Do i need anything else?

Hi, hope i have posted in the correct place, if not i apologise.

Stumbled across this page while reading in another forum, which said i may not need a SPL meter.
This sounded great as i dont have one, but i'm a little confused as to how i can use my microphones to take room measurments.
Think i managed to do a soundcard loopback test, though not really sure what info i can gather from that.

I have a M-Audio 2496 soundcard, Joemeek Sixq Preamp with S/pdif output, VTB-1 Preamp, Soundcraft Mixer for my analog outputs to feed my analog inputs on my soundcard.
For microphones i have Shure SM58 Beta, JoeMeek's, Samson and SE Electronics SE200
Oh and for monitoring i have a pair of Yamaha MSP5.

Will i be able to measure room EQ response with the above equipment?

Many thanks

Your microphone will need a calibration file and you also need an SPL meter.

Best to read The REW HELP FILES and this thread.

brucek

Right, thanks for those. And after much......much reading i "think" i'm starting to understand this a little better.
I'd just like to clarify a few things to make sure i do all this the right way.

First things first, i'm using this for a home studio setup with No Sub, just stereo speakers, and i'd like to measure my room response so i can either use corrective EQ in my music software or buy acoustic panels to achieve the desired flat response!

I see that i need the SPL meter to measure the 75db, as this seems quite hard to come by in the UK, other than ebay, is there another way to do this or a way to guess?

I have succesfully done a soundcard Cal, but was wondering if i need to include my mixer (preamp) in these tests, or to wait until i use the mic.

Also i plan to buy the ECM8000 to do my measurements as i cant find Cal files for any of my mics.

Many thanks again for any help with these questions

As I'm sure you know, above the subwoofer range, treatment is the desired route to follow. Perhaps some low Q filters can be used at higher frequencies, but room treatment would usually be advised.

Yeah, you can guess at a reasonable measurement level when you get to that part in REW setup (where it tells you to adjust your receiver to get 75dB at the listening position with an SPL meter).
After its set at a reasonable level, you click the REW SPL calibrate routine (which simply tells REW what your SPL meter is reading). You set that to 75dB. Yeah, maybe it's actually 79dB , but it is not too important.

Some people include their mic preamp into the soundcard cal loopback when they create the soundcard cal file. This then compensates for its shortfalls when measuring. If the mixer is of any decent quality though, it won't have that much effect. If you want to see the mixers actual response to make that decision, do a normal cable loopback soundcard cal file creation, then load that soundcard file into REW and then make a measure of the mixer itself (using the right channel of REW and cables to and from the mixer from the soundcard line-out and line-in). You have to mess with the levels a bit, but it will take a frequency response measure of the mixer and then you can see what it's like, and then decide if you need it in the soundcard cal file.
The mic itself, of course, is compensated by the mic calibration file (which you load manually).

I guess your listening position is the mixer chair position?

brucek

Cheers for the clear and informative response brucek!

Just out of curiosity, i noticed you say right channel and noticed on the mixer picture that it was also the right channel, is there a reason for this? Only metion this because when i record from my mixer i always use the Left channel.

Also (sorry for all the questions btw) i have another preamp which connects digitally to my soundcard (S/pdif) i assume i cant test it this way though?

Yes it is, right in centre between my monitoring speakers, though at present i'm not quite sure thats working too well, hence the need for a room EQ to see my problems , as i think i have some in the lower mid section (either that or my ears have finally given up)

I should hopefully be picking up my ECM8000 Tomorrow so the testing can finally begin, and no doubt many more questions will follow!

Thanks Again

When you use the (soundcard cal file method) as opposed to using the (cable loopback Left Channel Calibrate method), then yes, you have selectable control with REW on the Settings page to use either the Right or Left channel to measure with. I guess I get use to switching back and forth between these two methods, so I have a bit of a crush on using the right channel for measuring.

REW is specifically targeted toward using analog line-in and line-out. It is quite accurate when doing so. I would stay away from SPDIF. REW is only a tool to use to get a room setup and then you put it away. I tell everyone to suffer through the analog usage when using REW.

Yeah, we have many people who use REW for their home studio setup with great results. They get their room all setup and do some mixing and when they play it back, it sucks. So, they seek out REW and do a measure and find that it's their response at the listening position that needs work. Often it only takes a few panels or moving the near-fields or mixing chair a few feet and all is well....

BTW, it's very easy to create an REW calibration file for any studio mic you have, if you've got a response graph that came with it........

brucek

Now thats something i like to hear! Just hope it can do the same for me

Now this seems interesting, mainly because it'd save me money, and to have to go hunting for the mic tomorrow!! I also read on another post here that someone made one for their NT1A Mic.

So i had a look into finding some graphs for my mics and came across 2, one for the SE2000 (condensor) and one for the Shure SM58 Beta (Dynamic)
If you could please tell me which one would be better (if any) so i can have a look into making the Cal file.
If none are ok, then i'll just get the ECM.

SE2000:



Shure SM58 Beta



Cheers

Yikes..... that SE2000 graph shows that this mic is a very capable unit. I wonder if the graph is accurate.

Basically, that's a flat microphone.... from 20Hz to 20Khz It doesn't alter its output by more than ~3dB over that range.

If true, you don't need a calibration file for that mic if you're using it with that bandwidth. It is simply flat.

Forget the ECM8000 if you have a mic that is as accurate as the graph for the SE2000 shows.

brucek

Super.... Thats great to hear

Well i know it's quite a good mic for studio purposes, but didn't think it'd be any good for this type of work, so im over the moon you said that, saves me a few pennys anyways!!

Couldn't tell you if the graph is 100% accurate myself, but i pulled the graph from thier website, and i remember seeing it in my manual aswell, though i cant for the life of me remember where i put it

So i'm guessing i just follow the guide to do my tests, but don't load any Cal files for the mic?
I shall try it anyways, and see what happens.

I must thank you again for your time and patience in helping me out with this!
No doubt i'll be posting graphs for this, so should i just post them here or start a new thread?

Cheers

That's correct....

Here would be good...

brucek

Right.... Had a feeling i may have some trouble with this.
Tried some test's last night and this morning, and as you can probably tell from the graph below, something is not right, (as for one, they look nothing like anyone elses on the forum) though i'm not sure what i'm doing wrong?!

Also should the test sound be only coming through one channel (in my case the left channel) so for eg... you test the left channel for one graph, and then do the right channel on another graph?

Well here's the graph, hope i got the settings right for posting as i checked them with the thread you posted.

Thanks Again

Your graph is quite normal, albeit measured at a slightly low level. Try to measure a bit higher next time and be sure to run the REW SPL Calibrate routine to set the 75dB into REW's SPL meter (after you've set your room level).

Anyway, what looks like a horror show is actually comb filtering, which are reflections from the room's surfaces. When you measure full range, you need to apply smoothing after the measure is taken. Click the smoothing feature to about 1/3 octave, as this is more representative of how we actually hear.

The jaggedness is really too narrow to be perceived, and so to reveal the underlying trend of the signal and to show if treatments are effective, you need the smoothing feature on.

This is not the case for subwoofer measurements when creating filters for EQ's, but you're measuring full range. In that regard though, it wouldn't hurt you to take a look at the area from 50Hz-200Hz with simple axis change to your graph, so as to narrow in on any modal problems you may have to address (which hopefully would be solvable with positioning).....

brucek

Ok, thanks for that, and for the quick reply.

I've had another go, hope this is the right loudness level as i am guessing:



And the zoomed in version:



Also this is for the left speaker out only, as no sound was coming from the right speaker.
Looks like i have a good few dips and peaks though.

Overall it isn't too bad, but there are some issues.

I guess I would look at both speakers individually, and then play them together (simply use a Y-adapter to feed both channels of the receiver). The REW signal is obviously mono, so looking at each speaker by itself will reveal any specific problems to that speaker, but you also want to see what it looks like when the two speakers play in concert.

You have about ~ +/-5dB variation across the spectrum except for the <100Hz area. The dip at 70Hz and the peak at 60Hz will be noticable.

It's also concerning that the signal drops off so rapidly above 10K. I wonder if this is the speakers or the microphone. You would need one of them swapped out with a "known" before you could determine that.

I don't know if you've treated your studio area much, but I would say not considering the amount of relections you showed in the unsmoothed plot?

To try and smooth the response, the RTA feature of REW is very handy. Use Periodic Pink noise with a Rectangular Window and maybe a 1/24 Octave RTA with 65536 FFT and then simply move around the mic and watch the screen. It may reveal a better listening position.

Again, I'm not a studio guy, so I'm just guessing at this point. The guys in the Home Audio Acoustics section would know more about room treatment.

brucek

Thanks for the reply,
Yes, got a feeling that could be the mic to be honest, (so much for their frequency chart) i will buy the ECM and try, cause i'm hoping it isn't the monitors as they are expensive, and "supposadly" Mastering Grade! Though they do have high / low Trims on them so i should prolly try that aswell.

None at all at present, apart from the vocal recording area, which obviously doesnt apply tho the mix sweet spot. I didn't really want to go buying/building wall panels and bass traps, untill i knew exactly what problems i was dealing with, and with much thanks to this program and your much needed knowledge and guidance, i'm starting to see the bigger picture!

A point on the unsmoothed plot....... if my room was treated, (the right way) would i not need to apply the 1/3 octave smoothing?

Would i just play this as a wav file through my monitors and press the RED button in the bottom right corner of the Spectrum Tab window?

Maybe not, but but you've given me such an insight as to where i need to be heading with this, so thanks

Also before, i make a post in the Acoustics section, can i post a L/R and Stereo Graph of the measurements taken with a ECM (which will probably be tomorrow now as the shops have closed here ) just so i can have a professional opion on the problem areas!

Cheers

I can only draw on experience from reading posts from other members here who have studios, not my own, but I remember recently someone posted a plot of his small studio without smoothing and I could hardly believe how smooth it looked. He had just about every inch covered with treatment in the room (he showed a picture)...

No, you use the REW signal generator and press Periodic Pink Noise (not regular Pink Noise) and then go to the Spectrum tab and select RTA's......

Oh for sure. I only suggest the Acoustic guys, since they know about treatment. I think it's best to learn all about measuring techniques here and then you have that as a tool to evaluate any treatment you install.

It ain't a bad idea to get the ECM. It's a good all round omni measurement microphone for a very decent price. They offer a fairly consistent response between units, so the cal file we have here will be fine.... If you want to get crazy, you can get it calibrated, but for home use, it's pretty good with the cal file we have.

brucek

Sorry for my delayed reply, seems as though the ECM Mic is very hard to come by in Birmingham But it finally arrived yesterday so i have had chance to take some more measurements.

also in the time it has taken to arrive it has given me chance to build my acoustic panels which i am certain i will need

Just want to make sure where my problems lie before installing all the panels.

The following graphs are all stereo taken with the ECM mic,

the 1st is the original Speaker Position (spread quite far L-R from the sweet spot)


I thought the Speaker position may be spread too far so i brought them closer in and this was the result:


I've aslo been doing a lot of reading on speaker decoupling and tried some foam underneath the speakers on this following graph:


Hope this all kinda makes sense, and i know for certain i need some treatment but would you say the 3rd graph was the best of the 3?

Thanks again for your help!

You would be much better off measuring each speaker individually rather than both at the same time. When both speakers are running any path length difference between the speakers and the mic results in comb filtering, just like the comb filtering that results from a reflection.

If you use the "Measured" graph group your plots will all appear on one graph making it easier to see the differences between them.

To evaluate the effects of acoustic panels the energy-time curves can be very useful, with the Y axis set to the "% FS" units, as it is easier to see the spikes after the main spike which are due to reflections from room surfaces and how effective panels at the reflection points are in reducing the levels of those spikes.

Ok, thanks for the info, i've had another go and as i have 2 sets of monitors i thought i would try both, so on the 1st graph i have grouped the tests for the left side, and then the right side for the 2nd graph.

this test has be