread, as well as to (hopefully) kick Pioneer's ass a little bit in terms of fixing their VBR support (if that's what the data shows, which I'm sure it will).

Once it's posted I'll link to it here.

This may be a stupid question, but since it's relevant I'll ask it anyway: if I zip (or rar) the samples purely for the purposes of having a single file to download for the members (which I'm sure will include whatever mild amount of compression WinRAR can achieve), will this affect the quality of the unzipped/rared samples?  I assume no, because encoding something like a binary would necessitate lossless compression.  However, I just want to make sure I don't shoot myself in the foot before I start.

Thanks.

-r-