Trying to determine frequency response from speakers

MysteryMan

Active Member
I have a suspicion that my HiFi speakers are not playing sounds consistently across the frequency range, so, I am trying to investigate. However, I am wondering if my lack of some basic knowledge about sounds is causing me problems! I will explain what I am planning to do.....

I have got a copy of a program called NCH Tone Generator, which allows me to specify a frequency and then save a .wav file with a tone of that frequency playing for 10 seconds. I have laboriously gone through and created files from 10Hz to 400Hz in 10Hz steps.

Armed with a sound meter I plan to play the tones through my system and record what the decibel level is for each frequency to see if there are any obvious areas that it is lacking.

Assuming I am in the range that my speakers can reproduce would you expect them all to be at a similar level or do different frequencies play back at different levels?

The reason I ask I is that I have just played them back on my PC speakers and the level is in the 60s from 60-200 and then jumps to 80s 200-360 and then 74 at 380 and 78 at 400.

Would you expect this sort of variation as you move up the frequency range?
 

Ugg10

Distinguished Member
I suspect, for most hiif speakers, they should be fairly flat from around 200hz upward with the odd wiggle +/-5db would not be unusual. However under 200hz room modes that cancel and reinforce the sound will dominate depending on the room, position of speakers, furnishing and position of mic.

As a first stab I would download a pink noise file if you can find one and a free Spectrum Analyser for your phone (I use Audio Spectrum Analyzer dB RTA for my IPad), noting the mic on the phone is not calibrated and will top out at 15khz max. Play the file at around 75dB(A) at the mic position (quite loud to reduce background noise) and see whether you get a flattish graph.

Thus will give a better idea than discrete frequencies.

If you SPL meter can be played into your pc and read to it the you could use REQ software to create the test tones/sweep/noise and record the spectrum .
 

MysteryMan

Active Member
I suspect, for most hiif speakers, they should be fairly flat from around 200hz upward with the odd wiggle +/-5db would not be unusual. However under 200hz room modes that cancel and reinforce the sound will dominate depending on the room, position of speakers, furnishing and position of mic.

As a first stab I would download a pink noise file if you can find one and a free Spectrum Analyser for your phone (I use Audio Spectrum Analyzer dB RTA for my IPad), noting the mic on the phone is not calibrated and will top out at 15khz max. Play the file at around 75dB(A) at the mic position (quite loud to reduce background noise) and see whether you get a flattish graph.

Thus will give a better idea than discrete frequencies.

If you SPL meter can be played into your pc and read to it the you could use REQ software to create the test tones/sweep/noise and record the spectrum .
Thanks!
 

mushii

Distinguished Member
Unless you intend to use a Class 1 SLM then you may as well write off everything below 20Hz as it wont be capable of measuring them with any accuracy. Are you going to monitor each speaker in mono and near field or far field?
 

MysteryMan

Active Member
Unless you intend to use a Class 1 SLM then you may as well write off everything below 20Hz as it wont be capable of measuring them with any accuracy. Are you going to monitor each speaker in mono and near field or far field?
Yes, not expecting anything below 20Hz.

Was planning to do each speaker separately.

Can you explain what you mean by near field and far field, please?

Thanks
 

MysteryMan

Active Member
I suspect, for most hiif speakers, they should be fairly flat from around 200hz upward with the odd wiggle +/-5db would not be unusual. However under 200hz room modes that cancel and reinforce the sound will dominate depending on the room, position of speakers, furnishing and position of mic.

As a first stab I would download a pink noise file if you can find one and a free Spectrum Analyser for your phone (I use Audio Spectrum Analyzer dB RTA for my IPad), noting the mic on the phone is not calibrated and will top out at 15khz max. Play the file at around 75dB(A) at the mic position (quite loud to reduce background noise) and see whether you get a flattish graph.

Thus will give a better idea than discrete frequencies.

If you SPL meter can be played into your pc and read to it the you could use REQ software to create the test tones/sweep/noise and record the spectrum .
I am trying to find that app in the Google Store and there are a few with similar names, but, not that exact one. Can you tell me the publisher/author, please, and that may let me spot it?
 

Ugg10

Distinguished Member
It may be an IOS only app I guess -

 

MysteryMan

Active Member
It may be an IOS only app I guess -

Thanks....may have to borrow the wife's iPhone!
 

dogfonos

Well-known Member
Whilst it's an interesting excersize to measure your speakers in-room frequency response, I'm unsure of the usefulness of doing so. Having said that, many of us have a bit more time these days, so why not.

I have a suspicion that my HiFi speakers are not playing sounds consistently across the frequency range

Clearly, something you heard through your speakers prompted you to measure the frequency response. What did you hear?

Can you explain what you mean by near field and far field, please?

In terms of acoustics, near field means close up to the speaker, typically about one metre. Far field means further away, typically three metres or more - though the exact distances are open to challenge.

The further the position of the microphone is away from speakers, the more room acoustics (bass resonances/modes and reflections) affect the result. Conversley, the closer the microphone is positioned to the speakers, the less the room effects the result but, with multiple drive unit speakers (most hifi speakers are at least two-way), the exact microphone position becomes important because the two or more speaker drivers are, usually, physically distanced so which drive unit do you point the microphone at (see first link)?

In-room frequency response curves I've seen for quality speakers tend to fall away as the frequency increases (see second link).

 

Ugg10

Distinguished Member
Ooh, that looks fun :)

It’s ok for a general feeling, not analytical, and as said no data from my IPad Air above 15khz, I also use it to get a dB(A) balance between my two active speakers.
 

MysteryMan

Active Member
Whilst it's an interesting excersize to measure your speakers in-room frequency response, I'm unsure of the usefulness of doing so. Having said that, many of us have a bit more time these days, so why not.



Clearly, something you heard through your speakers prompted you to measure the frequency response. What did you hear?



In terms of acoustics, near field means close up to the speaker, typically about one metre. Far field means further away, typically three metres or more - though the exact distances are open to challenge.

The further the position of the microphone is away from speakers, the more room acoustics (bass resonances/modes and reflections) affect the result. Conversley, the closer the microphone is positioned to the speakers, the less the room effects the result but, with multiple drive unit speakers (most hifi speakers are at least two-way), the exact microphone position becomes important because the two or more speaker drivers are, usually, physically distanced so which drive unit do you point the microphone at (see first link)?

In-room frequency response curves I've seen for quality speakers tend to fall away as the frequency increases (see second link).


Thanks for reply!

What did I hear? Well......it all started when I was at the gym and listening on my phone and headphones to an old track by Humble Pie (I Don't Need No Doctor). It is not a great recording, but, there a wonderful bit during the intro where the bass guitar comes sliding in and is a really powerful bit of music. Sounded great so, I thought, this is going to sound awesome on my HiFi system. I got home and whacked it on......what a damp squib.....I could not even hear the arrival of the bass guitar. It just wasn't there! To hear it then you have to play at a loudness that I rarely do! having said that there are plenty of other recordings (Stanley Clarke, Prince) where the bass does seem to be fine!

The system is a good one and would cost £25K+ to buy new, including some very nice Audio Note AN-E SPe HE speakers. They have been tweaked internally with extra silver wire and are the actual ones that AN used in a few HiFi shows a while back. However, they are 10+ years old and I started to wonder if they need looking at. Hence, this is why I am trying to determine if they are failing to play certain frequencies properly. However, I would expect if something is failing/failed then it would be one speaker, whereas this"issue" appears uniform and I am starting to suspect speaker placement/stands/room interaction. Unfortunately, I am not all that knowledgeable in this area and may be measuring wrong/irrelevant things!

Just tried playing Pink Noise and using a Spectrum Analyzer app on my tablet to view. There was nothing obviously wrong. The response was not flat, but, a gentle slope from low end to 150Hz and then flattish and then dropping off around 12kHz. Both speakers looked about same.

I will try the individual test tones again later and go for near to speaker as I am attempting to test them rather than room effects. The speakers only have one drive unit I am interested in, plus the tweeter, so I should be OK pointing sound meter at the right source.

Will have a read of 2 articles you linked to later, when I have a bit more time. Sure they will help.
 

dogfonos

Well-known Member
I'm no expert in this area but a frequency sweep test might be a good idea.

Conducting freq. spot tests could miss the problem freq. if the peak or trough is sharp (i.e. high Q). However, sweep testing should enable you to weed out freq. response imbalances such as quiet or exhaggerated frequencies and speaker and room resonances. Choose a nice slow sweep or the problem freq. may fly past and be missed. Just use your ears but I imagine you could have your analyser running too. In an ideal world, the volume should either stay the same or change smoothly throughout the freq. range (although to some extent that depends on your hearing too). Any nulls or prominent tones can then be investigated further with freq. spot checks - assuming you have a suitable spot freq. to playback.

My best guess would be a room and/or speaker/listener positioning issue. Most folk regard the room as the second most influencial factor (speakers being the first), though sometimes, I think the room can exert the most influence.
 

JohnnyNapalm

Active Member
I wanted to test my speakers and was interested in the higher frequencies. I downloaded an app to my Samsung called ‘PA Tone’ and used headphone out to phono cable (Amazon) and connected it to my Pre Amp. I was then able to play specific frequencies through my system and could turn each channel on/off within the app. I used my iPad to view the frequencies being listened to by way of Decibel X and could compare each channel as it shows frequency and db’s. That is how I found out one of my tweeters was running lower than the other which confirmed my suspicion. I hope this helps. Cheers.
 

MysteryMan

Active Member
I wanted to test my speakers and was interested in the higher frequencies. I downloaded an app to my Samsung called ‘PA Tone’ and used headphone out to phono cable (Amazon) and connected it to my Pre Amp. I was then able to play specific frequencies through my system and could turn each channel on/off within the app. I used my iPad to view the frequencies being listened to by way of Decibel X and could compare each channel as it shows frequency and db’s. That is how I found out one of my tweeters was running lower than the other which confirmed my suspicion. I hope this helps. Cheers.
Thanks, that sounds interesting and I will try this, as well as dogfonos' suggestion.
 

BlueWizard

Distinguished Member
You can get 31 Step Tone and 63 Step Tones from 20hz to 20khz. Though you will have to search them out. You can also get Test Tone CD with not only Tones but Sweeps and various types of noise the most common being Pink Noise.

Before I ran any test tones I would use a 1khz tone to set the general level. It should be around 75db, a comfortable listening level, but not that loud. The thing is, if you turn up the Volume on the Low or High End you will then hear Tones that you normally (normal levels) couldn't hear.

If you are going to plot out the response by hand, best get some Linear/Log Graph Paper, in the form of Linear on the Vertical Axis (210mm) and Logarithmic on the Horizontal Axis (297mm). Very hard to find this in the store now, but you can buy it on Amazon or Ebay.

Phone Apps are good for Relative Frequency Response, that is, one frequency vs another, but the absolute value is questionable unless you have something to calibrate the App against, like an existing SPL (loudness) Meter. That's not a problem, as it will be fairly accurate across the frequency spectrum, but the absolute values can't be fully trusted. For example, if the App Meter reads 80db, it could be anything from 75db to 85db, but it will be the same across virtually all frequencies.

Next CAUTION, Sweeps and Tones are sustained sound, were as music is very dynamic. Those sustained tones can cause heat to build up in your drivers, so you don't want to run sustained tones for very long. When crude testing with a Bass Range Sweep (10 to 300hz) that gradually rises 1hz every minute, I pause every now and then to give things time to cool down. I have the Test Tones burned onto a CD even though I down loaded them, so Pausing is easy enough.

Sweeps, in my experience are not that helpful because you can't hear what you can't hear. For illustration you might have a small speakers so the sweep from 20hz to 40hz is pretty much inaudible. Unless you have someway to view or measure the sweep, or unless they are telling you each frequency as it moves up, you really don't know. More so on the High End, your speakers in theory could product 20kz, but it is possible if you are older that you high end hearing craps out at 15hz. So, if you are relying strictly on Hearing, you would never know that the sweep didn't stop at 15khz.

That's one reason why I like Step Tones, because I know what each tone is whether I can hear it or not.

Here are some examples of Audio Test CDs -

Amazon product
Amazon product
Amazon product
Amazon product
Amazon product
Although perhaps not necessary here, but worth mentioning, they make Surround Sound Test Discs too -

Dolby Surround Test CD: Amazon.co.uk: Music

I have no experience with any of these, they are just the first ones to come up on a quick Amazon search.

Just a few bits of info.

Steve/bluewizard
 

MysteryMan

Active Member
You can get 31 Step Tone and 63 Step Tones from 20hz to 20khz. Though you will have to search them out. You can also get Test Tone CD with not only Tones but Sweeps and various types of noise the most common being Pink Noise.

Before I ran any test tones I would use a 1khz tone to set the general level. It should be around 75db, a comfortable listening level, but not that loud. The thing is, if you turn up the Volume on the Low or High End you will then hear Tones that you normally (normal levels) couldn't hear.

If you are going to plot out the response by hand, best get some Linear/Log Graph Paper, in the form of Linear on the Vertical Axis (210mm) and Logarithmic on the Horizontal Axis (297mm). Very hard to find this in the store now, but you can buy it on Amazon or Ebay.

Phone Apps are good for Relative Frequency Response, that is, one frequency vs another, but the absolute value is questionable unless you have something to calibrate the App against, like an existing SPL (loudness) Meter. That's not a problem, as it will be fairly accurate across the frequency spectrum, but the absolute values can't be fully trusted. For example, if the App Meter reads 80db, it could be anything from 75db to 85db, but it will be the same across virtually all frequencies.

Next CAUTION, Sweeps and Tones are sustained sound, were as music is very dynamic. Those sustained tones can cause heat to build up in your drivers, so you don't want to run sustained tones for very long. When crude testing with a Bass Range Sweep (10 to 300hz) that gradually rises 1hz every minute, I pause every now and then to give things time to cool down. I have the Test Tones burned onto a CD even though I down loaded them, so Pausing is easy enough.

Sweeps, in my experience are not that helpful because you can't hear what you can't hear. For illustration you might have a small speakers so the sweep from 20hz to 40hz is pretty much inaudible. Unless you have someway to view or measure the sweep, or unless they are telling you each frequency as it moves up, you really don't know. More so on the High End, your speakers in theory could product 20kz, but it is possible if you are older that you high end hearing craps out at 15hz. So, if you are relying strictly on Hearing, you would never know that the sweep didn't stop at 15khz.

That's one reason why I like Step Tones, because I know what each tone is whether I can hear it or not.

Here are some examples of Audio Test CDs -

Amazon product
Amazon product
Amazon product
Amazon product
Amazon product
Although perhaps not necessary here, but worth mentioning, they make Surround Sound Test Discs too -

Dolby Surround Test CD: Amazon.co.uk: Music

I have no experience with any of these, they are just the first ones to come up on a quick Amazon search.

Just a few bits of info.

Steve/bluewizard
Thanks!
 

The latest video from AVForums

Podcast: Cleer Audio speaker + HiFi Rose Streamer Reviews & Movie/TV Show talk
Subscribe to our YouTube channel

Full fat HDMI teeshirts

Support AVForums with Patreon

Top Bottom