|
|
| Elliott Sound Products | Project 199 |
The ABC's (Australian Broadcasting Commission) broadcast of the New Year's Eve concert has, for the last three years, sounded like the entire band was behind a thick blanket. I don't know exactly how they managed it, but the high frequencies are well down on where they should be. I suspect (but have no evidence) that the live sound feed is equalised so the line array system used for the live sound don't tear everyone's ears off (a common problem). This year (well, last year actually 
) I decided I had to do something about it, because it sounded bloody awful!
The equaliser described is adjustable, so you can set the amount of treble boost to get a sound that you are happy with. Not only that, bit you can also adjust the frequency where the boost starts. Much as I'd like to include a sound clip of the 'before' and 'after' results, that would be a violation of copyright, so I'll show the spectrum and the EQ required using Audacity. This test confirmed that the high frequencies exist, but are pushed down by almost 10dB.
Somewhat unfortunately, it's difficult to 'reverse engineer' the exact equalisation that was applied before transmission, but an adaptation of the high frequency section of the bass guitar tone control circuit (Project 152 does allow the 'discerning' listener the ability to get the treble back again. It's not perfect, as that would require a far more advanced circuit, but the difference this year (compared to the painful, muffled sound from the previous two years) was quite remarkable.
The ABC's broadcast had hosts Zan Rowe and Charlie Pickering, and their voices sounded just fine without EQ. As soon as they start the band, everything sounds like it's been processed through a pair of old socks. It seems that most people don't notice (or if they do, they remain silent about it), but muffled and unintelligible sound drives me absolutely nuts! Doubly so when I know that there's no good reason for it, other than apparent indifference from the ABC (and I do expect better!).
Although this project is specifically aimed at Australian viewers of the ABC's NYE broadcast, it's possible that others may have found a need for something similar. As a 'single purpose' project, this is a very uncommon offering from ESP, but if it manages to improve the experience of anyone, anywhere, then it's served its purpose. Mine was built just before the broadcast when I remembered that I hated it last year, because the sound was so bad. The listening experience was greatly enhanced by the circuit, and I consider it to be a very worthwhile improvement.
The circuit is shown below. As mentioned above, this is simply a minor adaptation of the continuously variable treble control featured in the bass guitar amp project. Of course, it needs two channels because the sound from most TVs is stereo, although I suspect that the live feed is only mono. First, look at the spectrum of the sound, as received from the ABC's TV broadcast.
Figure 1 - Audio Spectrum Without EQ
The spectrum shown is not 'normal' for any typical audio source. Above 1.5kHz, the entire spectrum is suppressed, and by an amount that I would fully expect to be applied to the main signal send to the line array PA system (believe it or not, but the level is pushed down by around 8dB above 2kHz!). That's probably alright for the audience at the concert, and at least the line array (hopefully) won't sound overly harsh. The problem appears to be that the exact same send is used for broadcast. It should be a separate send without the EQ, but no-one seems to have noticed that the broadcast sound is shite!
Using Audacity, I equalised the signal, and ended up with the spectrum shown below. This was done by ear, since I have exactly zero info on the EQ that was applied, or how it was done. However, the end result is that when the spectrum looks more like what I expect from an audio source, it sounds a great deal better. At least one can hear the vocals properly, hi-hat, snare and cymbals all sound like they are there, and the overall balance is far better.
Figure 2 - Audio Spectrum With EQ
The above shows what the spectrum should look like. Over the years I've analysed a great deal of music for a variety of reasons, and when the signal is subjected to EQ it is more in keeping with what I've come to expect. The response only extends to 15kHz (as does FM radio), but that can't be changed because there's nothing there to equalise. The equaliser I built up can't compensate perfectly for the 'band in a sock' sound we normally hear, but it did make the entire concert far more enjoyable, both for me and SWMBO (she who must be obeyed ). The pots are both dual-gang types, and the second gang is used for the Right channel.
Figure 3 - Modified Equaliser Circuit (Left Channel)
One channel of the EQ circuit is shown above. The second channel uses the other half of each opamp. The variable capacitance multiplier circuit is optional, but highly recommended. The alternative is to have a 10nF capacitor directly from the end of R4 to ground, but then you have no control over the frequency where you start boosting the top end. I used the exact circuit shown, and it was tweaked a couple of times until I found that I couldn't get it any better. The problem is that you have to wait until the next NYE concert before you can verify its performance, since there usually isn't anything similar broadcast through the year. The output capacitor (shown as C2 (a and b)) just happened to be used because I had my bag of 33µF caps to hand at the time. Use a 10µF bipolar or whatever you have available that ensures that bass response isn't compromised.
The circuit is only ever needed to boost the high frequencies, and if Ropt is included, no treble cut is available. The response is shown below with Ropt included. I didn't have this option initially, but it's been added. Unfortunately, I have to wait for a year to hear it doing its job, but with the response shown I know that the normal pot setting will be at around 80-90% of maximum, which gives about 8dB of treble boost. The 'Frequency' pot is set for minimum resistance (maximum frequency) in the graph. Although the 'Frequency' section can be omitted and replaced with one or more switched caps cap, I don't recommend it because that limits the equaliser to a single frequency, which may not be appropriate. However, as noted below, the variable capacitance multiplier can be temperamental every so often (I've not found the reason yet, as it refuses to misbehave on the test bench).
Figure 4 - Treble Boost Vs. Pot Rotation (25% Increments)
The EQ is switched in and out of circuit using a miniature DPDT (double pole, double throw) relay, and the second channel uses the second set of contacts. The entire circuit is powered (in my case) from a 12V switchmode 'plug-pack' supply I had to hand, and the switch to control the relay is on a long lead so it can be switched in and out as needed. All commentary is broadcast without any top-cut EQ, and sounds horrible if the equaliser is in circuit the whole time. Because this circuit is only needed for about three hours each year, it just needs to be a quick project. No-one wants to spend a couple of days building something that's needed so rarely.
Figure 5 - Power And Relay Wiring
There's not a lot to the remainder of the circuit. An 'artificial' earth/ ground is provided by R1 and R2, bypassed by C1 and C2. These caps can be anything you have to hand, but I do suggest at least 33µF. The caps should have a voltage rating of at least 16V. All opamps should be bypassed as shown. The relay is a miniature DPDT type, and the one I used has a coil resistance of about 1k. You can use whatever you have handy, as it only has to handle signal levels. The remote is nothing more than a switch at the end of a wire, which needs to be long enough to reach your favourite chair. Make sure that the switch terminals are shrouded with heatshrink tubing or similar so that nothing can short out when it's coiled up and hidden away until next year.
There is one minor warning regarding the variable capacitance multiplier. You may find on occasion that it 'misbehaves', because it can't settle to steady-state conditions. These circuits can be temperamental, although it normally works faultlessly. If you find it to be a problem for you, you can replace the variable capacitance multiplier (U3A and associated parts) with one or two switched caps. I'd suggest around 22nF as a start, optionally able to switch in either 15nF and/ or 33nF. 22nF gives 3dB boost at 1.73kHz, 15nF gives 3dB boost at 2.55kHz, and 33nF brings the 3dB point down to 1.16kHz. these are all with the boost control set for 90% (which sounds radical, but you do need it!).
There are none, other than the bass guitar preamp project where I first showed the variable capacitance multiplier circuit. See Project 152.
Main Index
Projects Index