In my sensation perception/neuroscience class I came across an idea that I would like input on. For all I know this idea has already been implemented in programs and high end receivers/amps.
Let me first introduce you to the equal loudness concept behind my idea. When we listen at a specific decibel level we will perceive some frequencies to be more audible/louder than other frequencies. For example: when listening to a song at 40 decibels, lower frequencies (specifically between 500 and 4,000 Hz) are not perceived to be equally loud. You would be able to hear the high frequencies such as te treble but not be able to hear base. However, when listening to a song at 90 decibels we are able to pick up most frequencies equally including the base. Therefore it is necessary to turn up the volume to 90 decibels in order to pick up all the frequencies almost equally loud. This is why many sound engineers/music producers are forced to listen at levels of 90 decibels rather than 60. This is also why some radios, including in your car, have a "loudness" feature in order to boost the lower frequencies when listening to music at lower decibel levels.
My thought- If we were able to measure the decibel level in a given space would it be possible to create a program/device that actively boosts the lower frequencies amplitude to respond to the equal loudness curve. For example: When playing a song at 40 decibels the lower frequencies are boosted so that all frequencies are perceived to be equally loud. This would mean sound engineers/music producers will no longer be forced to listen at 90 decibels and could work longer amounts of time without putting as much fatigue on their ears. This also means music lovers would no longer have to boost their music to high levels to get the sensation of the full spectrum of frequencies.
This devise would most likely require an active decibel meter in order to calculate the amount of correction implemented on the equalizer. One program for all systems would not work because of the many variables such as different speakers/space/power/acoustics will produce different decibel levels even if playing from the same receiver.
In basic terms- it would be an active equalizer fluctuating to the decibel level in the room in order to give a flat perception of loudness in all frequencies. Has anything like this been done? Would this concept work?
For more info on equal loudness curves: http://en.wikipedia.org/wiki/Equal-loudness_contours
So they do have something along the lines of correcting this problem however from what I understand it would not work the same way as my concept nor would it do a good job of equalizing the perception of loudness at different decibel levels.
First, the weighting filters do not take in consideration the environment and the actual level of sound pressure. A fixed filter will not be able to predict how loud the equipment is actually producing at a given volume level. This is very important in order to match up the frequencies to an equal perceived level of loudness. At different levels of sound pressure, different amounts of amplitude at different frequencies need to be added.
Second, It would not fluctuate to the different decibel levels. The closer you approach 90 Hz it would not adjust to flatten out the amount of amplitude added. It seems to me that a weighted filter is fixed and not active in accounting for the decibel level that is so vital in the real world. The filter would have to have an equation that predicts the amount of boost in amplitude necessary to keep all the frequencies at the same perceived loudness at different decibel levels. At 40 decibels a 20 decibel increase in the 1,000 Hz range may be necessary but at 80 decibels only a 5 decibel increase may be needed to keep the frequency's perceived equally loud. As far as I understand, (I may be wrong) a weighted system would not do this but only be able to apply a fixed set of boost and designed for specific piece of equipment.
Even if you could somehow come up with a generic weighted program to respond to the volume of the HU/receiver, it would not respond to the actual decibel level in the environment therefor omit a vital variable.
Ben, what you're describing is "sound curve weighting" and yes car and home audio receivers/head units use a stereo mic where the listener sits, to read the environmental acoustics, then the computer adjusts the 1/3 octave EQs (one per channel usually) to compensate for the acoustic response, to give a C-weighted "flat" curve. For home audio, the system is called Audissey. For cars, there isn't one standardized company for it.
Adjustments are made for the EQ, time alignment, etc. This tech has been around in various forms since the 1980s
I would think mastering engineers take care of the curve weighting in the studio considering they are listening to it as it is mixed(for normal listening anyway). I wouldn't think there would be much of a need for the device* you described considering you can do the same with an EQ and have more control over it as well.
Thank you for chiming in Glasswolf. I have been searching the internet for "sound curve weighting" but I cant find anything on it. I know there is all kinds of weighting including a-weighting but those seem to be fixed weighting and look like they are all used in the recording of measurements and tests- not home audio. I can find systems that measure the room in order to adjust for panning but nothing about adjusting for the equal loudness curves at different decibel levels.
@Phil The operator would have to have some form of knowing how many decibels are being produced by the equipment in order to set his EQ. This would be a very timely way of adjustment leaving room for lots of errors. I think a program would serve much better. There is no need for curve weighting or anything like that when the volume is played loud enough at around 90 decibels. That is where all the frequencies are relativity evenly perceived. If they were to work at a lower level, say 60 decibels, they may adjust the lower frequencies too high because they cant hear them as well/equally. That is where a program/device would come in handy.
Ben, the engineer actually wouldn't need to know the db level are being produced(although i'm sure they do anyway) since this will be changed later in the overall mastering. mastering a song/album is a long process anyway and i doubt any engineer would give up the ability to eq each track, midi, or vocal to save a little time.
..they would need to know the decibel level being produced if they were to do anything about the equal loudness curve affecting their perception at lower decibels. There is no way around it. They cant guess how many decibels are being produced and then put it into their equation to figure out how much they need to adjust each frequency. At different decibel levels (all decibel levels but most importantly anything below 90 decibels) different frequencies of the spectrum need to be adjusted to new levels.
Now that I'm looking at the most "advanced" devices from Audyssey (surly not from the 1980s), I see they have done some remarkable things though nothing to adjust for equal loudness curves. They will boost the frequencies that are struggling to be heard on the decibel meeter, but that has nothing to do with how we perceive sound. Their newest technology maintain all the frequencies at a flat level but this is according to the physical aspects, not the psychological. Even if all the frequencies are at a flat level physically, we would not hear specific frequencies as well DEPENDING on the specific decibel level.
I'm sure soon enough they will have this technology will be implemented in their products.
Thats just the thing....they use their ear not a list of db levels of certain frequencies in a track. You can use you equation all you want and most tracks will still sound unnatural and in need of even more eq
yes, though off track. It would not affect the output from a sound engineer/mixer or who ever you want to call them.
The whole point of this program/device would be to listen to songs at a lower level without sacrificing your ability to hear all frequencies equally. It really has nothing to do with the results/product from a sound engineer. That means, a sound engineer could listen to his tracks at an average of 60 decibels and still be able to work on them because he/she could then PERCEIVE all the frequencies equally. As far as I am aware, this has still not been done.
If you don't know what I'm talking about try listening to a song at a really low level. Notice how you can't hear the bass as well as the higher frequencies? When you turn it up some more, somewhere around 80-90 decibels you will be able to hear the bass just as well as you can the other frequencies. Of course this is still pretty rough and the frequencies are still not perceived perfectly equally. Its a lot more complex than just having to boost the lower frequencies.
Phil, the source material is mastered to an anechoic environment, however, every speaker has it's own timbre, and every listening environment has a unique response, or effect on the auditory perception of that source material. Using a digital signal processor to align time delay, EQ the response to the environment to give a flat C-weighted audible perception to the human ear and it's own typical perception or weighting, handle phase shift, and so forth has to be done on a per-listening environment basis.
You can't just re-master a CD and make it sound the same in every listening setting. This is what was being asked.
I didn't realize he was speaking about listening environments. That in itself is an entirely different variable. From what i read in the first post he is concerned with how the ear perceives frequencies in any environment and different sound levels. Now my point is that why would a listener give up the ability to EQ the track himself(do the sound weighting himself) rather than allow a program to do it for them. To me a program which would make all frequencies the same level to the human ear would sound unnatural. Very similar to the Loudness wars in recording but bring it to the reproduction side. You would lose softer sounds and louder sounds since they would be tamed by this program so it is all at the same level. you would end up with songs sounding like that of the Artic monkeys or Red Hot Chili Peppers which are notorious for sounding terrible due to the fact that they played into the loudness wars.
To me a program which would make all frequencies the same level to the human ear would sound unnatural.
You're misunderstanding. It's about compensation for the listening environment's effects on the source material. What the software does, is output a white noise signal (same volume signal from 20Hz-20KHz). While doing this, the microphone maps the response curve of the audio in the car's cabin, or the listening room, and finds where the peaks and dips are in that response. After doing so, the software equalizes the response, so that the white noise signal is "correct" so all frequencies, at the same power output, are the same volume. What it can also do, is compensate for C-weighting, so that the volume sounds the same across the spectrum to your ears, which it won't do without weighting compensation. It's actually done so that the audio will sound natural. Listening to the stereo in a car without any equalization is what sounds unnatural, particularly when you also aren't correcting for phase shift per speaker, and time alignment based on signal path lengths.
The time correction function is very cool. When you sit in your living room listening to your stereo system, the best place to be, the "sweet spot," will always be centered between your speakers and between 6' and 10' away. The left speaker, right speaker and your head will form a perfect triangle with your head at the peak. The reason the sound is so good in that spot is that both speakers are the same distance away from you. They are at the same volume level, and the sound arrives to your ears at the same time. Sound does travel fast, about 1,128' per second, but even a few inches difference between the left and right speakers is enough to confuse your ears. Try this at home by moving left and right from the sweet spot and observe what happens to the apparent image. Instead of the lead singer coming from mid-air between the speakers, his or her voice will move to the speaker nearest to your ears, and the soundstage will collapse to a mono source. Now sit in your car. The nearest speakers are right next to you, from 12" to 24" away usually. The far speaker is at least 30" away, usually farther. Two things will happen in this scenario: the far side speaker will sound less loud and the sound will arrive to your ears later. More than likely you will only be aware of the near speaker and the subwoofer in your current system. Utilizing Alpine's equalizer and time correction functions, both problems can be corrected. Turn the level down on the near side and delay the sound a few milliseconds so that sound from both sides arrives at the same time and at the same level. Now you've got your sweet spot and the lead singer is coming out of the windshield right in front of you. How cool is that? Of course your audio sweet spot is at the expense of the passenger side, but she probably doesn't care anyway.
This concept is even more applicable when played back through any of the surround formats, thus providing you a "front and center" seat for audio and video programs. (Be safe--don't watch videos while you're driving.)The results were impressive. First of all, it doesn't hurt to have lots power to start with. The 100-watt x 4 amplifier was plenty, plus a 260-watt sub amplifier and a 12" subwoofer in sealed enclosure. My system sounded fairly good before the automatic adjustments were made. My front speakers are less efficient than the rears, so the only adjustment necessary was a little fader action to get the image up front. It really didn't need bass or treble adjustment after all. The sound was clear and the i mage was higher than usual. After the automatic adjustments, the image was high and right in front of me, just like it was supposed to be. The manual suggests that you can change the mic location to achieve optimal results, but I did not find that necessary. You can also make all of the corrections manually; however, you need to measure the relative distances of all the channels and dial in each one individually. The EQ is also very powerful, with independent or coupled adjustments for each channel, but the menu process is somewhat tedious. It would take you several hours to accomplish the same results that the PXI-H990 does automatically in a couple of minutes, and you might not get results as good as the automatic settings.
The Road EQ function is very cool. You know how you have to fiddle with the tone controls and the volume when you get on the highway or stop at a traffic light? Not any more, this system does it all for you. The PXI-H990 Multimedia Manager is constantly monitoring the ambient noise in the car through the microphone and adjusts EQ levels to compensate. The trick is to distinguish between your music and road noise, but since the audio signal runs through the manager anyway, it compares the microphone input and adjusts accordingly. You can set the Road EQ to three levels of compensation or turn it off. I went with the medium setting and didn't have to fiddle with the controls at all.
On the Bench
The equalizer and crossover curves were typical of the digital domain, picture perfect. The graphic EQ frequency centers were consistently shifted to the right, though. In other words, the peak at 40Hz was actually measured at about 43Hz, and a similar pattern was noted on each band. It's all mathematics in the world of DSP, so somebody forgot to carry the "2" or something. I removed the color from the Graphic EQ response graph to make it easier to look at. The crossovers lined up perfectly--it's interesting that the 6 and 12dB slopes are set to -3dB pts; the 18 and 24dB slopes, at -6dB; and the 30dB slope, at -9dB. The Parametric EQ bands also lined up correctly to their assigned frequency.
The Good•Enough control over the sound and time alignment to rule the world, or get yourself in deep doo-doo.•Very smart combo. It recognizes what you're playing and goes into the appropriate mode automatically.•Automatic EQ, Time Alignment and Road EQ.
The Bad•Sure it's got powerful DSP, but accessing them through the head unit's six buttons is a pain.
•I have to ship it back to Alpine!
This is one marvelous multimedia setup. It does everything a high-end home theater should do except for multiple zones (Alpine noted that multiple zones can had when the DVi-9990 is used with the TMI-M990 AlpineF#1Status monitor). It supports every surround format currently available and every source except SACD. More importantly, it really sounds good. You can fiddle with it if you can't resist, but take it from the most anti-automatic person on earth, me--use the automatic setups!
You are probably wondering how much all of the wonderful AlpineF#1Status products retail for. Here's the breakdown: the DVI-9990 source unit is $2,400; the PXI-H990 Multimedia Manager is $2,600; and the MRV-F900 100 watt x 4 amplifier is $3,500. I recommend two of the latter to use all eight outputs from the PXI-H990. This is almost what my car is worth these days, but a good value in my opinion. The range of media and the flexibility of the system make this a very functional "status symbol." To equal this range, you would need multiple high-end head units and a trunk full of standalone processors (imagine an outboard 31-band EQ for each of the eight channels).
Kudos to Alpine for faithfully following their game plan for the last 20 years. They know who they are and what they do, and so do most audio enthusiasts. The industry would do well to study and emulate Alpine's long-term business plan instead of chasing fads and losing their identities for short-term gains.
Test Measurement
•Output Voltage:4V (1.4V when using Aux In)
•Frequency Response:3Hz to beyond44kHz, -1dB
•THD:0.014%
•S/N Ratio>102dB
•The S/N Ratio was above that of my test equipment and therefore could not be measured. Alpine specifications list this at >110dB.
Not bad for a CD player, right? But there's more to the DVI-9990 than just the audio stuff. The head unit will play several formats, including Music CD (CD and MP3/WMA including ID3 tag information), Video CD, DVD Audio and DVD Video. It won't do photo discs, so you'll have to make a DVD if you want to look at your vacation pics on the move. Speaking of video, you get to choose from S-Video or composite output and three aspect ratios: 4:3 Letter Box, 4:3 Pan-Scan or 16:9 Wide Screen.
The PXI-H990 multimedia manager has eight RCA output connectors as well as a pair of auxiliary input RCA connectors allowing you to connect another sound source to your system.
You are probably wondering how the system works. To set the automatic Road EQ and time correction functions properly, you actually start the process and exit the vehicle (for best results, do not do this while driving). From outside the car you hear strange noises as the PXI-H990 equalizes the system while comparing arrival times for each of the speakers to make time corrections. These adjustments are based on the location of the system microphone, so you should mount it as close to that of your listening position as possible. I attached the mic to the driver's side visor a few inches from my forehead for the setup. When you climb back in the vehicle, the system will be set for your position. For equalization, the system analyzes the microphone readings and then cuts or boosts various frequencies to get the correct response. This equalizing process ensures that there are no huge peaks or dips in the frequency response.