Digital audio formats

quote:

Originally posted by JeremyB:
Beginning to get as irritated as Munch but on a different subject.

Is there any significant engineering talent left working in the audio field?

It looks like the only chance to get any improvement over CD for distributed audio is 88.2kHz or 96kHz and 24 bit.

Using over 3 times more storage and bandwidth than necessary to encode frequencies only half of which we can hear and 16 million steps 99.9% of which we can't hear.

What a waste. Are engineers working in the audio field too lazy to learn how to design any decent filters and too closed minded to use basic encoding rules known for 40 years to engineers outside the music industry?

Upsampling was an attempt at a solution to that but there's still something missing at the top. What you need to understand is that just because 44.1k can reproduce a 20k signal it wont do so in use at proper amplitude without compensation. They've been trying but sometimes the limit is the format. Another solution would be to record at higher sampling rates with a good algorithm to down convert to 44.1. In theory this could get more of the high freq. info back but there may be some time issues associated with harmonic structure and even in this case it can't be perfect on res. Tried this on some of the best programs and it's always not as good as the higher res file. You can rebuild the file back to 96 and it still loses to the original.

As for filters, it's like asking an engineer to make a steep passive speaker cross-over with linear phase. Dealing with square waves brings another level of complexity. I think seeing digital audio engineers as that incompetent is a bit short sighted. The rules they use are the same as those encoding and filter ones you talk about and dig audio engineers have moved well beyond them as their results were quite audible yet perfect to the theorists. It's that we imagine we here a difference that's the issue. LOL Thank god some smart people are as misguided as we are and continue to identify problems to be solved.

quote:

They've been trying but sometimes the limit is the format.

Audio engineers create the formats don't they?

Storage is cheap so why worry ?

I agree - I'd much rather have to store larger files to have a 'simpler' solution like higher sampling freq and bit depth than have complex algorithms which inevitable through up audible artifacts.

quote:

Originally posted by JeremyB:

quote:

They've been trying but sometimes the limit is the format.

Audio engineers create the formats don't they?

Yep and is why 24/96 is getting play. Sony/Phillips created CD sound which came with the 44.1 standard. Not all audio engineers. Some very good ones have been trying to make it work ever since.

Of course then you get DAC comparisons on Pinkfish where they all sound alike and perfect because they're using a limited source. Other sites can't here a 320 bit rate. LOL. To me, these differences are huge but you have to have your ducks in row.
Like the guys said, storage is cheap.

The video equivalent of your OP is for scaled 480p to look as good as native 1080p on a 100" screen. It can't or maybe it's not just the audio guys that are incompetant. The solution was more res just like in audio.

quote:

Storage is cheap so why worry ?

Check the title of the thread - distributed audio. It costs more to send all those extra bits and the telcos are capping this amount of data, soon they will charge per bit for data. Three times the number of bits equals three times the cost which equals three times the price to the customer, that is how it works. This wasted storage and bandwidth may not bother enthusiasts but my concern is that it wil prevent the format taking off, if the end product is $45 instead of $15 for example.

Jeremy - put in that context it is a waste. Going back to your original opener, there are plenty of talented Audio Engineers around - but all the big players dont want to invest. Joe public is happy with MP3 / low rate downloads so why develop better, higher quality codecs for a limited market. A great shame

James

quote:

Originally posted by james n:
Jeremy - put in that context it is a waste. Going back to your original opener, there are plenty of talented Audio Engineers around - but all the big players dont want to invest. Joe public is happy with MP3 / low rate downloads so why develop better, higher quality codecs for a limited market. A great shame

James

It seems that we all agree that there are no free lunches. I guess our perspective on that is somewhat different. HDCD and almost every type of filter, upsample etc. has been tried. I'm intigued by upsampling in conjunction with a minimum phase filter though I've never heard a great package of either. Eliminates pre ringing but with phase shift. The phase curves I've seen look very much like those from an all pass filter which while showing a lot of shift, do so in a very linear and gradual fashion across freq. and tend not to be very audible.

quote:

Originally posted by JeremyB:

quote:

Storage is cheap so why worry ?

Check the title of the thread - distributed audio. It costs more to send all those extra bits and the telcos are capping this amount of data, soon they will charge per bit for data. Three times the number of bits equals three times the cost which equals three times the price to the customer, that is how it works. This wasted storage and bandwidth may not bother enthusiasts but my concern is that it wil prevent the format taking off, if the end product is $45 instead of $15 for example.

I think this becomes more of an issue with streaming as opposed to a one time download of a hi-res file.

quote:

Originally posted by JeremyB:


Is there any significant engineering talent left working in the audio field?

What a waste. Are engineers working in the audio field too lazy to learn how to design any decent filters and too closed minded to use basic encoding rules known for 40 years to engineers outside the music industry?

YES.



MP3 sucks but people buy it. For Audio engineers to get on with a new format it has to be supported by the ones who write the check. There is no craptastic turd polisher 3000. To make a mp3 not sound like an mp3.

There are 2 different kinds of Audio Engineers.
One makes hardware and one makes music.

Hi Patevil/JeremyB

Most of the audio (music recording) engineers I know who learned their trade through tape-op jobs in the eighties have found it impossible to make a living from the business and have moved to other occupations.

The complete dissregard for the very idea of paying for music at all among the next generation of consumers has resulted in most productions never comming close to a 'recording engineer', with audible results. I trust there is no doubt about this particular subjective comment!

A selection of cracked plug-ins can be a very powerful recording solution but in the wrong hands it is capable of doing what we hear everyday..

However, these things tend to move in cycles. Maybe the success of Elbow with their 'play it louder' approach might make some headway?

The characteristics of audio formats generally are not decided by audio engineers, but by the bandwidth limitations currently limiting the potential for high resolution downloads. This does not reflect well on the AES (Audio Engineering Society). Back in the 80's and early 90's the Producers Guild in London debated many of these issues regularly at the BASF building - then mp3 happened.

I have a suspicion that the majors are now digging in to start a new cycle of 'improvement' by supporting the mp3 market now with a view to reissuing their catalogues in some new HDMI like hardware/software locked format..

Cheers

Peter

quote:

Originally posted by Peter Williams - Ainm Eile:
Hi Patevil/JeremyB

Most of the audio (music recording) engineers I know who learned their trade through tape-op jobs in the eighties have found it impossible to make a living from the business and have moved to other occupations.

That is why I have a day job now. (Used to work in Nashville) Now I work to support my habit of making music.

quote:

Are engineers working in the audio field too lazy to learn how to design any decent filters and too closed minded to use basic encoding rules known for 40 years to engineers outside the music industry?

I think you've missed the distinction between recording engineers and audio-focused EEs–also, quite honestly, implementation of some of the more sophisticated filter maths is only now becoming a possibility with new DSP platforms. The ear is far too good a transducer to mask over insufficiently-evaluated audio calculations, which has helped the cause of those with a less-than-adequate understanding (who would prefer NOS solutions, or NO solution) of the maths involved.

Also–as we've seen clearly with FLAC as a format, every encode / decode process has a price paid in computing cycles, with decoding delays caused by this directly affecting audio quality (frankly, the only benefit I see to use of this format is the embedded tags, not the storage compression).