Saturday, June 17, 2006

What Is ESSB ?

by John Anning - NU9N


ESSB
What is Lo-Fi - Mid-Fi - Hi-Fi and Extended SSB ( ESSB ) ?


Source: http://www.nu9n.com/essb.html - What Is ESSB Hi-Fi SSB Audio - ESSB Extended J3E Audio Bandwidth Defined

ESSB Hi Fi sounds ominous. These terms can be somewhat confusing since true "Hi-fi" is an old stereophonic audio term referring to true High-fidelity audio - e.g. 20Hz ~ 20kHz of flat frequency response with very low Total Harmonic distortion (THD) and good Signal-To-Noise (S/N) characteristics. By this definition, any Amateur Radio transmission (SSB or AM) hardly qualifies as absolute true Hi-fi.

However, speaking in a relative fashion as compared to standard SSB practice, I would like to define the terms we use as follows in the table below as a relative reference only:

Note: I consider "Extended SSB" to be any J3E SSB mode that exceeds the audio bandwidth of standard or traditional 2.9kHz J3E modes (ITU 2K90J3E) in order to support the fidelity required and desired for relative High Fidelity, full range clean and articulate vocal audio.
See table below:

ESSB Table


















In the scheme above, 3 kHz of audio bandwidth was the chosen criteria or threshold qualifying a signal as ESSB. The reason for this, is that high frequency audio from 3 kHz and above starts to support a significant difference in clarity, "openness" and fidelity of the audio signal that better reproduces natural energy found in the human voice. Even though vocal chord energy diminishes rapidly above 3 kHz, the all important high frequency consonants of human speech such as the "S", "T", "SH", "CH" "K" and "Z" sounds that are formed with various combinations of the tongue, roof of the mouth and teeth are well above 3 kHz. The accurate reproduction of these sounds are essential for high definition speech with less listener fatigue. See the excellent Polycom White paper on: "The Effects of Bandwidth vs. Speech Intelligibility".

The "E" in "ESSB" Stands For "Extended" not "Enhanced" or "Expanded"!

The term "ESSB" is being misinterpreted by many. I normally wouldn't say anything about it, but I have a couple of concerns regarding the "Enhanced" viewpoint, vs. the intended "Extended" version I designed as stated below from my website definition originally authored in January 2003:

While "Extended Single SideBand" is certainly an "Enhanced" form of SSB, "Enhanced SSB" is not necessarily "Extended" in bandwidth and fidelity! So, perhaps "Enhanced SSB" is a valid term for EQing the crap out of a 2.4 kHz signal, but it is NOT the "ESSB" that I established and defined. Maybe we should look into coining yet another phrase like Art Bell mentioned as: "BSSB" or "Better Single SideBand".

The main reasons I am concerned about how ESSB is defined are:

  • I originally authored the term "Extended Single SideBand" for a reason and it is copyrighted as such!
  • The term "Extended" is a quantifiable and definable term that is conducive to the bandwidth necessary for the fidelity it supports (as shown in the above tables)
  • The term "Enhanced" is a descriptive adjective that could mean just about anything, like "Smooth", "Pleasing", "DX Audio", etc...

Those who are told that any bandwidth (even 2.4kHz) can qualify as "ESSB" if EQ'ed properly, are being misled and likely to think that if they hear a signal that occupies 4, 5 or 6kHz of bandwidth, that something is wrong! I don't want to make excuses for ESSB, or the bandwidth it requires, but rather just define it and enjoy the fidelity that the "Extended SSB" bandwidth supports. The law of audio physics demand that audio fidelity requires audio bandwidth. --- We really can't have one without the other!

We need to be on the same page when referring to terms like this. 2.7 kHz of bandwidth will always sound like 2.7 kHz of bandwidth, no matter how aggressive we process it. Can it sound better than non-processed 2.7 kHz SSB? You bet! But it will NOT be "ESSB" that supports critical frequencies in human speech found above 3, 4, 5 or 6 kHz. Anyone can boost bass frequencies in a 2.7 kHz bandwidth... But that really doesn't produce the definition, fidelity and "openness" of a true "ESSB" signal as defined. That would only produce a modestly "Better Sounding Single SideBand" (BSSB)... HI.

To conclude, I just would like to keep the record straight regarding ESSB... I'm just concerned that what we all have been working hard for, doesn't end up in the toilet, where the ARRL can conclude that a 3k SSB bandwidth limitation (or less) would still support ESSB operations. IT WILL NOT !!!

Questions about ESSB

Should You Even Consider ESSB or Not?
If your desire is to have a more pleasing and natural sound than that of conventional SSB practice and a persona that is rich, smooth and highly defined, then you should consider ESSB. If you desire audio that does not sound like it originated from a tuna can, the bottom of a barrel or from the song "Winchester Cathedral, then ESSB may be for you.

However, if it is your desire is to "Cut" through QRM and bad band conditions, break a DX pileup or have contest grade audio, then forget about ESSB...it is definitely not for you! Stay with your Yaesu FT-1000-D and Heil HC-4... You will be much happier!

Are There Any Real Benefits To ESSB Hi-fi Transmissions and Reception?
ESSB
Hi-fi has many benefits, the most noticeable being a very pleasing and highly defined sound not inherent with traditional SSB audio. Also, it only occupies 1/2 of the bandwidth required by AM for the same audio quality, since SSB is a Single Sideband Suppressed Carrier mode vs. AM, which is a Double Sideband with Carrier mode and is therefor a more efficient signal and less subjected to the phasing and selected fading problems inherent with its AM counterpart.

Additionally, if you will be pursuing ESSB, you will also inherently be pursuing a clean, low distortion signal with low I.M.D characteristics, simply because you will not want these artifacts on your hard earned audio since it sounds bad, and this is, after all, why you are pursuing ESSB audio in the first place... To sound good!. This dimension alone will make you a better operator through experimentation that leads to engineering practices and technical excellence throughout your station such as RFI suppression techniques, cabling, grounding and in-depth equipment adjustments as well as many other factors that lead to a high quality signal. This will definitely expand your knowledge and harness skills otherwise not realized with simple plug-and-play SSB operations practiced by most.

Why Use More Bandwidth Than 2.4 or 2.8 kHz?
The simple reality here is this: Audio fidelity is directly proportional to audio bandwidth. And, with a clean SSB signal with low distortion and I.M.D., SSB RF bandwidth will be directly proportional to audio bandwidth.

Because of the way that our ears work, and the different audio frequencies that are supported by wider bandwidths beyond 2.4 kHz, the ability to reproduce certain sounds (especially high frequencies) becomes more realized as the bandwidth of a signal increases. For example, with a 2.4 kHz wide signal with a carrier set point adjusted for 200 Hz (200 Hz being the lower end limit of the total bandwidth), the highest frequency that will be heard will be at 2.6 kHz. While 200 Hz ~ 2.6 kHz will arguably suffice in reproducing the main envelope of human speech for intelligence and for bare minimum communications requirements, it hardly qualifies as a way to communicate accurately with any of the inherent frequency components found in the original signal that entered the transceiver, let alone the original voice itself.

As the bandwidth of a signal increases, improvements can be heard, making the audio sound more like the original source audio. When the bandwidth is increased from 2.4 kHz to 3 kHz, 600 Hz of the upper voice energy is recovered. Likewise, when the bandwidth is increased from 3 kHz to say 4 kHz, 1000 Hz of the voice's energy is recovered.

You may ask why 4 kHz of bandwidth or beyond is needed when arguably the voice can be sufficiently reproduced at 2.4 kHz as mentioned above? Hasn't 2.4 kHz been good enough for years? ... Well my answer would be a question; Is 2.4 kHz sufficient to produce the human voice accurately? My answer to that would be an emphatic NO! And just because is has been "good enough", doesn't mean that we shouldn't pursue some improvements. Isn't that, after all, in the true nature and spirit of Amateur Radio?

There are certain consonants that are very frequency dependent. For example, "S" and "F" will sound almost identical in a 2.4 kHz bandwidth limitation, as well as letters like "M" and "N" and also "C" and "E". When we look closely at the speech characteristics of a letter like "C" or "S" for example, there are natural components generated that are much higher in frequency than what the vocal cords are producing because of the "SSSing" sounds made by the tongue and roof of the mouth, producing high frequency air making up part of the speech process. If these frequencies contained in letters like "S" and "C" are suppressed, what we are left with are words that sounds like "shame" instead of "same" and "hey ho" instead of "say so", "hell" instead of "tell" or "J3AEM" instead of "KT8TN". Get the picture? For more on this, an excellent audio bandwidth "White Paper" article on this by Polycom is available. Also see "Broadcast Chain Tutorial" by Martin Wolters.

There is even more however... There are those who claim that there is no appreciable energy in the human voice beyond 3 kHz. I guess it would depend on how we define "appreciable". However, there is definitely energy above 3 kHz, and it is more than just harmonic energy as opposed to what some would argue. Look at the spectral graph below where I measured my own voice in a flat 20 Hz ~ 20 kHz environment using a good flat condenser microphone. You will see the energy content in my voice well above 3 kHz... Also notice that the human voice is not flat, but rather decreasing in amplitude as frequency increases.

NU9N Natural Voice











As you can see, I have some natural and dominant high frequency excitation occurring at 4kHz and 7.5 kHz due to the "SSSing" sounds as described above. If we remove these frequencies, especially the 5 kHz components of my voice, the 3 kHz component must carry the information by itself. Is a 5 kHz bandwidth required for the communication to be understood? Of course not. If I would have included the word "accurately" in my question, then the answer would be yes. I am only pointing out how we take for granted the natural energy that is really inherent with normal and natural speech! In my case (as with most male voices), a bandwidth of 4 ~ 5 kHz would catch the lower end of "SSSing" intelligibility before using the full bandwidth necessary for accurate reproduction that occurs at about 7.5 kHz. This is where the AM mode has it hands down when accurately reproducing full speech characteristics. The problem with AM however, is that to reproduce 7.5 kHz of audio with accuracy, the RF bandwidth would be required to be 15 kHz wide! (7.5 kHz x 2 sidebands = 15 kHz). For this site, we will stick to ESSB and RF bandwidths at or less than 6kHz.

Is Any RF SSB Bandwidth Beyond 3 kHz Legal?
The simple answer to this is YES! There has been much debate over this one. Debate it all you want. The simple fact is, THERE ARE CURRENTLY NO BANDWIDTH RULES stating hard numbers for the HF Amateur Radio service concerning A3A, J3E etc...! Look at Part 97 a thousand times if you like. YOU WILL NOT FIND A BANDWIDTH FORMULA OR RULE in place.

All you will find, are some very subjective and gray guidelines regarding "minimum bandwidth necessary". We could debate the meaning of words like "necessary", "quality desired", etc... all day long and not reach an agreement. This is where the vagueness of Part 97 gets a bit sticky. I have seen Amateurs debate this until Hell itself dropped 1000 degrees in temperature! HI. Let's stop this debate for awhile and just use some common sense...Read on...

The real questions is this... Is it always a good idea or acceptable to use more than 4kHz of bandwidth? Again the answer to this is simple... NO! If a band is very crowded, and there is nowhere to squeeze in a signal that is 6 or 5 kHz in bandwidth without interfering with someone else, then common sense and common courtesy would obviously apply here. Narrow it down to 4 kHz or less and I'm sure that most operators will be okay with this. We do have to compromise at times and Amateur Radio is no different than any other area of life where people are involved. If you can produce a good clean 4 kHz of RF and audio bandwidth, you will already be narrower than most SSB stations using 2.4 kHz of audio who's splatter and I.M.D. occupy twice that or more!

Is ESSB Really Necessary?
ESSB Hi Fi Audio is not absolutely necessary unless you consider accurate speech absolutely necessary. But other than emergency communications, is anything we do in Amateur Radio really necessary? No! Is DXing necessary? No! Are contests necessary? No! Is SSTV, FSTV, FSK, RTTY, Packet, AM, FM, SSB, Moonbounce, Satellite communications, Rag Chews, Nets or CW really necessary? No No No !!! So, who decides what has priority? I may be getting a bit philosophical here, but it seems to me that other than an emergency communication, NO ONE SHOULD HAVE SPECIAL PRIORITY OVER SOMEONE ELSE. And if they do, by what authority do they get their special privileges or priority?

John Anning - NU9N This is a shared service folks. I'm willing to share, I just hope that others not involved with ESSB are as well, despite their non acceptance of it. It's just another aspect of this great hobby of ours that should be treated as valid and with the same respect and credibility as any other interest in Amateur Radio!

73,

-John, NU9N