does every codec take that longto develop?? AS i has just discover that ogg started way back in 2000. And now is 2002. WIll RC4 be the last RC before 1.0??

What do you see as 'develop'? The decoder's finished, the available encoders are very good. They could have released 1.0 much earlier, it just wouldn't have been as good. Likely, it'll be developed further for several more years. The 'roots' of Vorbis development go back some 10 years I think.

The answer to your last question is yes.

--
GCP

Well not everyone takes this long to develop a product. Microsoft puts out alot of half finished products. But they get them out there rapidly.

As long there's a decoder that works and a halfway working encoder, the codec is virtually finished.

Did you hear what the very first MP3 encoders did output? They were considered finished that time. What mainly matters right now is tuning of the psychoacoustic settings.

I think what you are seeing these days is a shift backwards in version numbering. It would have been easy to slap a 1.0 on RC1, but as Garf said, it wouldn't be as good as it could have been... it wouldn't really reflect the finished and polished status which a 1.0 should (but most often doesn't) represent.

Another good example of this is mozilla. It has been in development for what, 4 years now? And it's been quite usable for at least a year or more already. It just wasn't "finished" as in what the developers wanted for truly stable and completed release.

Anyway.. I wouldn't put too much stock in version numbers. If you need to use the program, and it's stable enough for your purposes... might as well go ahead.

Original preliminary work and research for Vorbis started in 1994. Monty didn't make it a public project until August of 1999. An audio codec isn't something that you build overnight you know. It takes great amount of time, which some people don't have as much of. It also takes a lot of effort and hard work. As Garf stated decoder is finished for 1.0 and the physcoacoustics algorithm for the encoder is very good now, but needs a little more tweaking and in time will eventually become better like Tom metioned. What more could you want?

i just can't wait for version 1 !!!!

It's on par with MPC and AAC right now IMO, so you just can expect better results.

Most psychoacoustic codec's take a LONG time to develop, and most of that time is invested into tuning. If you wonder how Microsoft and Real manage to "optimise" their codecs so well, quite often they just scan the audio codec developers and license the technology, then wrap their disgusting DRM on the top and slap their name on it.

WMA is modified Vector Quantisation (according to Ivan Dimkovic), WMV is Mpeg-4.
Real Video (latest version) is rumoured to be On2's VP4. They have used a number of technologies for thier Audio portions, like ACELP, Dolby various and Sony's ATRAC3.

I'd say that the majority of those technologies took a fair while to develop, and those adhering to ISO (like mp3) would probably take a lot longer for the paper chain to catch up.

Ivan Dimkovic writes AAC codec's for his company, PsyTEL Research. He's a formidable coder and an absolute gentleman - but really he's just doing mpeg-4 a lot better than it is done already. Not that that should be reason to play him down though he is very very good at what he does. One look at the AAC spec will show you the amount of layers and precision he has to take on.

So don't be surprised if these things take time. There are some incredibly interesting things always happening. Thats why I love this scene!

Ruairi

Well, it just seems to me that Monty has been, for some time, especially *last year* and earlier, been saying things like (ad libbing only) 'working toward a full release', and 'should be available soon', and 'making final preparations before', and many *many* other phrases that are as distinctly vague *today* as they were then.

I'm not ungrateful to Monty, not by a long shot (I've committed to vorbis completely), but he was vague *then*, and he is equally as vague now. Will he be as equally vague next year?

It also seems to me that anything without the moniker "beta" attached to it would go a long way in encouraging hardware manufactures to incorporate the format into their equipment. A final 1.0 release is just that - a version 1.0! There can be a version 1.1, 1.2, 1.2345a, or whatever else you like - just as long as we can dispense with that awful 'beta' annotation which frankly, communicates to hardware manufacturerers that vorbis is not yet ready for 'Prime Time', which I certainly believe it IS.

He isn't "Vague" Monty doesn't work on the project 100% of the time he has. Like Ruairi mentioned below Ivan Dimkovic the lead developer Pystel AAC writes the Physcoacoustics model for his research company. He always constantly working on it as opposed the amount of time Monty has. the Ogg Vorbis project first started out as a research project. Monty only started working on the project publicly in August of 99'. In fact, there is still a lot of research that goes into the project even as we speak.



That "awful beta notation" is what allows for any "last minute" changes to the physcoacoustics algorithm especially in the low level libraries need it be neseccary. By all means I am not trying to be offensive, but it doesn't matter if "you" think that Vorbis is ready for "Prime Time" neither of us have control over that. Only time will tell.

That's all good and well, but what part of 'vague' don't you understand?

Vague = eventually; whenever; however; soon; around the corner; working towards; putting final touches on; imminent; just as soon as...; eliminating last-minute bugs; upcoming; hoping/wishing/praying; fingers/toes/eyes crossed; last year, this year, next year; on schedule, proceeding as planned.......

'Beta allows for last minute changes'? How many last-minute changes have we had so far? How many more do we need, do you suppose?

(I'll give them until this time next Sunday, or demand a Full Refund !! )

(That was a joke, HotShotgg).

Yeah, I can't help it, we're all a little impatient from time to time. In truth, I think Monty & Co. have done terrific, vorbis is *sensational*, and my hopes for the format's future success is unbounded.

I just think that the elusive holy grail (1.0) would substancially help to achieve much more quickly a wholesale incorporation of vorbis into home *hardware* equipment. Frankly, for my own part, I'd like to see vorbis playback capability a standard feature of standalone DVD players. (There's a lot to be said for navigating through a menu with a hand-held remote rather than hunched over a computer with mouse in-hand). I've burned a LOT of ogg vorbis data discs since discovering ogg just a few short weeks ago. I'm not much interested in portable hand-held battery-operated flash-memory-carded 'walkmans', but rather, would like to use something other than my computer to connect directly to my home stereo, but I will be delighted of course, to see vorbis built-in to any hardware device possible, portables included.

Now.......am impatiently waiting for Mike Smith to compile a Win-32 implementation of oggenc RC4 for us!

I'd be delighted to help out myself, but I couldn't code my way out of a paper bag if I wanted too. [shrug]

I hope there will be some super-duper ASM optimized version of oggenc after the 1.0 rls (like GOGO), or something like that.

Speed is always a good thing :=)

Not really. Indeed, Psytel Research is basically Ivan Dimkovic. It's not a big research lab. In fact, I would believe all their assets are Ivan's computer and headphones.

...And his army of dedicated fans!

Ruairi

Include me There

Not dedicated... [b]obsessed!

To be honest, I have made an effort to be as vague as humanly possible when it comes to the official release date of 1.0 on public forums and on IRC. It's not because I don't know when it will be out, it's because I would rather just release it when it's done and not put myself into a position where people will make investments of time and energy based on the date I tell them.

There are a lot of people out there waiting on 1.0 to write a bunch of tools to do a lot of different things. If I give them a time and date that turns out to be wrong, there's a very good chance that they've made a time and effort investment that puts them in an endless loop of waiting for us to get our stuff done. This might be okay for small free software projects, but when it comes to business users and people that want to plan and schedule releases on a commercial level, it means the very real possibility of lost time and money. That's bad mojo when you're striving to build open multimedia intended for the masses.

A friend of mine was running a small free newsletter for a group of friends that he called 'Mcoqfl Urtma.' The title was derived by banging his hands on the keyboard. He decided that 'Mcoqfl Urtma' would be the Klingon translation of 'Expect it when you get it,' since the newsletter was published whenever he felt like it. He published and sent out the newsletters with his own time, his own equipment, and his own money. When you're building things with your own resources, you get to take that kind of license. It's the only kind of luxury that organizations like ours can afford.

I appreciate the fact that people are foaming at the mouth for a 1.0 release, and I'm right there with you; I want to see it as least as much as you do. If you like what we're doing, please make an effort to understand that we're working as hard as we can, and we're just as interested in getting it to you as you are in getting it from us.

Emmett Plant
CEO, Xiph.org Foundation

Like Emmett metentioned below this is not at a commercial level. Time and money are not being lossed here. I know we would all love to see a nice fixed point integer hardware Vorbis decoder. If you like it though all that's asked for is a little more time and then you can be sitting around a nice pool somewhere drinking a Cappucino from Starbucks decoding your Ogg Physical Bitstreams on your portable with your headphones.






Find a nice tutorial on how to program in ANSI C and download the sources from xiph.org CVS vorbis low level libraries and read over them. It's science in the making. It may eventually even spark a few ideas. No matter how you split the dice I still find it interesting. Whether your using it to program yourself or your looking over it and getting a better understanding of it.

:sign:

Sorry for this reply after Emmet, but seems necessary.

As i said here i think that better co-ordination should solve thoose discussions (that seems are supplying negative influences around Vorbis).

The questions are: why Daily Binaries services are not on Vorbis.com ?
I know that developers don't recommand beta... bla, bla...
but common peoples can watch (uManiacs comments are really interesting...) the steps of the development.

I don't see any history-like on vorbis.com, why ?


(commercial mode on)
I think that could be a cool tools for capturing more visitors on Vorbis website....
(commercial mode off)

Then, what about BBC ?

Emmett could answer that for you.



that's more of less just focusing in on specific changes made to the C and C header files. That doesn't help much in understanding the actual source. Just how progress and change accumulate over time in the source is the whole purpose for that. For instance like if you where add a function pointer to motion estimation routine in line #x in the header motion.h. It would save people a significant amount of time that way they wouldn't have to browse through the whole header file looking for the changes made to that specific line.



I don't understand what you mean? the streaming/packet layer is one of the fundemental layers of the Vorbis? Im confused.

Patchworks says:

"The questions are: why Daily Binaries services are not on Vorbis.com ?
I know that developers don't recommand beta... bla, bla... but common peoples can watch (uManiacs comments are really interesting...) the steps of the development."

"I think that could be a cool tools for capturing more visitors on Vorbis website...."

We just don't have the resources right now to give people everything they want. We're an open source project with criminally low funding, so I need to make the call as far as what's best for the project overall. Unfortunately, this means that I'm a lot more focused on releasing and publishing good open multimedia tools than running a puppet show so that we can entertain web readers.

Sure, we could be doing a lot more stuff, and we will be doing a lot more, but at the moment, it's just not feasible to do much more than we're currently doing. At the end of the day, would you rather have Vorbis 1.0, or a big sexy changelog? Stan Seibert is currently the Editor-in-Chief of Ogg Traffic, and while it might not give you the granularity of detail that you're asking for, it might help you out a little.

Emmett Plant
CEO, Xiph.org Foundation

.. i try to explain better what i mean. (smoked mode on)



The point is *not* to help in understanding the source; if you're a developer u can grab the source from CVS, this in *not* the point.
It's only a psycological approach.
My sarcastic "commercial mode" explain this better.

The open source movement is not based on financial aspects (even if isn't true... ), so (i think) that the right way to reach the "success" of a project is based on 2 things.
1. Cool and open devops(ers) should co-operate ---> At the time seems functional (even if some sectors should collaborate mutch moooooore i.e. Audio/Video, MFC/OGG...)
2. Propaganda. ---> Half functional.
Even the best platforms/os (i.e. MAC, OS/2) aren't popular.
--> You can say: this is not the target of our project !
True and false. Popularity means movement of peoples (even win users )...



...that could be translated into money !

For example:



If i remember well, BBC was (i checked their website and nothing is functional...) streaming @ 32 K and
Which is the sense of this ?


Right.
But, how can i say, peoples are demanding features (and BBC too ?) and the "Modes tuned for 32 kHz are already in CVS" are not in testing (aka binaries)
I remember a different response time of Garf's BBC tuned stuff...

Don't you think that many projects (like speex) could contribute to the maturity ?
I think that the "poor" Monty (i love you) could be helped a lot !
(...i like FLAC)


OK, but if i can find anyway the stuff online, why you "will be doing a lot more" ?
In other words: why this page is *not* on vorbis.com website (even in Japanese only) ? [ Dee - ]

I think that a sort of "centralization" could benefit even research projects: developers could be *economically* helped, so they can focus on develop.

In the end: don't you think that a better co-ordination is favorable ?


Me too !

(smoked mode off)

Patchworks says:

"The open source movement is not based on financial aspects (even if isn't true... ), so (i think) that the right way to reach the "success" of a project is based on 2 things.
1. Cool and open devops(ers) should co-operate ---> At the time seems functional (even if some sectors should collaborate mutch moooooore i.e. Audio/Video, MFC/OGG...)
2. Propaganda. ---> Half functional.
Even the best platforms/os (i.e. MAC, OS/2) aren't popular.
--> You can say: this is not the target of our project !
True and false. Popularity means movement of peoples (even win users )... "

I agree with a lot of what you're saying, and that's a big reason behind the next big Xiph.org project. We don't do nearly as much as we can do (and not nearly as much as we should be doing) when it comes to getting the word out about our projects and working with other projects that use our technologies.

We absolutely positively need to do more. We simply can't right now due to real world constraints such as the number of hours in a day.

Stay tuned, we're working on a solution.

Emmett Plant
CEO, Xiph.org Foundation

From: "Monty" <xiphmont@xiph.org>
To: <vorbis-dev@xiph.org>
Sent: Thursday, July 11, 2002 11:44 AM
Subject: [vorbis-dev] build / release engineers' todo


> The code went 'gold' tonight. Vorbis 1.0 is in CVS and it's time to
> hand off to release engineering.
>
> I've updated the vendor strind in the libvorbis code, but all other
> instances of version number need to be updated as well. I know this
> is set in a million places, for autoconf, debian, redhat, Mac and Win
> projects, etc...
>
> This is the real 1.0. Mark as such before building.
>
> Monty

just got that mail too. It's just great. Maybe we'll see some hardware in the future!

> The code went 'gold' tonight. Vorbis 1.0 is in CVS and it's time to
> hand off to release engineering.

Bugger me.

(that's an expression, not a request)

I'm still going to wait until it happens - just in case it's a faked email.

a MSVC compile is already available on case's site:
http://www.saunalahti.fi/~cse/oggenc.zip

and on the vorbis IRC Channel its too!

Yes, it does seem very suspicious, doesn't it? But I did some looking around in the xiph.org CVS archives and found the following:

So it seems like it's for real... Either that or some prankster cracked into Monty's CVS dev account, but that seems very unlikely.

Yeah,

the new code is unbelievable ... Just did some test to look what ogg vorbis sounds with "worst" quality 0 ... I´ve never heard a better 64k encoded file, congratulations to the ogg vorbis team!

cu, pop3smtp23.

Wait...

I got auto-generated binaries from 5th July which are tagged pre-RC4, how can there be a 2nd July 'final'? Did I miss something?

So we get through RC4 and directly go forward to 1.0 release (pre maby)?

What's new in the code level? Can some good man who have studied the latest CVS give us a detailed description from a technical view?

[17:11] * Now talking in #vorbis
[17:11] * Topic is 'All 1.0 modes and tuning are now in CVS. Yes, the real 1.0. Test them!'
[17:11] * Set by xiphmont on Thu Jul 11 12:22:33

If you're referring to the date "02/07/11" then the 02 stands for the year 2002 and the 11 stands for the eleventh day of July.

Get out your black suits. It's time to have a funeral for Real, WMA and mp3pro.

Emmett Plant
CEO, Xiph.org Foundation

I hate to be a cynic but even the latest ogg binaries arent whipping WMA/Real at dial up bitrates. Yes, the high bitrates are damn cool and beating the majority of formats I'm just a tiny bit peeved at the very low bitrate performance. I'm sure within a month or two it'll be competitive. The 48kbps (44.1khz/stereo) experimental mode is excellent, but if they can push it one further by having comparable performance at 32kbps, then it'll be formidable.

*EDIT: WM8 @ 32kbps Stereo is at 32khz. Vorbis can do this.... wait a minute... im not sure it can... investigating...

Also bear in mind very very low bitrate audio is often muxed for streaming video, and with vorbis only performing respectfully at around 16kbps+, its a bit harsh for the dial up video users (for when the time comes around).

Don't worry, I havent lost faith. Vorbis is a wonderful format and I'm extremely impressed with the progress.

Ruairi

I agree with you - but it's a great start.

Put this together with the future developments in VP3/Theora, and in about a years time we'll have a streaming video *and* a DVD-ripping choice with no patent or licencing worries. Given how close we came to a DRM-only world, I think that's quite an achievement.

£10 will be winging its way to Xiph from be as soon as version 1 and the specs officially appear. It would be more, but I have to eat

I sure hope so. Any news on this front lately?

Ogg Vorbis Pre-RC4 already got ranking #1 in a review of audio encoders by a well known German magazine called 'Chip'


Chip Magazine (German language)

Although I do not take their tests very serious at all (measuring filesizes in CBR mode ??? ) , at least this will contribute a little to the popularity of the format.

and they link to my page

With my compiles, I believe?

sure:)

you can't image the traffic which was produced today

I probably can!! When I was trying to get the binaries out, I had ICQ conversations running with Roberto, RC55 and Emmett! And, I was trying to upload/mail at the same time!

RareWares is serving files like mad.

Hope Paul can handle all the bandwidth stress now that Ruairi left the boat.

Thanks a lot, Paul!

50MB today, no problems that I know of! Interestingly my log includes two hits from a Solaris machine, and one from Win3.1

Cheers, Paul

I just can't wait for the net radio stations to stream to ogg 64k:D

That would suck, but 96kb/s is good enough for me, 192 would own

Would I be right to assume that the only changes since the 20020704 library (pre-RC4) are bug fixes? Is it worthwhile to reencode pre-RC4 tracks with this final, stable 1.0 library?

Well, I would believe now is the right time for some serious listening tests.

@mithrandir



No, you'd be wrong! Numerous changes have been made. I would suggest re-encoding would be a worthwhile exercise, but you may like to test a few tracks before undertaking a wholesale re-encoding!