For those who are not subscribed to the MPC mailing lists, here my email from this morning :


I want to give a short summary of this long telephone conversation, here
the topics we discussed :


1. Subband coding vs. Transformation coding
2. Future of MPC
3. MPC in matroska
4. DVD players and their problems


1. I was very surprised to understand that Frank doesnt see any benefit
in musepack's subband coding, compared to modern transformation codecs
like AAC or even Vorbis. He told me that musepack's results in the 128
kbps listening test are a slap in the face of the other codecs, as
normally musepack should have performed worst at this bitrate. He is of
the opinion that musepack's only advantage is the psy model he made, or
the lack of there-of in the other codecs.
Not even in the decoding speed does he see a big advantage in subband
coding, he is convinced all of the other decoders could be optimized a
lot, especially with improvements in the lookup of the huffman tables,
with a proper indexing system as he has done with musepack once, gaining
a speed advantage of factor 9 compared to buschel's code.
He mentioned more than one time that MP3, from his point of view, is not
at all well specified and implemented and can even lead to drop-outs
during playback under certain conditions. He sees AC3 as a very good
standard for his needs, because of well existing hardware decoders in
external receivers, a proper and well done specification and
implementation, and because with DVD burners beoming more and more
popular, low bitrates for audio compression will be only interesting for
streaming in future, and he doesnt see the big market for streaming at
all. He told me he is maybe interested in making a proper AC3 encoder,
using his own psy model, that in principal can be transferred from one
encoder to another.


2. It seems that Frank is more and more into video, especially DVD
video, and my personal opinion is that musepack is not in the center of
his interest anymore. His opinion on 1. surprised me a lot, i didnt
understand why he had invested so much time into musepack when subband
coding has no advantages compared to other coding technologies, but as
so many other technicians i guess it was more like a prrof of concept
for him, after the LAME people had more or less pushed him out of the
team because of the many radical changes he had in mind.
After learning that, i honestly question the future of musepack in its
current form. It was maybe a better idea to make a good, free and
opensource AAC encoder, or to fork from Vorbis and making a new
compression format, if Frank ever finds the motivation again to invest
serious work into making a new audio codec. The other alternative is,
and that is what Frank obviously had in mind for musepack also, make a
clear cut with subband coding for the future and make SV8 a
transformation codec also. Frank told me that in his opinion the AAC
standard has a couple of quirks in it, and we could maybe make a new
compression format, still call it musepack and MPC, overcoming those
problems, but on the other hand staying decoder compatible with AAC ??

It appears to me that Frank alone can not invest the time necessary to
realize MPC SV8. On the other hand he could very well make the specs,
and invest time into his main area of excellence, the psy model. He
mentioned to me that he had invested too much time already into
explaining what has to be done to other people, but without any real
feedback from them after that. I offered to him to try to use the
existing GForge/Alexandria facilities on the corecodec.org server for
that, allowing you to assign tasks, explain them one time, and then
let me search for a developer who wanted to care about that. I am not
sure if i could convince him that this is a sensible thing to do, but he
promised to make a last attempt here and compile a list of things where
he could use help from outside. I will take this list and define taks on
corecodec.org for each of them. Hopefully that way we can organize
things a bit better in future.


3. Frank has doubts if our developers, mainly Toff, have really found a
way to read a complete block of data from a SV7 file for muxing into
matroska. He says the current bitstream is weired, there are sometimes
bits from one block attached to the next or to the preceding block, so
without calling the decoder on the blocks there is no way of finding
out. Even then, proper seeking might be impossible because in worst case
the decoding of a block can depend on up to 32 blocks before this block.
He promised to look into a repacking app, reading SV7, unpack the data
and repack them into SV8 bitstream, forming a kind of SV 7.5 version if
you want, as i convinced him there is no use of allowing subband coding
in SV8 at all, if he feels it doesnt have any advantages to
transformation encoding. With this app we could pack SV7 files into
matroska fine, just a new decoder is necessary that can handle the new
byte order as specified for SV8 ( big endian ).


4. Frank has a lot of ideas about improving currently existing DVD
playback chains, like by implementing the video decoder chip into the
TV, transferring the compressed video data from the DVD player to the TV
via firewire IEE 1394 . He also believes it should be possible to delay
the picture by several milliseconds, to allow the use of digitally phase
corrected speakers, introducing a delay by deafult.


Conclusion :

MPC is not dead, not at all. Even in its current form, its a great
encoder for movie soundtracks if highest quality for Stereo tracks is
the aim, and Frank promised to look into a way to be able to get clear
block boundaries for SV7 streams. The huge fangroup on HA.org, all music
lovers compressing their music albums with it, are a good prrof for
that. However, it seems that MPC only has a future in a complete new
form, and the man who could bring the format there doesnt have the time
to make this alone. IMHO its time to bring the project to corecodec.org
now and use the existing opensource facilities more evident. The way
things did work out until today, one man doing all the core things and
the other devs implementing these core elements into several apps, will
not work out for the 'new' MPC SV8 anymore.

Its time for us to ask ourselves if we want musepack to evolve further,
or if we are happy with what it is today. Whilst it is usable for music
compression already, its not for use with video. If we want musepack to
progress, we all have to work together to improve it. Looking forward to
hear your comments.

Christian
MPC and matroska project admin

http://mpc.corecodec.org

If development or just maintenance isn't carried on by any other than Frank the future for MPC doesn't look all too bright to me judging from that conversation.

Here lies MPC.
The best damn audio codec that almost was.
RIP

Amen!

New Musepack based on AAC is an interesting idea.

I don't think MPC is dead at all. If it does the job nicely, why change?

Ruairi

Why R.I.P.? Did you read that post at all? I think this is an opportunity for MPC to pick up steam again. But i'd also like to see a clear bottom line for SV7, something like 1.2 final, with hardcoded --xlevel and things like that. We know the psymodel almost reached perfection, so there's no big effort for him coding-wise to tie up the loose ends and bring SV7 to a closure. Then it could go on on corecodec, soundforge or whatever, when the further goals are clearly defined.

As said in the mailing list I'd love to see a unified bitstream that supports both subband data (which can also be losslessly transferred from SV7) and MDCT data. Then mppenc can produce both types of data, which sets a balance between speed and space savings.

Edit: Of course, in both modes the psymodel would be the same (and very up-to-date) module.

Frank said the MDCT data would reach the quality of the subband mode - but still, there's speed and compatibility. I think both modes are useful.

Also for muxing purposes, I think a unified, flexible, clean format is very important. Just think about mp3 and the numerous timing / gap / relaxed standard issues.. also the presence of two Musepack formats (ie: SV7.5 and SV8) wouldn't be good for muxing applications (we already have too many file formats IMHO).

Let's make a single format, where the latency (in microseconds or samples) and other properties are written in the audio stream, so the muxer can process subband and MDCT data transparently and easily, and produce a great Matroska stream with none of the stupid limitations or glitches of AVI and MP3 !

I'm looking forward to the time when two lossy video files can be joined gaplessly with the audio... (might already be possible with Vorbis though - I don't know).

Just my two cents

I second that desire to tie up SV7 with a final 1.2 version. While there might not be much difference between that and the current beta 1.14 (at least not enough to merit a reripping), I think it would be kind of untidy to leave things at "beta" stage. MPC is anything but dead, but I think it has reached the end of a particularly important stage that would deserve a definitive version that represents the summit of this phase.

I must agree (since I use 1.14 all the time) but do you think it would be worth the effort (from Frank's POV) to make an additional release within the SV7 limitations ?

If he really plans to make pre-SV8 final in couple of weeks, then everything created with 1.20 could be converted to that better format anyway.. which doesn't have scalefactors issues and such.

Apart from defaulted --xlevel and a more "final" stage, what improvements do you mean ?

Regards

Yup. Read every word. I'll believe it when I see it.

Edit: By the time anything significant happens with MPC, I will have a 1TB flash card in my portable, and it will be irelevant. Actually, now that I think about it, Vorbis is starting to head down that same path.

Yes, the "if" is the factor here. If a some sort of post-SV7 implementation makes good progress soon, that's excellent, but it will still take some time and beta-testing to have a truly improved alternative to SV7.



I'm thinking more of the frontend. All custom switches should be disallowed, only --quality should be allowed. --xlevel should be hardcoded, remaining --xlevel clipping warnings should be suppressed. Maybe Frank has more ideas, maybe there are also some bugs left to fix. This way we would have a finalized, simplified version that we could recommend to every newbie without concern and much further explanations as we have them now.

Well, that could depend on the time that he would be willing to invest; but, for instance, simply incorporating the improvements over 1.14 that 1.15r produced would make a good tie-up, IMO.

I am just waiting for Roberto to show up in this thread and add his 2 cents to the discussion

Then I suggest you keep your pessimistic views to yourself. Mpc doesn't need any downplaying, especially at this kind of juncture .

I don't see why that's such bad news even if MPC development is complete. It's already extremely well tuned as it is. I think SV8 would have been a better point to stop though but maybe some other developers could complete the to-do list?

I thought i was clearly stating that MPC is NOT dead, not at all !

It just came to the point of no-return, a lot has to change now or its going into an uncertain future, thats it. As Frank is into video more now, matroska implementation can prove useful to wake his interest again, who knows ? And with current SV7 bitstream this seems to be quite hard to do, so a move towards a fixed SV8 bitstream could help a lot already IMHO ......

EDIT : Frank was quick ! I just received an email from him in German, about a clear to-do list. I forwarded it to md`and Mosu, as i cant do anything about it for the next 3 days, sorry, as i am on an exhibition now . Anyhow, i see this as a good sign !!! Dont you ?

It sounds from this to me that what makes MPC gold is the psy model which needs a very, very skilled person to tune, so does Frank plan on tunning the psy model any further?

YES!!!!

waahooo

Is the psymodel really that portable between codecs? If so, then i'm surprised we haven't seen codec development split more evenly on between the psymodel and the encoder.

As another aside, whats so different about Franks psymodel?

It exploits post-masking and uses sophisticated tone/noise measurement methods. As you know it was carefully tuned and optimized by a perfectionist. Other than that, I couldn't say (but would love to know).

Frank's webpage mentions clear voice detection, but I don't know if it's used to gain space or to improve quality. He should write a book

Edit: Oh, and it uses an efficient mid/side stereo scheme (can be toggled independently in each of the 32 bands).

It's good to know that Frank is still active and generating plenty of ideas, however, I do hope that he will at least bring SV8 to completion.
Well, mpc is not dead as long as there are users and developers still behind it.

Yeah, it's frustrating when it's *so close* to being perfect.

I'd like to know exactly how Frank thinks transform codecs could be superior. Aren't they always going to be the other side of the 'quality/size' tradeoff than subband codecs like MPC? If he thinks AAC can be equally good with a lower size then great, but i'm extremely happy with the size/quality ratio of musepack as it is now and any trade-off seems a step back.

Whatever, it's nice to see he's alive!

Not trying to be pedantic but I'm pretty sure that a lot of groundwork of the psy-model, including the Clear Voice Detection, was done in the time Andree Busschmann was still working on it. One thing that was done is discard the standard ATH curves and determine more sensitive ones by doing actual measurements (anyone remember Filburt?).
But Frank took it to a higher level by optimizing, improving tonality estimation (that's what he was doing the last time an alpha was released) and changed it into something scalable (--quality slider). So we have scalable ATH's now.

OK that not the future of Musepack (more like history).

Personally I don't see Musepack as the ultimate codec for video (but I don't mind being proven wrong), there are already codec’s that support multi channel, are bitrate friendlier (more room for the picture) and have less syncing/muxing problems.
That said, for Music it is my absolute favorite .

@ChristianHJW Thanks for sharing this with us, we've been waiting for a life sign from Frank for a long time.

To what i understand by the post. The are so many limitation to what Frank wants to do with his idea.

Then why not frank just free his mind and create a super douper Audio Codec Mpc sv8. Departing from sv7, mp2 mp3... etc.. A completely different thing to what we have seen b4.

Or may be that is too much work.

To be honest, We have already got the best and near perfect phy model here. We could fit it in to other codec as stated. However, i personally like Vorbis but their organisation is some what wired and I don't like them as a whole. AAC is not good if there are no patent issues around it.

( Yes i am waiting for Roberto opinion as well )

Nope. According to one of HA's admins, I'm not supposed to have opinions on MPC's future. Or, at least, it seems I'm not supposed to post them.

I won't say who is this admin either because I'm not allowed to say his name.

So there.

Ha...

And the drama ensues....

Not that it will matter much to those who are quick to make assumptions, but this is really pretty much nonsense.

I was really surprised by this, so I contacted Roberto. Turns out that he's angry at CiTay because of what he said in this thread:
http://www.hydrogenaudio.org/forums/index....showtopic=14741



CiTay asks Roberto not to use his real name, not to cover "who gave the order" like one could think from the Roberto's message above. Also in the thread above CiTay in no way says that Roberto is not supposed to post to the MPC threads anymore. Roberto also bases his claim only on this thread, there were no other "orders".

Roberto told me that he even got support email based on the missunderstanding he creates in his message and sympathy that admins silence him here... I asked Roberto to clarify his message several times, nicely, but he refused, so he gave me no choice but to explain this publicly.

Obviously, since I'm recently working quite hard to make HA even better place, I'm not very amused to see this from Roberto, especially because imo he gives simply wrong information and bad image of us.

Fwiw, I just got the feeling Roberto was being sarcastic -- I didn't get a feeling like he's giving HA (and/or admin) a bad image. IMHO it wasn't strong enough a statement that it feels that way.

As far as incorrect information, that part's open to interpretation anyway...

This goes off-topic so I hope nobody continues this here anymore.

I do consider allegations that he's instructed by one of the admins not to post to MPC threads anymore wrong information.
Considering that he started getting support emails against the "HA admins who silence him" few minutes after his post, there are surely lots of people who didn't see his message only as sarcastic. It's also funny how soon people start sending support mails after incidents like these to Roberto, but I have no hope of getting any support or simple unprompted thanks PMs pretty much ever. All I get is messages which mean work for me, for the benefit of the members or forum.

Try to be admin of this big forum yourself, and see what you think when some of the regulars imply that you possibly gave this kind of "orders".

Anyway, no more about this. Back to the topic, please.

I'd love to be able to add something more on topic, but I do have to point out that it should be pretty obvious why Frank doesn't like posting to this place, because of "whatever his choice of words were", mine would be "full of unconstructive drama queens".

Anyway, I agree with Frank on everything he says. And he has generated an excellent amount of good idea's, it seems to me. I am feeling positive - and, like many here, have no understanding of, but an unbelievable amount of respect for, the newest holy grail, MPC's psymodel

i am a MPC fan (the only one i use)
i think that robertos's posts are based on facts clear and objective
this forum has been using scientific statisticals tests to evaluate the codecs
opinions are always welcome, (freedom of speech), but the conclusion can only made on the ABX tests protocols

The most interesting thing in this thread (for me) is the idea that subband codecs are not automatically superior to transform codecs. Not that there's a reason why they should be, but I think that it's been an unwritten assumption around HA for a while?

I can imagine that a transform codec could be made perfectly transparent with an intelligent psy model. Just like a sub band codec, this would mean dramatically increasing the bitrate on very difficult signals.

Could it be that, if equal attention is given to ironing out the problems of each, then both would be roughly equal in terms of bitrate/quality? I doubt we'll ever find out - not for many years, by which time codec design may have changed beyond recognition.

So, the users just have to use what's available. This is currently mpc. However, if AAC can be tuned so that we can trust it as much, it's the obvious choice. I wonder what Frank sees as the problems with AAC?

Cheers,
David.

P.S. JohnV and the other mods: you do a very difficult and very good job. I'll email to tell you, if you really want some fan email! But hopefully "Thanks" will do, so thank you.

Unwritten assumption? Not necessarily. As far as I kept track of things, the general idea was that 1) transform codecs are more efficient in low-bitrate scenarios (~<130kbps?) while 2) subband codecs scale better at higher bitrates and 3) it is easier to avoid pre-echo issues with subbanding. (but maybe it's just me. )

Is it along the lines of transform codecs being more complex that means tweaking is more critical to get decent quality? In which case, i'd stick with mpc as we're still tweaking the hell out of mp3 and aac, and with both these codecs quality is incredible variable between implementations. MPC as a subband codec gives me more confidence that the compressed version will sound decent, i'd be wary of moving to aac regardless of who is working on it.

And yes, I agree the mods do a great job, but am guilty of assuming my continued presence on the boards is proof enough of how much I appreciate it - thanks to all of them.

Well, if Musepack development really completely stalls, then let me suggest that a group of ambitious developers from HA pick up AAC or Vorbis as it is, iron out the stupidities in the specs, and create a new derivative codec, exactly like what Andree Buschmann (I hope the preliminary spelling is correct) did with mp2 and gave HA its most cherished codec. It would be a very interesting challenge tweaking a transform codec, as some of you have stated. Then add some assembly optimisations to it, and we get a codec superior to any other, with the same patenting muddle as Musepack.
It's.... just an idea.

That's what I was thinking sld.

But if the result is a codec with similar performance to mpc, what is the point - apart from as an interesting academic exercise?


As for transform vs sub-band. Let me point out something that a very clever person pointed out to me:

A sub band codec is based around a filter bank. A transform codec is based around a time to frequency transform. But the two can be mathematically identical. It can be two different ways of performing the same process. It's not necessarily two different processes...

So, you can (theoretically) make a very high resolution filter bank (say, 1024 bands or filters). If you decimate the output of this filterbank (throw away redundant samples), you can be left with an output equivalent to the output of an FFT of those same samples. How? well, the only important information is the amplitude and phase of the output of each filter - if you know this every 1024 (or 512) samples, you can predict (calculate) what the time domain waveform is. So the only useful information at the output of each filter is the same as the information which comes out of a time>frequency transformation: amplitude and phase, every 1024 (or 512) samples.

Here's an opposite example: In place of a very fine bank of filters (e.g. the 2000 or 8000 that are used in COFDM digital broadcast systems) you can just use an FFT to split the signal into 2000 or 8000 bands.


However, it's perfectly possible to design a filter bank which isn't equivalent to a simple time to frequency transformation. And each real world time to frequency transformation has its own little quirks. Still, it's worth getting your heard around this issue to understand what's involved (and I don't claim to have thought it through completely!).


Cheers,
David.

Why is there a "MPC ist dead, no shut up, MPC lives" debate? Its close enough to 'perfection' that any additional code isn't going to achieve a great deal.

The mentioned interest in AC3 as a starting point sounds much more interesting. Thinking of the improvements that went from MP2 -> MPC, I wonder what APC would end up sounding like for high bitrate 5.1? All things going well you could have something that DTS was meant to be, I couldn't imagine what Widescreen Review would write about it, the authors would all spontaneously combust with joy

WOW, I think this would be the best option, assuming it can be done. An MPC compatible with AAC decoding!!! There are already AAC decoders so if we can get the benefit of wide industry support and the quality of MPC (or MPC's psy model) on standard players, that would be amazing.

Maybe Nero would even be willing to pay him?? I doubt apple would but it would be worth looking in to.

This is a good news for MPC fans!!

Christian keep us updated and tell us if the community can do something (for e.g. beta testing, ABXing, etc.)!

Fr4nz

I have to agree. Amen.

Your point being? I'm very happy with my personal 100 GB of mpcs... RIP on my HD Drive Or RIP every CD you can put your hand onto?

Sorry, but as dead as it is, its the best codec out there (no HW support but you can't have it all...)

edit: ok, not dead, it's more like a zombie case here

IIRC, Franks biggest rant about AAC was its inability to store a decent encoder latency time, making it difficult to append 2 AAC tracks into one. Other than that, i may not have understood what he was trying to tell me.

I dont know if this would require a huge change, but maybe it could be done on the container level even, so the raw frames were identical and all existing decoders could be used for decoding MPAAC ?




Why should they be interested in this ? They are pushing AAC and the MPEG4 Standard, so their bigger interest was in a better tuned AAC encoder i bet ....

You can put whatever you want into an MP4 - all you need to do is decide where to put it AFAIK the QuickTime file format has full support for metadata on both files and tracks.


Reading from your original post, isn't this what Frank was talking about - more or less? Creating an AAC encoder using his superior psycho-acoustic model?

A few days later, and a few more glasses of wine...

I only said that because, from what I was able to determine from that conversation, it sounded like Frank was not going to work on MPC at all, but rather, turn it over to others. Maybe I'm wrong about that.

I was always under the impression that Frank (and probably Andree) were the only ones that understood MPC enough to actually improve it. Maybe I'm wrong about that as well.

Some would argue that it's already good enough, and therefore, does not require any improvement. I can almost agree with that, but not quite.

Anyway, If Frank is done working on MPC, and nobody understands it as well as him, then I think that's not a good thing.

On the other hand, maybe I misunderstood the conversation, and if so, my apologies.

I think many of us understood the original post as just this. If it were to be compatible with AAC decoding (basically appling his psy model and a few tweaks to AAC I presume) then why wouldnt nero be interested. They currently pay a number of developers to work on projects like this an maybe with a little money frank could dedicate a little time to at least help the nero developers (menno, ect) to implement his work. And if he does have an increased interest in video maybe he could find a niche with nero for that too. He is clearly extremely talented and it would be shame to loose his talent.

An open source high-quality AAC encoder focussing on transparency, using Frank's great psymodels would be WONDERFULL! MPC is a great format except for the hardware support.... this would be damn near perfect!

Now, be careful here. As clever as Frank is, some of these posts almost sound like they're suggesting that Frank is God, and current AAC encoder developers are stupid!

The idea that Frank can port is psy model to AAC, and magically it'll be perfect sounds a little strange. I'm sure the friendly AAC encoder developer we know is already doing a good job.


I agree it would be cool for Frank to get (re)involved with any audio coding project - I was just a little worried that some of the comments in this thread might offend the people who are already working on a certain codec. Though they've probably got better things to do than be offended by such gentle comments.

Cheers,
David.

You put the words right out of my mouth! I can´t agree more