Current Trends in the Recording Format Arena P1

Part 1: Is Louder Better?

As many of us have noticed over the past ten years or so, CD recording quality appears to be going on a downhill trend. Remarks are commonly made by consumers and professionals that many earlier recordings on CD sound better than much of the newer material. Are these the limitations of the present format? Of course there are exceptions to the above and some of the newer recordings are stunningly good, but one might guess that if the recording technology is supposedly getting better, why aren't the bulk of newer recordings getting better?

One of the major problems that have cropped up in many of the pop music releases in the past 5 to 6 years is level control. I have painfully listened to my step kids CDs and heard the shrilling sounds they conveyed. I wasn't surprised when I hooked up the CD player's analog outputs to my oscilloscope and observed levels so high that they were clipping on a consistent basis. One normally thinks of clipping as a power amplifier problem, yet clipping in the digital domain can be just as bad. But before we go further let us define a basic explanation of what digital clipping is and how it relates to audio applications.

We know that if we overdrive an amplifier the amplifier will limit the signal and we typically see a flattened peak at the output. This is called clipping, and thus results in gross distortion. When we try to drive a CD player past the 0db full-scale point, two things can happen. The CD player will go into clipping just like an amplifier or some players (due to their digital filter/converter structure) can have a little more headroom above 0db full scale. If we drive them a little harder they will go into clipping also. CD players differ in one major respect from amplifiers. At 0 db full scale the player is at its lowest distortion; anything above that whether it is clipped or not results in gross distortion. The author knows of no CD player that can exceed 0 db full scale without gross distortion. If that's the case why are we seeing this happen on a great majority of recently released CD's?

It's a deliberate act on the part of the producers, and sometimes the musicians, who are trying to put forth the meanest and baddest recording they can make. Consumers who demand loud in-your-face recordings also drive some of this. The louder and more distorted they can make it, the more they like it. I suppose this has an appeal to a certain genre or age group, but certainly not all of us audiophiles enjoy this.

Now that we are beginning to understand that what we are hearing is a deliberate violation of generally accepted practices from the past 100 years or so, we can move on to compression. This is a technique commonly used in not only recording but in Radio Broadcasting as well. In simple terms compression is defined as bringing up the lower level signals while generally, but not always, maintaining the higher amplitude signals at their present levels. This is not a new technique by any means and if properly applied it can help a recording sound better. When this technique is deliberately abused we end up with what is commonly known as hyper compression. In the good old days, a typical pop recording had an approximate 20-db peak to average ratio, with hyper compression we now have recordings that have only a 6-db peak to average. Keep in mind here that really good recordings of acoustic sources can have a 60-db dynamic range, and a well-designed outboard DAC's can have a 96 db dynamic range for 16 bits, while uncompressed pop recordings can exhibit a dynamic range of approximately 30 db, not including fades. When this hyper compression is combined with digital clipping you get the gross distortion from the constant digital clipping and the loudness, from the hyper compression with it.

Both of these techniques, or more properly described as abuses are typically combined on many new releases. No wonder why many audiophiles and music lovers are so repulsed by it. For those of us who have some well-recorded CD's in our collections, scratch your head no longer. What's really going on here is a gross abuse of a good medium. The author wonders that if the present trends continue, will we have a generation of children who are totally unfamiliar with reasonably undistorted music?

The following are quotes from the AES Journal, September 2000. The article the quotes come from is titled "Integrated Approach to Metering , Monitoring, and Level Practices."

"Part 1 Two-Channel Metering". The author is mastering and recording engineer Bob Katz.

"The average level of popular music CD's continues to rise. Popular CD's with this problem are becoming increasingly prevalent, coexisting with discs that have a beautiful dynamic range and impact, but whose loudness (and distortion level) are far lower. What good is a 24 bit 96-kHz digital audio system if the programs we create only have a 1 bit dynamic range?"

"Since the average program level is now closer to the maximum permissible peak level, more compression or limiting must be used to keep the system from overloading. Increased compression or limiting is potentially damaging to the program material, resulting in a distorted, crowded, unnatural sound ."

"By 1997 some music clients were complaining that their reference CD's were not "hot enough" a tragic testimony on the loudness race, which is slowly destroying the industry." Each client wants his or her CD to be as loud as or louder than the previous "winner", but every winner is really a loser. Fueling that race are powerful digital compressors and limiters, which enable mastering engineers to produce CD's whose average level is almost the same as the peak level. There is no precedent for that in over 100 years of recording. We end up mastering to the lowest common denominator ."

"The psycho acoustic problem is that when two identical programs are presented at slightly differing loudness, the louder of the two often appears "better" in short term listening. This explains why CD loudness levels have been creeping up until the sound quality is so bad that everyone can perceive it. Remember that the loudness "race" has always been an artificial one, since consumers adjust their volume control to each record anyway."

"Consumers find that their volume controls at the bottom of their travel, where a small movement produces awkward level changes."

This oscilloscope photo was taken from a recent Jennifer Lopez Pop CD release . The author has found this to be typical of many new pop music releases:

Current Trends in the Recording Format Arena P1 - page 2

Part 1: Is Louder Better?

The following are excerpts from a thread that includes some of the best Recording Engineers and Manufacturers of studio gear today.

Q: What is one of the biggest problems with CD production/mixing methods today? Please note the distortion figures below and try to remember that at 0 db full scale a typical CD player can typically have -90db or better distortion

A: General Misconception on maximum recording levels for 16bit systems

"Actually, it's the only way to do it, insofar as a distributed medium is concerned. I didn't say you had to clip or limit! You could have a single sample at -0.3dBFS and only the oldest 16-bit DACs would have any problem at all, if at all. Any DAC with 18bit processing, or better, has no problem".

"That's precisely the misunderstanding I would like to correct. Most recent pop CD releases contain frequent signals that are severely distorted in nearly *all* consumer devices. DACs are only designed to accommodate signals up to the level of a sine wave at 0 dBFS asynchronous to the sample rate. In the real world, Anno 2002, level above that can be present more than 30 times per second.

Average distortion figures over a variety of players are:
-32 dB (2.5%) for a +0.69 dB intersample peak,
-25 dB (5.6%) for a +1.25 dB intersample peak,
-18 dB (12.6%) for a +3.00 dB intersample peak,

To make things worse, there is an aftereffect (latch-up, recursive filters etc.) which prolongs the distortion period, so you end up with an ear-fatiguing sound because the equipment is in more or less a permanent state of distortion. A net result that may not be better than a highly data reduced MP3 delivery without (that kind of) artifacts.

Because most new releases contain peaks above +3 dB, you would have to lower the limiter threshold to -3 dBFS for the method you propose to make sense.

Point is: Single sample peak detection is futile, and I expect us to shortly publish more papers and audio examples of this to get it across."

Thomas Lund
T.C. Electronics

Q: How can a Producer guarantee the finished CD sounds as good as the Master Tape? The gentleman below is not the only one to complain of this problem.

A: "I don't think I've ever had a CD come back from a plant that sounded as good as the PMCD I sent them (really unscientific claim, sorry...but in an a/b comparison using the same d/a converter the difference is obvious to the CLIENT). I actually do believe that 16bit/44.1kHz can sound really, really good, but rather than fine tune the medium we seem to be jumping onto the next big thing without getting the best out of what we have. Bob K. has suggested that a consumer CD player with a built in up sampler would go a long way toward improving playback of ALL existing CD's, but as far as I can tell, we're tossing the medium as flawed because there's something "better" up ahead. I can understand the appeal of surround, but that doesn't negate the stereo (or mono) production process. If more data is better all the time, than a 5.1 mix of a song will be better than a stereo mix (wow, six times more information!), and no one in their right mind (who reads marketing hype) will be buying CD's anymore...so I guess I should just sell all this stuff and go flip burgers."

Tom
Thomas Eaton recording

First find a pressing plant that will work with you. If you ask a potential plant what speed they burn their glass masters at and you get a "dah" at the other end of the phone hang up quickly and find some one who is knowledgeable and willing to provide a quality product at a reasonable (notice I did not say "cheap") price. Too many times today people are going for the bottom line and they want to get 1000 CDs pressed with artwork and jacket for $1,000.00 and then they complain about the final quality of the project. There are lots of really good duplication plants out there and someone, the likes of Glenn Meadows, could probably give you a list of ones to use and ones to avoid.

I know what your frustration is all about. We recently did a project and the test pressing came back sounding NOTHING like the master I had send the plant. When I contacted the plant I was told to talk to their "mastering" engineer but I would have to call back after 3 pm since he did not get out of high school till then. When I talked to him he was very nice and very concerned about the quality of the product the plant was producing but his boss was more concerned about getting projects out the door than worrying about individual projects or the concerns of one or two people that could hear the difference.

The "mastering" engineer told me that he did the transfers for the company and that his boss had told him on numerous occasions that "digital is digital" and it all sounds the same anyway. He further told the "mastering" engineer that the most people who could hear the difference were NOT buying their products.

The student told me that the job was better than flipping burgers at MacD's but he was trying to find another job that was more rewarding and that he could feel good about.

There are a lot of really bad pressing plants but hopefully as people trade stories and listen to the final product the bad ones will go away and the good ones will stay around but as long as people want "cheap" pressings that is EXACTLY what they will get. I am not saying this is your case but a lot of the "bottom feeders" we deal with want quantity not quality.

Best of luck in your quest for the perfect replication plant.

Tom Bethel

Note: Some studies, though informal, have found that higher burn speeds of the glass master at the CD production plant cause the "digital pits" that are inscribed on the CD to be less uniform. The lack of precision will cause higher timing jitter and increased errors. Typically timing jitter will cause more distortion of low amplitude signals, and more distortion of higher frequency signals. There are ways of insuring that what you put in to the production house is what you get. Comparing the digital files at each step of the way is one way of insuring the continuity.

Current Trends in the Recording Format Arena P1 - page 3

Part 1: Is Louder Better?

Q: What are your feelings about SACD (DSD) becoming the next defacto standard for audio encoding?

A: I saw the writing on the wall at the 1998 Tonmeistertagung. There in a large auditorium Sony was presenting it's reasons for introducing DSD as an originating format. I thought--"why should we have to dump our entire studio" just to be up to date in a few years--why should be change to DSD? To the tough German audience of classical recording people, Sony argued, in part, that we should change because the record companies want a more secure medium and besides, the sound of DSD is loads better than PCM 192. The audience guffawed. And I guffaw now!

In our tests here, when comparing DSD to PCM 24/192 (equivalent data rate) and comparing both to the two mix of the mike feed, there was no discernable difference between PCM and DSD. And we have done this many times.

For major mastering people I see that DSD is an endorsable format. Clients may come in with that format and they need to be serviced. Certainly it's important for Sony because of a captive licensing market and brand new revenue streams. But what good are DSD and SACD for the engineer and producer working today without the major label support? (It's gone in classical)

Let's take a look at what would happen if we decided to work in DSD format tomorrow:

25 years of knowledge of pcm---dump it
High end pcm workstation---dump it
800 GB of hard drive---don't dump it but buy 800 more
Expensive multiple A/D and D/A's --dump 'em
New digital studio wiring---rip it open again and add 4 times as much
20 years of engineering thinking about new digital microphones---forget it---dump that too
Expensive outboard gear--wrong format---dump it and re-quip (with what?)
Review fax from major labels that says just send us your pcm work and we'll convert it to DSD. ---Weep!

It's a total joke and we refuse to let the dullards in hi end hi-fi drive our investment. Those who push us to do this re-equipping will not guarantee our bank loan will they? This is even before one reviews the research of Lipshitz and Vanderkooy into the drawbacks of DSD.

Sony didn't tell you that you could record DSD on a regular PCM multi-track did they? No, they said we have some new gear to show and install for you.

We have an increasing load of work for DVD-A. Now someone tell me (a tough customer who has done the listening) why we should change to DSD.

Ealing mobile recording, Ltd., Chicago

Note: The above comments show us that in order for a studio to fully produce SACD discs they will need to fully re-equip their studios with a lot of new hardware and software. This is a very expensive proposition. Many recording studios are in the middle of an economic recession, and with the possibility that the SACD format will have limited distribution, many studios have serious reservations about buying into the SACD format.

 > continue on to Part 2 Current Trends in the Recording Format Arena <