A new CD Audio format?

[elektro80]

And if not CD.. what?
There are already "new" formats out there, but do they make any sense for consumers?


There have been many complaints re how bad CD Audio really is. But is it really that awful? What is wrong?

Is stereo bad?


16bit/44khz is of course not cutting edge, but would 24bit/44,1khz be acceptable?

[opg]

I don't know much about DVD Audio discs. What's up with that? This reminds me of the discussion about how digital is bad for your ears because of the changes in complete lack of noise to sound - and analog (like cassette tapes) always have the self noise that eases this. I always thought this was a bigger issue than some crazy bitrate. Do we really need to record 5Hz and blow out speakers? How good is a person's hearing that digital needs to be better?

And yes - stereo sucks! I'm doing everything in mono now. Stereo was overused on those old LPs in the 50s, and panning is just gratuitous nonsense.

[elektro80]

You know, the bad sound of CDs..
Obviously a tad lower than 20 hz would be nice, a tad more up there would be nice.. but IMO 24bits instead of 16 would be really nice.

Re that bad sound of CDs, I still think that main problem has been bad transfers from old masters, bad production, bad mixing, and the brickwalled masters, which are too popular today.

[Kassen]

OPG; the idea that digital has no noise is completely falase. Digital does have noise, it's just a different kind of noise. There are always rounding errors, there's jitter, etc, etc etc.

It's not as obvious as tape hiss but it's there and it does affect the way you hear things.

Anyway, I think new formats are mostly about re-selling the same old things again. The bottleneck at the moment, IMHO is a aesthetical one; CD or even vinyl with it's 60db noise floor have much more potential for dynamics then what's used at the moment. THAT affects sound quality, not even DVD or serious reel to reel will rescue you there. Another bottleneck is cheap speakers on consumer setups. I don't think there is any use in going DVD as long as those limitations are in place.

Oh, well. Seems like most people attach more value to what they hear in marketing then to what they hear in music. The specs gaming consoles like to promote are quite amusing if you rip some audio from the disk and see what rate those are sampled at....

[DrJustice]

I'd be happy not to see too many new formats for a while... Until now I have actually thoroughly enjoyed the sound of quite a few CD's. At one time I did hold hope for music DVDs with 24bit @ 88.2kHz / 96kHz as the new CD. The higher sampling frequency would make good working conditions for the reconstruction filter. But that's not going to happen, is it?

The trend is looking more like SIM cards with 32kb/sec MP3, or sumthin' - perhaps all music in the future will be distributed as ring-tones only

DJ
--

[elektro80]

[opg]

[elektro80]

[DrJustice]

[mosc]

Personally, I thing that for stereo the 44.1K 16bit standard is quite ok. I like 24 bit for recording and sound and music construction, but with properly mastered 16 bits sounds just fine. On the final product you don't need the headroom.

In a future format, there needs to be multichannel capability. There are some standards already out there. I think Robin Miller has one of the best standards there are. It uses 8 channels - 2 front, 2 back, and 4 for the ambisonic signals. Thus, it supports nicely any playback system from 2 to more that 12 channels. It supports 3D as well as plainer surround.

[elektro80]

[Kassen]

I'm not at all sure about how usefull that tag would be.

I like CD's (or anything else) to go into my amp as hot as they can without clipping, I then use this very advanced technology called a "volume dial" to set the volume to my liking, given constraints like neighbours.

To me it would seem like a very bad idea to have people like Alec Empire (to take a obvious pick) set this for me.

The issue with pop isn't in the volume per-se, it's in the lack of dynamics. This in turn gets interperted as loudness by your brain which is partially due to your ears working as a sort of compressor when confronted with steady streams of high-volume material. (at least with regard to preceived volume on it's own, obviously there are other factors involved as well). Tags won't fix that though higher bit-depths should enable material with more dynamics to sound better...

I have little hope for that. Vinyl recordings often benefit from limiting dynamics due to the -60db or so noise floor which CD-fans think is pathetic but what is the actual noise floor of modern listening conditions? What is the practical noise-floor when playing back material close to a computer with fans when using cheap speakers? What about systems in cars or indeed typical conditions in urban homes or headphones in public transportation?

Multi-channel is a interesting idea but I'm gonna sit the next few "standards" there out as well. Still too gimicky for my tastes and I find the end results unsatisfactory so far on everything on the consumer market. It's interesting when you can design your own system and compose for that but then multi-channel wave is good enough as a standard.

So; basically I don't think it matters what standard for recordings is cooked up at all since the bottleneck is in other factors. Putting a spoiler on your lawnmower won't make you mow the lawn faster either. A very large factor, I suspect, is that attentive listening as a activity on it's own (much like reading a book) seems to have been pushed aside by a demand for a continual stream of visual input and a preceived increased in stress (for lack of a better word). I don't think any of that is inherently "bad" but we do need to take it into account. The future might be there or in a new format that will adapt to the activities performed *while* listening to music.

[Stanley Pain]

interesting thread, but it seems impossible to have this discussion without it turning into another discussion about the "loudness wars" and ethics of mastering.

[kvnvk]

[opg]

[elektro80]

Well, I still think certain standards for mastering/product quality should be considered. If the content industry thinks it is OK to brickwall masters, the industry has also pretty much decided:

1. that audio quality is of no concern
2. that there is a consensus on a standard volume level
3. this standard will be forced upon the consumers

Implementing a perceived standard playback volume while keeping audio quality and data integrity would be best done in the playback equipment and not in the data on the CD.

The probably best place to implement this in playback equipment would be in the CD player ( or the digital media player be it software or "hardware" ). The idea is however to pass this control data to the analogue circuitry so all gain adjustments would be done in the analogue domain. I imagine each track could have its own tag.

As for tags, you guys probably already know the TIFF file format. One might easily imagine a similar approach for audio files. Such tags could also contain say instructions for multichannel playback and more. Only problem here is of course how the audio stream is laid down on the CD and how it is played back .. but still we have CD-TEXT now and tags could be stored the same way.

I´m not thinking of a gain riding autopilot here, but a set-once per track functionality.


Back to the "problem".
The reasoning behind the

[MrHope]

24 bit would be nice.

[Antimon]

I have a problem with levels related to the connection between iTunes and iPod. I recently bought an iPod to use at work, and I had already run iTunes for a while on my Mac. Some records that were more silent than others I had adjusted the level for (you can do that per song in iTunes, in the info window). I felt that I needed to boost some old records to 200% in order for them to be the same level as newer records.

Anyway, when I transferred my iTunes stuff to the new iPod, and then listened to shuffled songs at work, I had several scares as old tracks started playing at incredibly high volumes compared to previous tracks - it seemed that the difference between 100% and 200% in iTunes is different than in the iPod. I don't know if this can be affected by the listening device (monitors or headphones), or some subjective thing. It would be nice if this was solved by some standard.

I think maybe the new stream concept can speed up this process, with the required shuffling between tracks. How do internet radio stations like Soma FM (or radio stations in general) solve this - do they manually check all tracks for levels?

/Stefan

[elektro80]

iTunes/iPods.. there is a hidden feature in there somewhere which analyzes all the tracks and then sets a tag for each track.. in order to make all tracks sound the same level wise. I guess you have manually adjusted as well as the auto thingie activated.

See iTunes prefs.

Broadcast... arrrgh...

See: http://www.orban.com/

Many stations will simply smack on some compressor/limiter and set this to a standard level ( which usually means nothing sounds really good at all ).

There is however more to this than that..

[Kassen]

Well, there is a perfectly fine standard for that, it's called db spl weighted. The weighing is to take into acount how loudly it sounds to human ears, adjusted for frequency distribution depending on volume.

For some reason people always use "a" weighing which makes NO SENSE AT ALL for the majority of the contexts wehere it's used since "a" weighing is meant for very low volumes.

Since Mac products often cover the whole process, from the software to the final speaker, it should be possible for them to adjust volume in terms of the preceived end-result in many cases. I think this is a bug or a mistake (or lack of planning, not so much a lack of a standard

[elektro80]

[jksuperstar]

There's a big push in the software world to have programs that run off of a USB stick (for example), so there is no installation, and you always have the program to deal with your data together. To me, this is akin to object oriented programming, where the method & data can be tied together.

Anyway, extend this to CD's (or however the audio is distributed), and the intended method of decoding the audio, be it from mp3, flac, ogg, or multi-channel raw, is always with the audio. Maybe your player can be smart, and decide that the decoder it has might be better (a user setting), but this allows the engineer/mix/master guys to help determine how the music is played on your 2 channel stereo (while the original contains 8 channels of audio).

So, basically, the "CD Player" has some level of virtualization of software needed to read the CD.

[mosc]

In a way, that is what the surround sound capable receiver does. It just needs more decoders. In some cases, it makes sense to keep the player very simple so it doesn't know much about the decoder.

Oh well. We don't need a better format, we need better music.

[opg]