What is the status of ogg highbitrate coding???? im trying to be patient...

You living under a rock? They are working in other projects like Theora, no recent fine tunings planned are until then, and its eta is about middle next year at best. On the other hand "Vorbis" IS finished, the finetuning corresponds to the encoder used, which may not necesarilly mean oggenc. Third parties can create and fine tune their own Ogg Vorbis encoders, or help improve oggenc.

And this ice cream with cheese thing... yuck Take Vanilla ice cream with black beans instead, heh, very good (Japanese style which means sweet black beans)

GT3, we are all waiting !

According to the tests I did GT2 still kicks Vorbis 1.0's ass in the bitrates it provides (128, 160 and 350). Vorbis 1.0 only is better at bitrates <128kbps, if you want anything higher I'd at least try GT2.

dev0

Have to say Vorbis is a stupid audio codec! People like it because it's open source! All surround info is damaged until a bitrate above 224 kbps is used for Vorbis......Urghhhhh

As for as I'm concerned (though I'm not directly involved in Vorbis), that's not the top priority. Some people mentionned Theora, but there's also lots of work being done on making Ogg (the container) better, including mixing codecs in the same stream. Most people (except on HA) just don't care about high bit-rate... but DO care about low bit-rate (<64 kbps, streaming, etc).

As for what bitrate people prefer, I don't see really any proof either way: are your statements purely speculation?

For myself, I can never envision the time when vorbis or any other codec would sound acceptable at <64 kbps, but I would certainly love to see vorbis improve at ~128 kpbs, with better stereo and less high-freq 'harshness'. Obviously the vorbis devs have already given up on any bitrates higher than that anyway...

I'm not talking about archiving your whole CD collection here. I'm talking about Internet radio type of quality. Most of the ones I've listened to are crap. Besides, most of the time I listen to music, I've got all kinds of noise around me, so it doesn't matter that much. I won't rip my CD's at 64 kbps, but I'll listen to radio and things like that and I most likely won't even notice any degradation.

Another way of saying that is "I prefer listening to the music than listening to the codec". Maybe with good headphones, and lots of attention you can "catch" artifacts in Vorbis around 160 kbps. Now, if you sit down on your couch, stop focusing on artifacts, listen to the audio with your speakers, where's transparency now? Add to that the normal noise you get in a house, and you'll see why most people don't care. (or just see how many people encode 96 kbps MP3 using Xing and don't find it that bad)

I can NOT confirm this !

Your Dolby ProLogic decoder is obviously crap and has problems with small phase shifting. Even the simple DPL decoder in my Dads TV sounds excellent, when playing my OGM files with Vorbis q = 1.5 .

Once Monty has time again to look into Vorbis and add a nice 5.1 channel coupling profile to it this will be history, as Vorbis 5.1 should be feasible in very good sound quality at bitrates 140 - 180 kbps, with discrete channels.

( BTW : There was some work done on an on-the-fly Vorbis 5.1 to AC3 encoder, so 5.1 Vorbis could be sent to external surround receivers via SPDIF, using AC3 only as a transport layer at max. bitrate 640 kbps to link PC and surround receiver digitally ).

you mean encoding a 448 kbps ac3 track to 140-180 ogg 5.1 track, then realtime transcoding back to ac3 at 640 kpbs for use with spdif? I'll reserve judgement til I hear such an implementation, but the at first glance it seems like that idea would produce horrible sound quality, especially for all the extra effort required.

You have a very valid point (only a relatively small user base has much use for high-quality Vorbis audio, although perhaps not as small as you think), but your argument (that everyone should just chill out and stop listening for artifacts) will not sit well at all on this board.

The high-frequency harshness that floyd mentioned is an obvious problem with Vorbis at 160kbps (q5). Maybe you're good at ignoring artifacts in music, but I lost that ability a while ago, for better or for worse.

There are plenty of good reasons why Xiph.org isn't tuning Vorbis for high quality right now. But "audiophiles should just zone out until they don't hear anything wrong" is not a good reason.

You are comparing apples with peaches mate.

A 5.1 Vorbis audio track can not be compared with the original 448 kbps AC3 track, for sure. If you want optimal sound quality you have to stick with the AC3 and either buy the DVD, make a 2 CD rip or get a DVD burner !

But IMHO a 5.1 Vorbis track could be a very nice alternative to a Dolby Pro Logic ( DPL ) encoded Stereo track with q = 1 - 3, and thats the goal.

About the on-the-fly AC3 encoding to use SPDIF : a 640 kbps AC3 track should be so much superior in sound quality compared to a rather low bitrate Vorbis that i can not see any possible degradation in sound quality by doing so, depending on the quality of the encoder implementation of course.

Hmm - I think the best idea is to leave AC3 encoding from DVD as-is, unless you want to get really low bit rate. AC3 is a perceptual codec (and not a state-of-the-art perceptual codec), with loss of quality in every encoding step. So, every transcoding from AC3 source (or any source encoded with perceptual lossy codec) would degrade sound quality further.

Even worse - transcoding from 448 kbps AC3 to 140-180 kbps Vorbis, and then reverting back to 640 kbps AC3 would make things even worse - that's another distortion step (Vorbis encoding) plus again AC3 encoding.

Ivan, of course there will be a degradation of sound quality when transcoding an AC3 track into a low bitrate Vorbis, but as a matter of fact you have to accept that if you want a 1 CD rip. For sure you know what a 1 CD rip is ?

The reencoding to AC3 640 kbps on-the-fly MAY add some new sound degradation, sure, but do you think this is really noticeable, especially when being compared to the first loss ? Those who want to get the last bit of sound quality out of a 1 CD rip can always hook up their 5.1 soundcards to their surround receivers via 5.1 analog input, so they override the AC3 encoding stage. But for those who dont have a 5.1 card, but only a soundcard with SPDIF out and an external surround receiver with Dolby Digital decoder, they for sure will appreciate it, compared to a DPL encoded Stereo track ...

Another option is of course to use the 192 kbps Stereo AC3 tracks coming with most DVDs here. I'd be keen to make a comparison what will sound better

- 5.1 Vorbis 190 - 200 kbps, encoded from 5.1 AC 384/448, and reencoded to 5.1 AC3 640 kbps

or

- 2.0 AC3 192 kbps with DPL surround

Maybe very much depending on the movie soundtrack, dont know. For musicals, etc. i'd probably go for option 2, taking the degraded effects into account for less artifacts, but for a normal action movie i'd probably go for option 1 .... we will see once we are there, and to return to the original topic, i cant see a proper 5.1 channel coupling profile anywhere down the road, as Monty is busy with Ogg Theora right now as we all know, so its probably a bit hypothetical what we are talking about.

There is completely no sense in -reencoding- it to AC3 (especially when no good software AC3 encoders exist). And i do believe Ivan knows what i "1 CD rip" is ...

Much as I am very much looking forward to Theora and am looking forward to it's release, I think that sorting out a good implementation of 5.1 Vorbis at sensible bitrates (< 200) should be higher priority. It would be a real incentive to switch over to using Vorbis audio (and therefore ogg container) for movies, as it would finally make a 5.1 1CD rip possible. The only other feasible option at the moment is WM9 (urghhh).

This in turn would mean Theora would get more attention at it's lauch because the average user would have heard of Vorbis and ogg.

Just my 2 cent's worth.

there is. as far as i understand it the idea is to make the vorbis 5.1 stream usable for external ac3 decoders, if you don't have a multichannel soundcard for example.

but i guess it might prove to be difficult to create a decent on-the-fly ac3 encoder that does consume few enough cpu cycles to allow smooth video playback at the same time..

Ever heard of digital output to an external decoder?
Did you not read what ChristianHJW wrote about work being done on a software encoder?


Even if done on an interface card, PCI might not even have enough bandwidth for the audio. Best bet seems to be the NForce/NForce2 APU with the MCP-T Southbridge which does realtime AC3 encoding.

Well, what could turn out to be very helpful here is the fact that bitrate is no issue, so you can spend bits like crazy and dont have to care much about the PSY model used, consuming at least some part of the used CPU power AFAIK on a normal encoding process . I honestly speaking have no idea about the encoding principle AC3 is using and if there is any DCT involved on encoding or not, but i feel any Athlon or better should be capable of doing that. The only remaining problem then will be that the very same CPU has to decode the 5.1 Vorbis and the MPEG4 video ....


Yup, AC3 hardware encoding would be excellent, but will it allow to output via SPDIF ?


We cant decide that, we just have to believe Xiph people know exactly what they are doing. Emmet and Monty have proven in the past they have a good instinct what users need and want, so lets wait what they come up with ...



edit: fixed quotes.

Would there be a point to ac3 hardware encoding otherwise, if the apu doesn't have spdif output? Of course there's spdif output

That is plain wrong lossless stereo coupling starts at q6 which is about ~= 192 KBps

No, Ice cream with C0ke

What is this GT2?

Garf Tuned 2
which is based on a vorbis version before the code went gold