This thread has been split from here, according to Rule #1, and Note #1 of the diyAudio rules, and in respecting the wishes of the original poster.
*****Please avoid any disagreements you all may have, it will only cause me more confusion as I am just trying to learn!
Okay...
Melanie is crystal clear, except for a half a dozen time she overloads the mic for one or two words. When she goes to french "ils ont changé ma chanson" the word change is pretty rough as was the word "brain" in the second verse. (A common problem back then)
Mason Williams is pretty good (my offline copy is better) I can hear his guitar all the way through, even among the other instruments. In his solos you can even hear him fingering the strings. The brass section is amazingly dynamic.
Stepenwolf (magic carpet ride) is totally clear. I hear all the instruments all the time, only a couple of words are masked when he trails off. The good old Hamond B3 is at it's best and you can tell the drummer is using riveted symbols. It is impressive they did that without any special effects at all... not even a plain old fuzzbox.
I don't know what anyone else might say...
especially if they have offline copies to use.
All that aside, speaker positioning is not done with music. You need to be sweeping in REW and looking at your response graphs... the flattest response wins.
Last edited by a moderator:
In my experience, the best REW sweeps almost always correlate with the best sound after. It's pretty rare that I see a bad sweep with good music or vice versa.
Could you tell us what frequencies we are talking about and why those frequencies?
You do understand the concept of a flat frequency response... Don't you?
Or are you being deliberately obtuse with me?
Okay... so you are being obtuse. (be sure to tell me if I pass the test...)
My wording suggested an observation.
The audio spectrum generally runs from 20hz to 20khz. I didn't decide this, a whole series of experts did and it's been the accepted standard since the mid 1950s.
The goal in hifi systems is to be able to accurately reproduce an electrical waveform within that frequency range. The term "accurately" is not my word, it is the underpinning of the term "high fidelity".
Speakers belong to a class of devices known as "Transducers". These are devices that covert energy from one form to another. For example the heating element on your electric stove converts electricity to heat. In the case of an audio system the speakers are converting electricity into air movement which the human ear interprets as sound.
Recorded on each of our sources (vinyl, tuner, tape, cd, file, etc.) is a model for an electrical waveform that (hopefully) represents the sound we are about to hear. This waveform is passed, component to component, as a fluctuating voltage. In order to maintain the High Fidelity standard, this waveform needs to be reproduced accurately at each step of the way, finally reaching our power amplifiers.
Now we bump into a little artefact of Ohm's Law. Current is the result of applying voltage across a conductor. Thus if we change the conductor, we change the current, but not the voltage. If we switch to a conductor with a higher impedance current is reduced. If we change to a lower impedance current increases. Except for catastrophic failure the voltage remains constant.
Thus, an audio power amplifier is producing a fluctuating voltage across a speaker. This voltage literally tells the speaker's pistonic cone where to move to. It is the transducer's job to do this accurately if the High Fidelity standard is to be obeyed. That is to say that the speaker's cone is expected to move in very predictable ways in respect of the waveform it is being fed by the power amplifier.
As should be obvious, airflow in a room is anything but uniform. Furniture, soft and hard surfaces, doors, windows, corners, etc. all come into play. Sound is essentially the movement of air and it is also affected by the room's characteristics.
If we are to maintain the goal of High Fidelity we need to create rooms and position speakers in such a way that the pistonic motion of the transducers can faithfully reproduce the sound represented as a waveform on our source recordings.
Within the agreed bandwidth of 20hz to 20khz each frequency present in our source recordings needs to be reproduced accurately. We should hear the same level of sound at all frequencies... which is referred to as a "flat frequency response".
Failing to do this, alters the recorded sound and is thus not accurate reproduction.
Now in practice we will never get this perfect. There are just too many variables in the process. But, if we are to hear a reasonably accurate reproduction of what is on our source recordings we need to get as close to that ideal as we can.
Thus the goal of speaker positioning within a room is to find the spots at which we get the most accurate reproduction...
This is most easily accommodated by using test equipment (or software) to send out a sweep of all frequencies in the audio range and chart the result. Comparing the charts helps us find the best position for our speakers... Per the High Fidelity goal the spot with the most accurate frequency response will also give us the most accurate reproduction of our source recordings.
Does that answer your question?
@ Douglas. Everyone is entitled to an opinion, but when facts are correct they stand up to examination. Usually only simple responses are required and indicate an understanding of the subject.
This is a technical forum, expect technical matters to be questioned..
This isn't the thread for this discussion. If it continues here I'll need to split it to another thread to keep things clean for Kustomize.Douglas Blake said:
This is a technical forum, expect technical matters to be questioned..
I believe trying to achieve a flat FR at the listening position is best pursuit as @Douglas Blake has done an outstanding job illustrating. You can always make slight adjustments based on preferences. The OP will learn this with more experience with measurement gear. It has taken me some time to determine that my preference is a couple dB lift in the bass region. This leads to a somewhat descending FR. Just my 2 cents from someone still learning. No hard and fast rules.
Best Regards,
Rich
Best Regards,
Rich
This is right.
Sometimes I suggest an in room measurement as a starting vector. However I'm careful not to suggest it is a definitive process because as I understand it, it misleads. There is a time and place. Results can be ok.
I used to do only in room measurements but was not able to achieve the same level of success as I do now.
I'm confused -- has someone suggested that adjusting a system by ear is superior to doing so by measurements? It seems like an argument without an opponent. I mean, I suspect some keen listeners might be able to do so, but I wouldn't make a generalization either way, personally.
My overarching question would be "why frequency response?" and as a followup, "how do you determine which of two sweeps in the frequency domain is better?"
I mean, if one sweep is +/- 3 dB from 20 Hz to 20 kHz but it has a sudden trough at 200 Hz, and another is +/- 4 dB but is quite smooth throughout, which is "better"?
And why is "frequency response" more important than "phase response", etc.?
My overarching question would be "why frequency response?" and as a followup, "how do you determine which of two sweeps in the frequency domain is better?"
I mean, if one sweep is +/- 3 dB from 20 Hz to 20 kHz but it has a sudden trough at 200 Hz, and another is +/- 4 dB but is quite smooth throughout, which is "better"?
And why is "frequency response" more important than "phase response", etc.?
Of course it's the only way to go.
AND... it can only be done by test equipment. With variations in hearing, confirmation bias, wide variations between rooms, furnishings, doors, windows, equipment, sources etc. only a complete fool would try to tune up a high quality system by ear.
Last edited:
Phase is an issue when designing speaker enclosures and crossovers, there is no doubt about that.
But, in a room the speaker phase is far less important than room nodes, reflections, resonances and echos... all of which happen in the frequency domain.
Allen challenged me twice to explain why I would ask for test measurements.
I explained, as you saw at the beginning of this thread.
To get the rest of the story, also see THIS THREAD
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.