AVS Forum banner

1 - 20 of 36 Posts

·
Premium Member
Joined
·
3,258 Posts
Discussion Starter · #1 ·

Grammy-winning recording engineer and speaker designer Allen Sides shares some great stories about artists he's recorded in his Ocean Way studio and talks about his preference for high digital-audio sample rates and analog equalization, Neil Young's high-resolution-audio Pono project, the negative effects of audio-data compression, dynamic-range compression, multichannel movie and music recordings, speaker directivity and time alignment, the differences between professional and consumer speakers, answers to chat-room questions, and more.

 

 

·
Registered
Joined
·
8,204 Posts
Lots of good info in this episode. I never would have thought that CD sound could be effected by it's pressing but once it was explained it made perfect sense. Scott please have Allen Sides back sometime in the future.
 

·
Registered
Joined
·
735 Posts
A great episode Scott, but when Allen said he has been unable to copy a CD and have the copy sound as good as the original requires much discussion. it goes against everything I've heard or read.

Never mind ripping a CD to wav or flac how can that ever work if you cant even copy a CD. Would love a panel to come on the show with Allen to discuss his position.
 

·
Registered
Joined
·
331 Posts
Wow. Just fantastic. Sounds like this guy could talk audio for hours! Great to have somebody say that something may be difficult to measure, but yet they're very , clearly audible.

I've always felt this way, but you always get shot down by the "snake oil" argument.. so to have someone with such experience and credibility say that, is helpful.


Please have him back in the future! Loved his enthusiasm too.. with me pursuing at audiophile greatness at a young age, once he said he started selling speakers at 16, I was like "Man, I gotta pick up the pace!"

Thanks Scott, well done.
 

·
Registered
Joined
·
1,011 Posts

Quote:
Originally Posted by dargo  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides#post_24543379


A great episode Scott, but when Allen said he has been unable to copy a CD and have the copy sound as good as the original requires much discussion. it goes against everything I've heard or read.

That's simple dargo. Everything that you have heard or read is wrong!
 

·
Registered
Joined
·
50 Posts

Quote:
Originally Posted by Player3  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides/0_100#post_24543549


SNIP... Wow. Just fantastic. Sounds like this guy could talk audio for hours! Great to have somebody say that something may be difficult to measure, but yet they're very , clearly audible.

I've always felt this way, but you always get shot down by the "snake oil" argument.. so to have someone with such experience and credibility say that, is helpful.
It certainly is! This has been my experience transferring audiophile quality LP's to CD and giving up due to their poor quality and later starting again transferring to DVD-A at various sampling rates and finding the 24-bit/192kHz recordings via diskWelder BRONZE was the answer, producing great DVD-A discs, worthy of the original quality of my old audiophile LPs. Even a few friends could hear the difference. Also agree with his comment about about Blu-ray audio, as I have a number of SONO Luninus Blu-ray audio only discs which I've been buying and play at 5.1 DTS-HD 24-bit/192kHz and they are really stunning. Great show, please have him back to share more with us.

-Rod
 

·
Registered
Joined
·
12 Posts
Scott,

Really enjoyable interview. Two things I would have liked to ask are:


1) What does Mr Sides think of binaural recording? With the great mike collection he has, I would think it would lend itself to a very interesting production.


2) What movie multi-channel audio format does Mr Sides prefer since it sounded like he was down on both DTS and Dolby Digital?


Also, he mentioned with the advent of Blu-Ray there was no need to compress movie audio. How could one tell if the audio was compressed?


I agree with others you should have him back to discuss more, very engaging and knowledgeable.
 

·
Registered
Joined
·
8,204 Posts

Quote:
Originally Posted by dargo  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides#post_24543379


A great episode Scott, but when Allen said he has been unable to copy a CD and have the copy sound as good as the original requires much discussion. it goes against everything I've heard or read.

Never mind ripping a CD to wav or flac how can that ever work if you cant even copy a CD. Would love a panel to come on the show with Allen to discuss his position.
Part of the problem is CD's are not digital, the laser reads the pits and lands and the signal is converted to digital, this is where problems can start.
 

·
Registered
Joined
·
735 Posts

Quote:
Originally Posted by tenthplanet  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides#post_24546943


Part of the problem is CD's are not digital, the laser reads the pits and lands and the signal is converted to digital, this is where problems can start.

Stop the clock, if CD's are not digital they are analog??? so the logo on CD's Compact Digital Disc is a lie??? where exactly is the music converted to digital???

If CD's are not digital, then they can not accurately reproduce the original master??? my head is spinning.
 

·
Registered
Joined
·
553 Posts

Quote:
Originally Posted by tenthplanet  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides#post_24546943


Part of the problem is CD's are not digital, the laser reads the pits and lands and the signal is converted to digital, this is where problems can start.
Only if your player/copier/ripper has no ability to make use of the error detection capabilities inherent in red book audio.

Yes, something can happen that causes problems reading the pits in the medium. But...


Each track on a standard "Red book" CD is divided into sectors of 3234 bytes of data. Of that, only 2352 bytes is actual data (audio data in our case). So what is the rest for? Well 98 bytes are just cueing overhead that enable track-to-track skipping. The rest, 784 bytes, is error correction coding. The coding scheme used can basically correct errors of up to 500 bytes in length and can detect MUCH larger errors. To make this physical, a scratch of 1/10 inch or less is completely correctable with no information loss because the encoding scheme carries with it enough redundant information that it can re-create with perfect accuracy any data loss resulting from errors of this size.


So what happens if you have a bigger scratch? Well, even then, while you can't recreate what was lost using the on-disc scheme, you can detect that it was lost. Many good rippers and copiers use this information to attempt multiple retries reading a bad sector. Sectors with errors on CDs can often read different data from one attempt to the next because of minuscule changes to how the light reflects off the disc inside the transport. If any one of those retries yields something that does not show an error detected for that sector, then that HAS to be the correct data, down to each individual bit.


The net of this is that a good copier/ripper will be able to use "forward" error correction for small errors, "backward" error correction for larger errors and will either

A) Be able to correct them all - perfectly or

B) TELL YOU when it can't. (say - too many failed retries).


As such, you can rest assured with a good copier/ripper that if you aren't told that there were errors, you have a perfect bit-for-bit copy of the original material.


Now - writing to the new shiny disc (or a file) is really the only remaining question. Well, guess what!! You write the error correction information too! Guess what happens if the data gets screwed up while writing? A re-read will result in an error!! So as long as you are intelligent about using the ECC information on the disc as designed, there is basically a 100% probability that the audio data on the destination disc is an accurate copy of what was originally placed on the source disc, even if the source disc now has small errors due to physical abuse.


The only logical explanations, therefore, for a digital copy "degrading" are the following:

1) Use of hardware or software that did not make proper use of the the error detection/correction scheme inherent in CD audio or

2) Subconscious (or conscious) expectation bias altering how the output is perceived by the listener.
 

·
Registered
Joined
·
2,064 Posts
great video and good insightful information and a lot fun as well.
 

·
Registered
Joined
·
1,011 Posts

Quote:
Originally Posted by LairdWilliams  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides#post_24547504



As such, you can rest assured with a good copier/ripper that if you aren't told that there were errors, you have a perfect bit-for-bit copy of the original material.

This is precisely the problem that Allen Sides mentions many times in the interview: it simply isn't so.

In reality, each time you record or copy or write digital data there will be errors or changes made. Because there is some degree of error checking it does not mean that the process is perfect. Indeed, it cannot be perfect, either in the digital realm or the analog realm.

As Allen stated, CD pressing plants vary in their ability to take digital data from a master and precisely or exactly reproduce that data onto each copy they make. Then there is the problem that cpu's use probability and algorithms in their operation and with data compression schemes being added to the mix there are additional sources of potential error which either fall below or outside the error correction schemes.

Computers do make mistakes!


Transferring digital data onto a physical medium such as a CD or DVD or a hard disk is not some magical perfect process. It can only be done as well as the limits of the technique used allows, error checking or no error checking. Just like there are cars and Rolls Royces, so there are high quality transfer processes and less high quality transfer processes. Each and every stage in the transfer from the original recording of the film / music event to the final output on CD or blu-ray etc will introduce some change or manipulation. It is impossible for that not to happen.

Here is a link to an excellent interview with DP Geoff Boyle where he talks about how information is being thrown away and being guessed at using algorithms at the very first stage of capture. The same process continues thru-out the chain until we see and hear the final result on our projectors, tv's, speakers or whatever.

Quote:
Originally Posted by coolscan  /t/1309492/4k-by-2k-or-quad-hd-lots-of-rumors-thoughts/3630#post_24468105


Here is a short interview with British Cinematographer Geoff Boyle where he talks about resolution,RAW, compression and the future resolution of cinema.
 

·
Registered
Joined
·
553 Posts

Quote:
Originally Posted by catonic  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides#post_24549662


This is precisely the problem that Allen Sides mentions many times in the interview: it simply isn't so.

In reality, each time you record or copy or write digital data there will be errors or changes made. Because there is some degree of error checking it does not mean that the process is perfect. Indeed, it cannot be perfect, either in the digital realm or the analog realm.

Ummmm...then explain how we manage to copy large files of all kinds of data, from banking data to military information - exclusively digitally, and with no degradation from one copy to the next? And please remember that - once digitized, there is NO difference between audio and video data and any other kind of data. Unless, of course, you are UN-digitizing it at some point, which is NOT what happens when a digital file is copied.
 

·
Registered
Joined
·
2,064 Posts

Quote:
Originally Posted by LairdWilliams  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides#post_24549937


Ummmm...then explain how we manage to copy large files of all kinds of data, from banking data to military information - exclusively digitally, and with no degradation from one copy to the next? And please remember that - once digitized, there is NO difference between audio and video data and any other kind of data. Unless, of course, you are UN-digitizing it at some point, which is NOT what happens when a digital file is copied.

its more a long the lines of when you make a copy of cd to another cd and take that new copy and make a copy of that, and repeat, over time you will have so much degradation that it wont be able to copy anymore... its only when you turn it into a physical medium that has to be created, and in the case of a CD, it goes from digital to optical reflection and that is where data can be lost or damaged. Like mentioned in this video interview.
 

·
Registered
Joined
·
1,011 Posts
^^ One reason is as follows: When we copy some digital data then we do some sort of checking mechanism (say checksum, as a crude example) that is intended to show that the copy is the same as the original, that test itself is not perfect. It is only capable of checking or testing certain parameters. If we are checking a document involving text and maths etc it may well be good enough to show if all the figures are same as in the original. If we have a high definition, high quality movie file with a complex soundtrack the ability of the checking software to see if an exact copy has been made may not be up to the task. It may report that an exact copy has been made but that will only be true within the parameters it is capable of testing.

What Allen Sides is saying, and IMO he is well placed to know what he is talking about, in practice, especially with sound and video / film, there are discernible differences that show that an exact copy has not been made. No doubt my ears would not be able to pick up the differences that he can but everything I know about computers, software and film and sound recording and copying / transfer tells me he is correct.

Many computer people, like people in general, tend to have an exaggerated belief in the capacity of their tools to do the job they claim they do. Some expert film and sound people are aware of those limitations. Unless we are also expert in those areas and well aware of how cameras, microphones etc capture in digital form what they are intended to capture and how that data is then transferred to other media then we will be most unlikely to pick up things that someone as experienced as Allen Sides can.


At an even more complex level, anyone one interested in this matter may choose to check out this Home Theatre Geeks show:
Quote:
Originally Posted by Scott Wilkinson  /t/1521459/how-low-can-you-go-with-keith-yates#post_24452347


Theater designer Keith Yates talks about how to get the best bass in all seats, including classic room-dimension ratios and how they don't really apply to real theater rooms, how the walls in modern room construction act like diaphragmatic bass absorbers (which is generally a good thing), the importance—and expense—of computational fluid dynamics in determining a room's bass performance, how four moderate subwoofers can sound better than one really great sub, listening to the sound of a room on headphones before it's built, answers to chat-room questions, and more.


Using software as complex and sophisticated as computational fluid dynamics (CFD) is going to give a different result to software that is linear based when it comes to testing how sound is reproduced in a room. The same principle applies to testing a highly complex digital file such as a high quality film and soundtrack and trying to see if subtle nuances in colour, tone, shadow and sound have been accurately reproduced.

We are still at the beginning of the digital age and there is still much room for improvement in out digital tools, imo.

People such as Scott Wilkinson, Allen Sides, Joe Kane etc point us in the direction where we can learn why and how we can make those improvements.
 

·
Registered
Joined
·
553 Posts
While there may indeed be flaws in digital encoding of "colour, tone, shadow, and sound" as a result of digital encoding, that has nothing to do with digital copying. I agree that we can probably find some improvement, but making perfect copies of digital information is not one of those places.


Your assertion that other forms of data are less sensitive to small errors is complete nonsense. Please don't go into a hospital if you believe this. And you probably should start keeping your money in a mattress. Oh - and don't use any electricity from the grid either. All of these things and many many more rely on the ability to make perfect digital copies, petabytes of them a day, every day. And we won't start on what would happen inside that computer you are using to make your posts if digital copies errors could go undetected there.


A bit is a bit. If even one gets flipped it is significant error, and most systems are well-designed enough to detect errors whose probabilities are even vanishingly small (like "1 out of the number of grains of sand on all the beaches in all the world" small probabilities). Some, like the playback mechanism on a CD player, are designed to ignore them to some degree in return for robustness, so one spec of dust on your disc won't cause the entire playback to halt - but that does not apply to copying. It applies to playback.


You also need to spend a little time with the mathematics of information theory and error correction. Your grasp of the field, which I studied, is so shallow that it is causing you to make incredibly bad assumptions about how error detection and correction work.
 

·
Registered
Joined
·
7 Posts

Quote:
Originally Posted by LairdWilliams  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides#post_24550570


While there may indeed be flaws in digital encoding of "colour, tone, shadow, and sound" as a result of digital encoding, that has nothing to do with digital copying. I agree that we can probably find some improvement, but making perfect copies of digital information is not one of those places.


Your assertion that other forms of data are less sensitive to small errors is complete nonsense. Please don't go into a hospital if you believe this. And you probably should start keeping your money in a mattress. Oh - and don't use any electricity from the grid either. All of these things and many many more rely on the ability to make perfect digital copies, petabytes of them a day, every day. And we won't start on what would happen inside that computer you are using to make your posts if digital copies errors could go undetected there.


A bit is a bit. If even one gets flipped it is significant error, and most systems are well-designed enough to detect errors whose probabilities are even vanishingly small (like "1 out of the number of grains of sand on all the beaches in all the world" small probabilities). Some, like the playback mechanism on a CD player, are designed to ignore them to some degree in return for robustness, so one spec of dust on your disc won't cause the entire playback to halt - but that does not apply to copying. It applies to playback.


You also need to spend a little time with the mathematics of information theory and error correction. Your grasp of the field, which I studied, is so shallow that it is causing you to make incredibly bad assumptions about how error detection and correction work.

I fully support this post. Digitally copying a block (or sector or track) of audio or video data is no more and no less complicated or error-prone than any other block of data. We don't just flip bits every once in a while in IT. This would indeed be a very rare ocasion and any modern checksum algorithm will detect these errors.


Computational fluid dynamics (CFD) concerns the solving of non-linear partial differential equations (Navier-Stokes) by numerical means (finite element/volume/difference methods). These numerical algorithms are error-prone per se as you try to represent fractions and real numbers with a limited number of bits. It is a totally different field in computer science.
 

·
Banned
Joined
·
3,903 Posts

Quote:
Originally Posted by Daniel Chaves  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides#post_24550038

Quote:
Originally Posted by LairdWilliams  /t/1524739/digital-audio-and-analog-speakers-with-allen-sides#post_24549937


Ummmm...then explain how we manage to copy large files of all kinds of data, from banking data to military information - exclusively digitally, and with no degradation from one copy to the next? And please remember that - once digitized, there is NO difference between audio and video data and any other kind of data. Unless, of course, you are UN-digitizing it at some point, which is NOT what happens when a digital file is copied.

its more a long the lines of when you make a copy of cd to another cd and take that new copy and make a copy of that, and repeat, over time you will have so much degradation that it wont be able to copy anymore... its only when you turn it into a physical medium that has to be created, and in the case of a CD, it goes from digital to optical reflection and that is where data can be lost or damaged. Like mentioned in this video interview.

Please, stop. You are wrong. I am a computer scientist and you are dead wrong if you believe digital copying of CDs is in any way similar to photocopying a photocopy ad infinitam. The first CD you copy is precisely as accurate as the Nth copy of that copy, by the transitive property. Either error correction works or it doesn't. And it works. Therefore you can make an infinite amount of copies of copies of CDs without a single error from the point of view of bit-by-bit comparison once you re-rip it.


I normally love watching AVS interviews but based on the falsehoods that he espouses, makes me completely ignore everything else he has to say on this topic.


Error correction on CDs either guarantees that you have a bit perfect reproduction, or that you are aware there is a significant error and therefore cannot trust CDs as a medium. I've made thousands of copies of CDs and DVDs in my life, and copies of those as well, and I cannot exclaim how utterly wrong you are without putting on capslocks. And the interviewee as well.


This thread is a total facepalm.
 

·
Registered
Joined
·
331 Posts
It just doesn't make sense to me, that you guys are willing to debate a guy like Allen Sides with so much authority. I mean this guy is as credible as it gets. He has loads of experience.

He even gave references to Mariah Carey throwing away half a million cd's? Did she really do that for NO audible reason? Doubtful.


What Allen was saying is exactly the opposite. You're doing the same thing they were referring to in the video, "because you can't prove it" blah blah, well what reason does Allen have to be biased? Wouldn't it be easier for them to all be perfectly, audibly the same? Seems if anything, you'd be biased in the other way naturally!


Now I personally know very little to nothing, but I'm just coming from a logical standpoint.


I'd trust gobs of experience over theory any day.
 

·
Registered
Joined
·
1,011 Posts
One last try.

http://en.wikipedia.org/wiki/ECC_memory


Electrical or magnetic interference inside a computer system can cause a single bit of dynamic random-access memory (DRAM) to spontaneously flip to the opposite state. It was initially thought that this was mainly due to alpha particles emitted by contaminants in chip packaging material, but research[1] has shown that the majority of one-off ("soft") errors in DRAM chips occur as a result of background radiation, chiefly neutrons from cosmic ray secondaries, which may change the contents of one or more memory cells or interfere with the circuitry used to read/write them........


.........Work published between 2007 and 2009 showed widely varying error rates with over 7 orders of magnitude difference, ranging from 10−10–10−17 error/bit·h, roughly one bit error, per hour, per gigabyte of memory to one bit error, per millennium, per gigabyte of memory.[2][4][5] A very large-scale study based on Google's very large number of servers was presented at the SIGMETRICS/Performance’09 conference.[4] The actual error rate found was several orders of magnitude higher than previous small-scale or laboratory studies, with 25,000 to 70,000 errors per billion device hours per megabit (about 2.5–7 × 10−11 error/bit·h)(i.e. about 5 single bit errors in 8 Gigabytes of RAM per hour using the top-end error rate), and more than 8% of DIMM memory modules affected by errors per year.....


And that is just for memory.


CPU's do not operate exactly, checking everything they do in exact, perfect detail. They use probability and algorithms, making guesses and assumptions.

Computers make mistakes.

Hardware, including hard drives, are not perfect.

Software, including error checking software, which goes thru revisions, upgrades and updates like most software, is not perfect.

Human beings are not perfect.


On the other hand, our tv's and projectors and sound systems and computers are gradually improving so that is great.
 
1 - 20 of 36 Posts
Top