Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Future of digital music distribution. (Read 12855 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Future of digital music distribution.

Hi,

While some time ago Audio CD was almost the only way of digital music distribution there are some growing alternatives nowadays – DVD-audio/video, SACD, via NET. All of them are very different in sound quality and the latter (via NET) is very flexible in this sense. Here I would like to share some of my preliminary observations of the subject and would be glad to get your comments, opinions and just thoughts.

1. Distribution via Net. Concidering:

•   Success of on-line music stores;
•   Impossibility to stop file sharing in real world;
•   Wide spreading of broadband internet connections;
•   Continued demonopolization  of audio production and distribution (independent labels, authors,       DJs, home studios),

I think that distribution of audio content via Net (legal or illegal) will dominate over all other channels in the nearest future.

2. Lossy vs. Lossless. All consumer audio will be lossy compressed as like images and video already. Technology always tends to give more for less. In our case it means that for any target bitrate compression algorithm has to provide the best possible perceived quality and if the goal is some level of quality compression algorithm has to provide it at the lowest possible bitrate. Lossless compression is not effective in this sense and is useful only in professional sphere where possibility of further editing or research is essential. 

3. Audio CD is lossy as well. In most cases output of audio production looks like this – 24/32bit@48/96/192kHz and is converted to 16bit@44.1kHz at the final step of mastering. This conversion even with some sophisticated dithering applied is the simplest and roughest of all bit reduction schemes. Today’s state of the art codecs could compress those masters with better quality and at lower bitrates.
The same time so-called “CD quality” is very important anchor point on quality scale, because it is well known to majority of listeners.

4. Quality classes. Considering results of different listening tests and real-world usage of various codecs all audio could be divided into four groups (classes) according to perceived quality criteria:

PREVIW QUALITY
   - Slightly-to-not annoying artifacts
   - Low bandwidth applications, demos
   - Mono/stereo
   - ~24-64 kbit/s
CASUAL QUALITY
   - Transparent to majority of listeners
   - Home stereo/video, DAPs, radio
   - Stereo/multi-channel
   - ~128-192 kbit/s
TOP QUALITY
   - Not worse than “CD quality”
   - Attentive (involved) listening, audiophile’s subculture
   - Stereo/multi-channel
   - ? kbit/s
PROFESSIONAL QUALITY
   - Best possible quality
   - Audio production
   - Mono/stereo/multi-channel
   - Lossless
   
5. Top Quality Distribution. While most of listeners are pretty satisfied with “128 kbit/s, best codec” quality and almost happy with “192 kbit/s, best codec” quality there is a minority of listeners with dedicated rooms and special audio/video installations. Yes, they also want to be happy and need audio content with higher quality margins for that.

Sources of such high definition audio are:
•   24/32@48/96/192 studio masters
•   DVD Audio/Video, SACD, Blu-ray Disc and so on
•   Good-old Audio CDs which can sound very very good being correctly recorded and mastered.

The necessity of effective distribution of such content via Net is obvious. It seems reasonable to deliver it in some high definition lossy compressed format. The level of compression has to provide perceived audio quality not worse than “CD quality” for the most critical content. I know it sounds not very technical but I’m talking in principal now. If the source of Top Quality content is Audio CD it has to be compressed losslessly.

6. Codec for Top Quality distribution.  Actually any modern codec capable of handling 24/96 multi-channel audio is suitable for the role (after confirmation that it can provide above mentioned level of quality). The same time some special features would be helpful for such codec:

•   possibility of lossless compression for 16bit@44.1/48 kHz
•   some kind of quality protection mechanism (QPM) which would indicate (on decoder side) that the audio is the first generation compressed material

In order to be reliable QPM has to be an integral part of an audio content. Such mechanism could be realized, for example, by means of watermarking – adding of some special signal inside audio content far below threshold of human audibility for the price of some extra bitrate. In fact two layers of watermarking is necessary – one, very subtle, easily ruined by any transcoding and second, very firm on the contrary, not destroyed even by multiple transcodins. Presence of both marks (and only two) will indicate that audio content is not transcoded and is compliant with Top Quality specs. Yes, this protection scheme prevents post transcodings only – the source for the first encoding can be of any kind. May be incorporation of some tools like “true-audio analyzer” into decoder would be a solution, but I’m not sure that any ideal QPM is possible and really necessary because quality of delivery is only a part of the problem, another part is quality of production which is naturally uncontrollable. Say, QPM is helpful but not necessarily obligatory.

If to dream a little bit more (why not?) then it’s worth to add to the codec’s wish list such possibility as repacking Top Quality audio into Casual Quality audio without transcoding. 
   
7. And finally some thoughts about DRM and copyright. As it was previously noted there is no possibility to stop or even limit file sharing inside any social community like family or friends; and we all are brothers and sisters, you know  . For that reason any copy protection will be hacked for sure. Now this sport is very popular, even more popular than virus writing, so it’s a question of time. For that reason in turn on-line music stores (with their prices) can’t be considered as permanent solution as well.

On the other hand we all understand that work of authors, producers and sound engineers has to be paid. New forms and principals of economic cooperation between audio production and consuming parties are necessary to find. Some of them are in the air already:

•   Donation model (is under test now on commercial-free internet radio stations)
•   Commission model proposed by Fraunhofer IIS (http://www.potatosystem.com/info/en/index.html)
•   I wonder if somebody could convince those audiophiles not to buy shamelessly overpriced equipment in favor of deliberate buying of slightly overpriced Top Quality audio content, compensating this way damages from free sharing of Casual Quality audio content  .

Some other models may be proposed yet. In any case a wide and open discussion of the topic among all parties involved is important. Some solution (or several ones) will definitely be found because publishers and consumers are interested in each other.  Unfortunately big labels are moving in the opposite direction. Indeed, prosecution activity just consumes their precious business time which doesn’t “go slowly” for them any more.

Sorry for too long post.
keeping audio clear together - soundexpert.org

Future of digital music distribution.

Reply #1
Quote
Here I would like to share some of my preliminary observations of the subject and would be glad to get your comments, opinions and just thoughts.


Is this really a scientific question? I mean everyone has their own take on this. I personally feel that consumer tends to be shifting towards digital audio downloads on the internet and DAP's.  The audiophiles will never except anything less than lossless for their golden ears.  I guess it depends on what medium you plan to store music on. If you are a net person then streaming audio and perceptual audio coding for your portable player with keep you satisified. If you are an Audiophile with a large CD collection then you are probably going to want to make lossless backup's of your disc to store them on your HD or CD-R/CD-RW that has a long shelf life (Tayo Yuden).  I never see the SACD/DVD-A standard taking off, due to the fact that the average consumer doesn't probably have a 5.1 reciever in their home and RIAA did a great job not making the specification backwards compatible.  The only people who really will take a liking to 5.1 mixes are Audiophiles and those who are hardcore about the entire experience.  The entire optical media industry these days is a disgrace, because everyone is trying to corner the market with their own type of disc. They have these "brilliant" ideas for discs except how they plan on using them.
budding I.T professional

Future of digital music distribution.

Reply #2
Quote
Is this really a scientific question? I mean everyone has their own take on this. [a href="index.php?act=findpost&pid=373074"][{POST_SNAPBACK}][/a]
If I’m wrong with the forum, admin, I hope, will correct my mistake.


Quote
The audiophiles will never except anything less than lossless for their golden ears.[a href="index.php?act=findpost&pid=373074"][{POST_SNAPBACK}][/a]
They are also changing (slowly though).  If all audio is from Net, then some compromise will be inevitable.
keeping audio clear together - soundexpert.org


Future of digital music distribution.

Reply #4
I think it will be on-demand, subscription based mostly. I don't think either the users or the distributors would bother with DRM because not only it's useless but also finding the music and downloading (even with an intent to break the law) would take time and would be inconvenient. Your subscription plan would cover the playback on your computer and portable devices anyway (in lossless of course).

I have built a 400GB archive (mostly MPC but also lossless) in which there's quite some rare material. Although I enjoy it very much daily at the moment, I imagine within 10 years I probably wouldn't use it as much. This is because I expect I'd be paying a monthly fee for all my music needs.
The object of mankind lies in its highest individuals.
One must have chaos in oneself to be able to give birth to a dancing star.

Future of digital music distribution.

Reply #5
May I ask you, atici, where do you usualy get info about new music, players, groups, etc.?
keeping audio clear together - soundexpert.org

Future of digital music distribution.

Reply #6
Quote
May I ask you, atici, where do you usualy get info about new music, players, groups, etc.?
[a href="index.php?act=findpost&pid=373593"][{POST_SNAPBACK}][/a]

How is this related to the discussion? Well for the new albums I buy, I learn them from friends / reading reviews from various sites / through other music genome sites such as last.fm or Pandora.com. What exactly do you mean by "players"? Digital audio players or artists?

BTW I listen to classical, jazz or ethnic music only -- as there's not much genius elsewhere. I think the answer would probably depend a lot from person to person based on taste.
The object of mankind lies in its highest individuals.
One must have chaos in oneself to be able to give birth to a dancing star.

Future of digital music distribution.

Reply #7
I don't see why net distribution in the future would only use lossy codecs.  Honestly I think it's bad enough that most CD's are mastered so poorly, and then to degrade the quality more with a lossy codec and then sell it for the same price as the CD sounds like robbery to me.  If people are fine paying for inferior audio (buying an mp3 versus the actual CD) then so be it.  I have and never will pay for lossy audio.  Of course masters may be of higher bitrates and such than CD's, but that dosn't mean CD's are lossy, just different sample rates/bitrates.

With the increase of internet speeds & hard drive sizes lossless is not only reasonable now, but will only become more so in future.  The introduction of DVD-A and SACD doesn't really undermine this simply because most people haven't bought into the technology.  Several releases appear to be simply CD audio resampled and stuck on a DVD.  SACD's can only be listened to on SACD players.  Consumers will always want and need portability.  Lossy will be around for some time for that purpose I think.  But Apple and the DRM nuts (IMHO) are stealing from everyone who shops with them.  It's a (mostly) free world though, so I can't complain.  Also about DVD-A and SACD; as best I can tell the only real audible differences seem to come from different mastering, or 2ch versus 6ch.  I've only listened to other peoples though, as I don't own any.
What do I know though?  Every codec has it's place.  I just see lossless becoming more popular in the future.  To each his own though.  Some will always want their mp3's, others their oggs, and others thier flacs. 

Future of digital music distribution.

Reply #8
Quote
How is this related to the discussion? Well for the new albums I buy, I learn them from friends / reading reviews from various sites / through other music genome sites such as last.fm or Pandora.com. [a href="index.php?act=findpost&pid=373611"][{POST_SNAPBACK}][/a]

Are you sure that the “on-demand service”, you’ll be subscribed, will be able to offer you all the music you’ve learned from frends/reviws/sites? Especially taking into account that music market tends to be broader and more varied.  I think searching in the internet for some music you like is not boring or inconvenient. It could be great creative fun.
keeping audio clear together - soundexpert.org

Future of digital music distribution.

Reply #9
Quote
I don't see why net distribution in the future would only use lossy codecs. [a href="index.php?act=findpost&pid=373621"][{POST_SNAPBACK}][/a]

Simply because lossy compression can provide better than CD quality if compressing directly from studio masters (with file sizes compared to losslessly compressed CDs and even less). I mean conversion to 16@44.1 is not necessary any more.
Quote
Of course masters may be of higher bitrates and such than CD's, but that dosn't mean CD's are lossy, just different sample rates/bitrates.[a href="index.php?act=findpost&pid=373621"][{POST_SNAPBACK}][/a]

That exactly means that CDs are lossy because bit/sample rate conversion is irreversible or lossy. Try to downsample CD audio to 8bit@22050Hz – the difference is shocking. Of course, 24@96 to 16@44.1 conversion is not so audible but it’s definitely lossy.
keeping audio clear together - soundexpert.org

Future of digital music distribution.

Reply #10
Quote
Quote
I don't see why net distribution in the future would only use lossy codecs. [a href="index.php?act=findpost&pid=373621"][{POST_SNAPBACK}][/a]

Simply because lossy compression can provide better than CD quality if compressing directly from studio masters (with file sizes compared to losslessly compressed CDs and even less). I mean conversion to 16@44.1 is not necessary any more.
Quote
Of course masters may be of higher bitrates and such than CD's, but that dosn't mean CD's are lossy, just different sample rates/bitrates.[a href="index.php?act=findpost&pid=373621"][{POST_SNAPBACK}][/a]

That exactly means that CDs are lossy because bit/sample rate conversion is irreversible or lossy. Try to downsample CD audio to 8bit@22050Hz – the difference is shocking. Of course, 24@96 to 16@44.1 conversion is not so audible but it’s definitely lossy.
[a href="index.php?act=findpost&pid=373656"][{POST_SNAPBACK}][/a]


Down-converting to 44.1/16 will not have any audible consequence if done properly. Doing so properly is no issue; even common software such as Adobe Audition can do so properly with the right settings. In fact, anything over 44.1/16 this is a waste of file space. Sample/bit rates exceeding 44.1/16 do nothing to enhance the audible quality according to known credible perceptual research.

-Chris

Future of digital music distribution.

Reply #11
Quote
Down-converting to 44.1/16 will not have any audible consequence if done properly. Doing so properly is no issue; even common software such as Adobe Audition can do so properly with the right settings. In fact, anything over 44.1/16 this is a waste of file space.[a href="index.php?act=findpost&pid=373696"][{POST_SNAPBACK}][/a]

Could you specify at least one reason for 24/96 to 16/44.1 conversion? Instead you could lossy compress the original directly saving more quality and file space.

Quote
Sample/bit rates exceeding 44.1/16 do nothing to enhance the audible quality according to known credible perceptual research.[a href="index.php?act=findpost&pid=373696"][{POST_SNAPBACK}][/a]
Could you name some?
keeping audio clear together - soundexpert.org

Future of digital music distribution.

Reply #12
Why would you keep 96 kHz while a lossy encoder probably will do a lowpass way beneath 48 kHz?

There are several threads about why such high sampling rates and bitdepths are not necessary. Here is one.
"We cannot win against obsession. They care, we don't. They win."

Future of digital music distribution.

Reply #13
Quote
[
Could you specify at least one reason for 24/96 to 16/44.1 conversion? Instead you could lossy compress the original directly saving more quality and file space.


Loss compression is by definition 'lossy'.  I don't understand how you will save more quality when the 16/44.1 contains all that is known to be important. The 16/44.1 does not exclude anything of relevance to the human auditory system. There is no known reason for 24/96 for playback purposes. If you find credible research concluding that it is needed, via valid perceptual tests, please let me know.

Quote
Could you name some?


Do a search on this forum and you shall find the relevant threads that are full of useful information pertaining to these issues. I have provided such references many times in the past, but it will be more informative to you if you read these threads in their entirety rather than doing the same thing all over again. It would be preferable if you would ask questions as follow-ups to those threads.

-Chris

Future of digital music distribution.

Reply #14
Quote
Why would you keep 96 kHz while a lossy encoder probably will do a lowpass way beneath 48 kHz?

There are several threads about why such high sampling rates and bitdepths are not necessary. Here is one.
[a href="index.php?act=findpost&pid=374060"][{POST_SNAPBACK}][/a]

And why not, if audio content is initially in this format. How do you think what conversion chain is better:
1.   24/96 => 16/44.1 => lossless
2.   24/96 => lossy
As you know there are some technical advantages of high definition format concerning mastering and Digital-to-Analog conversion. And yes, advantages for human perception are very questionable.
keeping audio clear together - soundexpert.org

Future of digital music distribution.

Reply #15
Quote
And why not, if audio content is initially in this format. How do you think what conversion chain is better:
1.   24/96 => 16/44.1 => lossless
2.   24/96 => lossy
As you know there are some technical advantages of high definition format concerning mastering and Digital-to-Analog conversion. And yes, advantages for human perception are very questionable.
[a href="index.php?act=findpost&pid=374076"][{POST_SNAPBACK}][/a]

Interesting, this is a question I'm thinking about several years. I'm happy to see it pop up here.

Future of digital music distribution.

Reply #16
Quote
And why not, if audio content is initially in this format. How do you think what conversion chain is better:
1.    24/96 => 16/44.1 => lossless
2.    24/96 => lossy
As you know there are some technical advantages of high definition format concerning mastering and Digital-to-Analog conversion. And yes, advantages for human perception are very questionable.
[a href="index.php?act=findpost&pid=374076"][{POST_SNAPBACK}][/a]

If bitrates are roughly equal, I would definitely go for door number one.

While resampling is, well, lossy (or destructive  ), 16/44.1 is more than enough for me, and I don't consider resampling nearly as dangerous as the loss that could be introduced by a lossy codec. In the end, both will probably sound transparent though...

And seriously, you want to use lossy encoded 24/96 for editing and mastering purposes? 


[edit: i have spellbound, so why dont i use it?/]
"We cannot win against obsession. They care, we don't. They win."

Future of digital music distribution.

Reply #17
Quote
If bitrates are roughly equal, I would definitely go for door number one. [a href="index.php?act=findpost&pid=374176"][{POST_SNAPBACK}][/a]
The same as for me except may be 16/48 instead of 16/44.1 as more natural and easier for resampling itself. But what if the second door provides lower bitrate and higher perceptual quality?
Quote
While resampling is, well, lossy (or destructive  ), 16/44.1 is more than enough for me ...[a href="index.php?act=findpost&pid=374176"][{POST_SNAPBACK}][/a]
... but not for audiophiles. And I think if they come to the internet new authors and music publishers could receive additional investments. Now all that money is spent entirely on marketing departments of “hi-end equipment & accessories” manufacturers and music production of big labels.
Quote
..., and I don't consider resampling nearly as dangerous as the loss that could be introduced by a lossy codec. In the end, both will probably sound transparent though...[a href="index.php?act=findpost&pid=374176"][{POST_SNAPBACK}][/a]
This can be tested.
Quote
And seriously, you want to use lossy encoded 24/96 for editing and mastering purposes?  [a href="index.php?act=findpost&pid=374176"][{POST_SNAPBACK}][/a]
See p.2 of the first post.
keeping audio clear together - soundexpert.org

Future of digital music distribution.

Reply #18
I don't think it is possible that lossy compressed 24/96 will sound better than properly resampled 16/44.1 (or 16/48 if you wish, but you didn't give that as option ). I'm not a dog, I can't hear frequencies around 30kHz. Actually I can't hear anything above ~17 kHz.

24/96 would just be a marketing trick (more is always better, right?) which I'm not willing to pay for. I'm really not interested in the useless things that so-called "audiophiles" want to spend to their money on. Mentioning the word "audiophile", I doubt an "audiophile" would be satisfied with anything less than uncompressed 32/192. That, or vinyl.
"We cannot win against obsession. They care, we don't. They win."

Future of digital music distribution.

Reply #19
Quote
I don't think it is possible that lossy compressed 24/96 will sound better than properly resampled 16/44.1 (or 16/48 if you wish, but you didn't give that as option ).[a href="index.php?act=findpost&pid=374195"][{POST_SNAPBACK}][/a]
Ok. Would lossy 24/96 be attractive for you if it sounds the same as lossless 16/48 but saves more file space?
Quote
Mentioning the word "audiophile", I doubt an "audiophile" would be satisfied with anything less than uncompressed 32/192. That, or vinyl. [a href="index.php?act=findpost&pid=374195"][{POST_SNAPBACK}][/a]
Actually they are not so stupid. They are passionate I would say
keeping audio clear together - soundexpert.org

Future of digital music distribution.

Reply #20
Resampling is not harmless and is definitely lossy (you can’t get the original content back).

Also even if for playback it’s not really necessary, it is interesting to have a bigger resolution for processing and processing may occur during the playback (digital crossover, resampling, frequency correction, …). Also DAC is easier to make properly with a higher sampling rate because the anti aliasing filter can occur at a higher frequency (creating less phase issue).

Concerning the bit depth, some codec use 32 bit float internally (mp3, …). While in 16 bit you should apply a dithering before the encoding, you should encode the noise introduce by the dithering (making the encoding process less optimised) and you will probably apply a second dithering after the decoding (if you use foobar for example). Encoding from a higher bit depth should avoid the dithering before the encoding and let the dithering at the very last stage of the playback, which is recommended.

Future of digital music distribution.

Reply #21
Quote
Resampling is not harmless and is definitely lossy (you can’t get the original content back).

No one is saying that resampling isn't lossy. But I sincerly doubt anyone can hear the difference between 24/96 and 16/44.1 resampled (if done properly).

Quote
Ok. Would lossy 24/96 be attractive for you if it sounds the same as lossless 16/48 but saves more file space?

Smaller with same quality is always ok, better quality with same size is always ok.
"We cannot win against obsession. They care, we don't. They win."

Future of digital music distribution.

Reply #22
Quote
Resampling is not harmless and is definitely lossy (you can’t get the original content back).


While it is *lossy* by technicality, it is not known to be *lossy* audibly, since the only thing lost is audibly irrelevant information. A good re-sampler does not produce artifacts of any significance; the artifacts from a good re-sampler are difficult to measure at all, excepting the frequency response difference, of course.

Quote
Also DAC is easier to make properly with a higher sampling rate because the anti aliasing filter can occur at a higher frequency (creating less phase issue).


What phase issue? A linear phase FIR filter has no such issue, and even if the out dated analog type filter were used that did have phase shift that is commonly touted as being why 44.1 is not sufficient, credible and peer-reviewed perceptual [1]research does not demonstrate the degree of phase shift present in such an analog filter to be of audible significance.

-Chris

[1]Perception of Phase Distortion in Anti-Alias Filters
D. Preis and P. J. Bloom (1984)
JAES, vol 32, number 11, p842, November 1984

Future of digital music distribution.

Reply #23
A reminder:
The human perception is limited somehow around 20 bit/44.1-48 kHz, so that 24/48 is already overkill, or on the safe side so to say.
24/96 or above, well already discussed here lengthy.
Scientific citing about these limits you find at Musikhochschule Detmold / Tonmeister (sound ingeneers) studies/diploma works which are published partly at their websites. of course, they were referencing to other studies/papers regarding these topics. I think, if you search here at HA by detmold , you will find the links, or try google.

 

Future of digital music distribution.

Reply #24
Quote
A reminder:
The human perception is limited somehow around 20 bit/44.1-48 kHz, so that 24/48 is


Of course, the 20 bits is overkill for the actual dynamic range found in even the most dynamic classical music(excepting perhaps a special variant of 1812 Overture, if one wanted to actually record a canon at close to mid range, as opposed to the far range perspective of the current versions; this would also result in auditory pain if the speakers could actually reproduce such SPL), and almost no speakers(this includes audiophile high dollar speakers), excepting very large horns or line array systems, can possibly use even 16 bits to it's potential.

-Chris