Before sigma-delta DACs/ADCs became common, most DACs and ADCs would often work at (e.g.) 4x the sampling rate internally. So a DAC would take your 48 kHz signal, up-sampling it (digitally) to 192 kHz internally, and then convert to analog. At this point, the analog filtering is indeed trivial because there's no signal between 24 kHz and 168 kHz. The same idea applies to ADCs. The bottom line is that you get the benefits of 192 kHz sampling (simpler analog filters) without the pain of having to store all those samples.
Why 24/192 is a bad idea?
- Thread starter audioguy
- Start date
Before sigma-delta DACs/ADCs became common, most DACs and ADCs would often work at (e.g.) 4x the sampling rate internally. So a DAC would take your 48 kHz signal, up-sampling it (digitally) to 192 kHz internally, and then convert to analog. At this point, the analog filtering is indeed trivial because there's no signal between 24 kHz and 168 kHz. The same idea applies to ADCs. The bottom line is that you get the benefits of 192 kHz sampling (simpler analog filters) without the pain of having to store all those samples.
I surely do not have the credentials, nor the engineering prowess of some of our esteemed members (and I am honored to be able to be a part of this community) . But one thing that escapes me is this, "Although the higher sampling rates encode audio outside of the human hearing range, the net effect of higher sampling rate and conversion technology improves the audio quality within the ideal range of human hearing."
So, more bits and pieces that are beyond my hearing, will in effect improve the sound that I cannot hear? The more I read from numerous forums, mainly Computer audiophile, the more I believe in 16/44.1. Almost everyone is so focused and concerned on the highest this and the highest that, and what is new, that they have forgotten that it is how you build something that is most important. We have not even begun to explore all that 16/44.1 can offer. I will take a SOTA CDP that is old tech. over some of this cheap, disposable, computer driven crap that is market driven. If you do not like the sound of 16/44.1, then you are listening to either inferior players, or inferior recordings. I really LOL, when I read something from a recording engineer hat sniffs he can "only" accept 24/192. To me, that says, he can only drive fast if he has a 1,000 hp car.
So, more bits and pieces that are beyond my hearing, will in effect improve the sound that I cannot hear? The more I read from numerous forums, mainly Computer audiophile, the more I believe in 16/44.1. Almost everyone is so focused and concerned on the highest this and the highest that, and what is new, that they have forgotten that it is how you build something that is most important. We have not even begun to explore all that 16/44.1 can offer. I will take a SOTA CDP that is old tech. over some of this cheap, disposable, computer driven crap that is market driven. If you do not like the sound of 16/44.1, then you are listening to either inferior players, or inferior recordings. I really LOL, when I read something from a recording engineer hat sniffs he can "only" accept 24/192. To me, that says, he can only drive fast if he has a 1,000 hp car.
I have asked the question twice and seems to be impossible to get you to share your opinion with us. So let's see what is written on this topic. First, from your site: http://www.vorbis.com/faq/
"Vorbis is patent and license-free, so you will never need to pay anyone in order to sell, give away, or stream your own music."
You honestly think that if I implemented Vorbis in my device, someone could not come knocking on my door, tell me I am violating their core IP which you step on and ask for royalties? If so, will you personally indemnify me and pay for any and all damages per your FAQ? Or will I be holding the bag instead?
Here is the Wiki page on this: http://en.wikipedia.org/wiki/Ogg_vorbis
"The Xiph.Org Foundation states that Vorbis, like all its developments, is completely free from the licensing or patent issues raised by other proprietary formats such as MP3. Although the Xiph.Org Foundation states it has conducted a patent search that supports its claims, outside parties (notably engineers working on rival formats) have expressed doubt that Vorbis is free of patented technology.[47]
The Xiph.Org Foundation has not released an official statement on the patent status of Vorbis, pointing out that such a statement is technically impossible due to the number and scope of patents in existence and the questionable validity of many of them. Such issues cannot be resolved outside of a court of law."
So which is right? That you can never know as in the Wiki or that you do know that I won't have to pay licensing fees to anyone per your FAQ?
As I noted in my comments on AVS, the whole debate is about nothing. Few people in the studio are capturing and processing 16bit at 44.1 Khz. It is always something higher. We used to have to downconvert this to 16/44.1 so that it would comply with CD spec. With digital distribution we no longer have to do that. I say give the original "masters" to the consumers. If they think they need nothing more than 16/44.1, they can do the conversion themselves. For others, they can keep it as the original.
Here is the nice thing about that. The high resolution master is usually free of the crazy rush to heavily compress the audio on its way to the CD. So you solve the "loudness war" while you are at it.
I think these are very compelling arguments to give out high res sources which have nothing to do with the debates around sample resolution and rates. Why would we want to put so much energy to stop such transition?
The patent problem is the same for everyone, whether you're writing free codecs like Xiph.Org does or an encumbered one like MP3 and (AFAIK) WMA. In all cases, there's a risk (that you want to minimize) of having to go to court about a patent. When writing an encumbered codec, you "invent" a bunch of new algorithms and patent them. You also have a look at other people's algorithm and decide whether you want to use them and that decision is based on whether the algorithm is patented and if so, whether you're willing to license the patent.
In the case of free codecs, there's two minor differences. First you don't patent what you invent (or in the case of Opus, you patent it, but give a free license). For other people's algorithm, you just decide that you're not willing to license anything patented algorithm.
The risk -- for both encumbered and unencumbered codecs -- is that you missed something. And you obviously want to minimize that. So far, I think Xiph's done pretty well on that front, considering the number of free codecs under the Xiph umbrella and the number of patent lawsuits (zero).
Last think, who will indemnify you if you get sued for patent infringement on MP3 or WMA?
Not internally, to my knowledge. They would rely on an external digital filter. SAA7220 in the case of TDA1541(A) for example.
It depends on the performance of the digital filter - the SAA7220 for example put out a fair bit in its transition band before settling down to below -50dB or so over that band. If sound quality (rather than measurements) is an issue then no analog filter design could accurately be called 'trivial'
So the game of Calvinball continues
. First you say " His bibliography isn't any thicker (most of it references other papers he wrote himself)." Now "most" becomes 11/27? The list is who is who in audio and signal processing: Professor Hawksford, Jullian Dunn, Vanderkooy, Lipshitz, on and on. How many should he have put in there to be credible there in your opinion? I counted 13 AES paper references. What should it have been? How many AES papers did you cite?His "core premise" is similar to yours. In that sense I find your negative views of his work illogical.
But again, I am very comfortable with what you stated in your article: that he wrote what you wanted to write. Unless you are taking that back, I think we are done with this point.
Not really. If you use WMA on Windows, you get Microsoft to defend you. And it comes with the backing of their patent portfolio and assets. You bring none of that. I will be on my own, defending myself and paying out of my pocket.
If there is a risk, then I suggest changing the FAQ to reflect that.
That is not how it works but it is not worth arguing as their technology is not on stage here. Nor are they shouting to the world that their codec is free to use.
Problem is, you don't know all that has been patented. And this field is covered with patents. Any patent search in the field of audio compression is going to pop up countless claims. You will be dead, buried and broke by the time you investigated all of them
. I am pretty sure you didn't do this anyway when you started. The main reason for lack of attention is that the big boys haven't licensed it. Companies like Microsoft, Apple and such are targets people want to go and sue. An implementation here and there doesn't represent enough damages or attack surface for someone.
I answered for WMA. For MP3 if you are on Windows and are one of Microsoft's OEMs, Microsoft will again defend you. Here is an example: http://en.wikipedia.org/wiki/Alcatel-Lucent_v._Microsoft
"Lucent Technologies Inc. v. Gateway Inc. 470 F.Supp.2d 1180 (S.D.Cal.,2007) is a patent case between Alcatel-Lucent and Microsoft litigated in the United States District Court for the Southern District of California and appealed to the United States Court of Appeals for the Federal Circuit. The litigation money involved was $1.53 billion for the first patent infringement case. Its final verdict was given in August 2007 at the U.S. District Court for the Southern District of California in San Diego. The damages award was reversed on appeal in September 2009 and returned for a separate trial on the amount of damages.[1]"
FHG/Thomson have a very strong position in MP3 patent portfolio yet that was still not enough to fend off Lucent. Fortunately Microsoft managed to reverse that verdict. It is an ominous warning to people who naively think just because they have looked to the left and to the right and didn't see someone with a patent on their technology, that they should sleep easy
. There are a lot of bad guys in this space. You have been fortunate that you have not had to meet them . Yet!
You've asked an entirely different, rather insulting question twice. But since you've now asked a new question on which I have plenty to say:
Short answer: Yes. No one could practically get away with it, though someone might think about it, maybe even try before getting laughed out of court.
nuanced, tl;dr version: A party can sue anyone for anything. I could sue you today for killing my cat, even if I never owned a cat. Barratry is not illegal in the United States; this is part of why the patent system is becoming dysfunctional. Xiph commented in detail on this problem in an advisory whitepaper paper to the FTC: http://xiph.org/press/2011/ftc/ftc-comments-20110614.pdf
Vorbis itself was searched and vetted in 1999/2000 by Ed Kelly at the direction of iCast (the primary sponsors of the Ogg project before Xiph incorporated). He issued a formal opinion of non-infringement. This vetting process has since been repeated, in my personal direct knowledge, by Mozilla and Google... and apparently Microsoft: They've been shipping Vorbis in games since 2000. Several changes we made to libvorbis were at Microsoft's request.
Vorbis is installed on most desktop PCs in the field today, owns the gaming market, a large fraction of mobile and embedded devices, and more than 100 makes of portable player. It is possibly the most widely deployed codec in history (if you count all applications, not just music). There have been no lawsuits.
No. It is against standard industry practice. Ask MPEG-LA about indemnification some time.
I'll take our own legal counsel and a 14 year track record over a Wikipedia article.
That's correct. Patent law has no provision for the legally binding statement you're demanding. Surely you knew that before you asked the question.
Monty
Xiph.Org
Oh, and "I am not a licensed attorney; none of the above should be treated as legal advice or counsel."
Just curious, where's the public statement that Microsoft will do that if you get sued?
Would this be enough:
WARNING: IF YOU MAKE, DISTRIBUTE, OR USE ANY KIN OF SOFTWARE FOR ANY PURPOSE, THERE'S A CHANCE OF BEING SUED FOR PATENT INFRINGEMENT.
Problem is, you don't know all that has been patented. And this field is covered with patents. Any patent search in the field of audio compression is going to pop up countless claims. You will be dead, buried and broke by the time you investigated all of them
I'm also pretty sure Microsoft did not investigate every codec patents either.
Right, it's not like Microsoft has pockets that are that deep anyway...
Yes, Microsoft currently ships not only Vorbis, but also Speex and (soon) Opus. Google also ships at least Vorbis and Speex. Apple and Adobe ships Speex, and so on. Of course, these are all tiny companies that aren't worth suing, as opposed to...
You are missing my point, or perhaps my point is not clearly stated. I don't care about whether music is recorded in 16/44.1 or 24/192. I am intrigued by what the hardware engineer can do to minimize the recording limitations. I have more respect of a designer who makes musical audio gear that can turn a sow's ear into a sumptuous feast, rather than someone who can turn filet mignon into well a done hamburger. Who cares what you process in the studio, as long as it is done with quality of the recording process. A lot of recording engineers need more rates and bits is to make up for their poor recording skills. I love it when the hardware guys can make that shite sound great. To get back to the original thread topic, no I don't think 24/192 is bad. I don't think 16/44.1 is bad either. It is what you do with it.
All of this technical talk is fascinating, really, but I'm a bit short on time. Can anyone point me to the scholarly paper that details the discovery that middle-aged men with expensive audio toys can now hear beyond 20khz?
Tim
Tim
This seems logical, but I don't want to down load a huge file for no good reason, I don't want to convert a huge file for no good reason, and hi res is not spared the pain of "loudness" because it is hi-res, but because of the niche market it serves. The moment it becomes the standard, it will suffer the indignities of the mass market product.
You're getting close, though. We, and Neil Young, should be demanding better recording and mastering of the better music (let them compress pop tarts if they must) not formats we're physiologically incapable of hearing.
Tim
Personally I feel this all comes back to how it is implemented and importantly differentiating between the different processes (studio or DAC at home), hardware related coding such as seen inside the Sabre chipset and oversampling chips, and software algorithms relating to the filters.
The reason I am saying this is that in the real world we have measurements showing poor implementation, or that the music is not are a true representation of hi-res.
There are plently of examples showing poor upsampled/downsampled music files available for purchase and others that are perfect true/"native" 24/192.
Then again there are examples where the clocking or receiver chips are poorly implemented for OS-upsampling that then gave poor results, however other products show near perfect results.
Some chipsets have certain trends or potential challenges to overcome as again can be seen in measurements such as that from Hifi News, also a trend is for distortion in some DAC chips become worse higher the FR above 20khz or have a distinct trend at 0 to -20dbfs.
I remember quite awhile ago John Westlake when discussing the ongoing development of the Audiolab DAC-CD player identifying a low/hardware level issue within the ESS Sabre chipset that he decided to develop a workaround for - sorry it would take forever to find the specific post as those forum threads reached many 100s of pages.
So it comes back to implementation and what one is talking about or focusing on technology-architecture wise in such discussions as these.
IMO 24/192 is a good idea for the reason I mentioned before, especially as native-studio recorded music and not upsampled-downsampled that usually measure with more anomolies due to implementation and critically the process.
This is why IMO the engineer/s involved are essential, and why I mention people such as John Westlake (along with his colleague) often as their knowledge even down to low level coding-hardware is exceptional, or those heavily involved in developing measurements such as Paul Miller and Keith Howard.
So IMO real world examples showing 24/192 is a bad idea are down to poor implementations-processes-architecture, and these are seen in the measurements that also highlight why there are benefits to native 24/96 and higher from a technical and engineering perspective - discussing audible and perceived benefits are pointless and will just go round in circles without being conclusive to all.
Cheers
Orb
The reason I am saying this is that in the real world we have measurements showing poor implementation, or that the music is not are a true representation of hi-res.
There are plently of examples showing poor upsampled/downsampled music files available for purchase and others that are perfect true/"native" 24/192.
Then again there are examples where the clocking or receiver chips are poorly implemented for OS-upsampling that then gave poor results, however other products show near perfect results.
Some chipsets have certain trends or potential challenges to overcome as again can be seen in measurements such as that from Hifi News, also a trend is for distortion in some DAC chips become worse higher the FR above 20khz or have a distinct trend at 0 to -20dbfs.
I remember quite awhile ago John Westlake when discussing the ongoing development of the Audiolab DAC-CD player identifying a low/hardware level issue within the ESS Sabre chipset that he decided to develop a workaround for - sorry it would take forever to find the specific post as those forum threads reached many 100s of pages.
So it comes back to implementation and what one is talking about or focusing on technology-architecture wise in such discussions as these.
IMO 24/192 is a good idea for the reason I mentioned before, especially as native-studio recorded music and not upsampled-downsampled that usually measure with more anomolies due to implementation and critically the process.
This is why IMO the engineer/s involved are essential, and why I mention people such as John Westlake (along with his colleague) often as their knowledge even down to low level coding-hardware is exceptional, or those heavily involved in developing measurements such as Paul Miller and Keith Howard.
So IMO real world examples showing 24/192 is a bad idea are down to poor implementations-processes-architecture, and these are seen in the measurements that also highlight why there are benefits to native 24/96 and higher from a technical and engineering perspective - discussing audible and perceived benefits are pointless and will just go round in circles without being conclusive to all.
Cheers
Orb
Good or bad, unless your business is out united states, you have to abide by the laws of this country. The day I get sued, my defense cannot be, "Judge, people get sued all the time. Please let me off the hook." Typical patent litigation costs which includes a jury (as it often does when a big company is sued), runs into millions of dollars. This is *before* any damages that may have to be paid. Yes, companies do leverage their patents, sometimes unfairly against some parties. Part of the solution to it is to have RAND licensing: reasonable and non-discriminatory terms. This means that you get charged what everyone else gets charged. This is a condition of participation in many standards groups. So if you use one of these standards, you know you at least can get a reasonable license.
Not so with Vorbis and the rest of the so called "open source" technologies. Since no major patent holder is consulted or paid for their use, nor are there any RAND terms in play, when the day of reckoning comes, the costs may easily put you out of business.
Look at the mess Google is in with Android. They had to pay $12.5 *billion* dollars to buy Motorola's cell phone division just to get their patent portfolio. And the major bidding war to buy patent portfolios before that. They did that because unlike you, they are not just hoping the FTC will bail them out or complaining about unfairness of the US law. Those will be reasonable defenses to make, especially against foreign companies but they won't be sufficient as anyone who has had to look at the Apple, Oracle and Microsoft attorneys in the eye can attest.
To be fair, I absolutely share your sentiments that this is a broken system when it comes to allowing innovation at some level. But as I said, we live in this country and we have to live by its laws. And therefore, you work to get yourself better situated. While we were developing our own compression algorithms at Microsoft, we also contributed to major standards and sat at those tables to establish their rates. By making sure those were reasonable, then comparative solutions otherwise would potentially cost less to license. In that manner, we helped you without you even knowing it!
.That is good in that you won't be accused of willful infringement and subject to treble damages. But will amount to nothing the day I am sued and have to defend myself without you or that attorney defending me for free.
Microsoft games <> Microsoft. Please don't confuse the two. When I was there, my group would constantly be educating people who are confused by marketing collateral such as your FAQ thinking open source = patent free. Sometimes they would be pretty far down the road before we would stop them and remind them of the huge liability they were creating for the company. I don't know who does that or how careful they are since I left four years ago.
The best test of any of this is whether you personally stand behind anyone getting sued. It is clear you are not.
First of all, you would not know about all the "lawsuits." It is not like you are a licensing and provide an indemnity and hence will be called on every one of them. For all you know, folks could have gotten threatening letters and settled. Your contention that most desktop PCs ship with it is also substantiated. Windows doesn't ship with it and neither does the Mac OS. So you are at the mercy of some app the OEM may have installed. I doubt that there are many of those that include Vobis. As to 100 makers of portable players, that is nothing compared to AAC, MP3 and WMA whose device counts is into thousands. But yes, you have achieved some success with it and in that regard, are protective of that franchise. I get that. But you can have all the people in the world shipping it but it does not matter the day *I* get sued. The fact that five other guys shoplifted, doesn't become my defense for shoplifting.
It is not against industry practice. There are technologies that are provided with indemnity. Or equiv. thereof. Have your attorney explain what "non-assert" means. This is a provision where if you are a licensee of a technology as such, you cannot sue other licensees (non-assertion against other licensees). This means that if a standard becomes very popular, the universe of people who can sue you becomes extremely small. It essentially takes out all the major companies out of the running and leaving you with so called "patent terrorists" to come and after you. These are individuals/law firms that buy patents and try to monetize them by holding you hostage until you pay.
There is no non-assert for open source. You are given something for "free" but in return you shoulder all the risks. In many areas, such as generic software, that is very manageable. Not so in signal processing where the way every electron moves is often patented! Or claim to be patented. I have lived this world. You say you have never been sued or know about a suit. Well, then you don't have the education. You did not work for a big company that was a popular target so had to come up with solutions and strategies to deal with these things.
You should but the fact that Wiki says what it says, and you are aware of it and its references to detailed discussions on patents that Vorbis may be stepping on, may create some issues with respect to the non-willful infringement defense. So you may want to check with your attorney on that. If you have and he is OK with it, then that is cool.
I am not an attorney but if I got sued and had shipped the software based on assurances you provided in your FAQ that I would never ever have to pay anyone, if you had any assets, I might try to pull you in to help share in my costs.
It won't be. But if I were you, I would not engage in any such discussions. I would get what my attorney wants me to say, which may turn out to be nothing, and leave it at that. Indeed, I feel bad that I have gotten this deep into the topic as I honestly do not want any harm to come to you or your technology. All I wanted to know was that if my impression of the old you where you thought open source meant litigation free was still true. I have that answer. You are much more sober and realistic now
.Then the publisher can, as they do today, provide you with both versions. The server cost of storage is nothing in this day and age.
Loudness war is not due to 16 bits being too little. It's due to a "louder is better" attitude. No matter how many bits you have, these folks will still make sure they mix the CD as loud as the format allows it.
Well, why stop at 192 kHz/24 bit? Hey, let's all distribute 64-bit digital audio sampled at 1 MHz! Why would we want to put so much energy to stop audio going into the MHz?
True but unrelated to the point I made. When the music is being mixed it is not subject to loudness compression. It is only when it is being mastered for CD that the atrocities are committed
. Here is an example of why this is true: http://mixonline.com/mixline/reierson_loudness_war_0802/"The Loudness War is Over
Feb 8, 2011 2:22 PM, By Greg Reierson
Making loud CDs will become just a bad memory.
I was at the AES show in San Francisco last November and I came back with renewed hope for the future of the music industry—not just from a business perspective, but from a recording-quality perspective as well. Besides the usual discussions about gear and recording techniques, there was a lot of talk about high resolution digital downloads surpassing CDs as the dominant delivery format within the next few years. Optimism is growing as more and more engineers are seeing a way to finally get past the loudness war.
Greg Reierson is the owner/chief engineer at Rare Form Mastering in Minneapolis. Visit him at www.rareformmastering.com. "
I am not stopping at anything as that is not my point. Whatever pros produce is what we should give to the consumer and with it, bypass this silly argument. I would be hearing the same signal that powered their electronics/speakers that way. Anything else and I am at the mercy of them understanding things like dither, resampling, etc. And worst of all, compressing the heck out of it.
I could. Don't think I've ever been able to tell the difference, but most of my collection came from CD, and hard drive space is cheap enough that lossless makes sense. 24/192 is much larger, but still the HD space is not the issue as much as the download time.
Tim
Similar threads
