Correct your speakers with REW, UMIK-1 and Equalizer APO (Room Correction Tutorial)

Hey, Julian Kraus here and in this video I'm going to show you how to measure and correct the frequency response of your sound system with RoomEQ Wizard in combination with the U-Mic One measurement microphone and equalizer APO. So the first thing you have to do is to download the calibration files specific for your microphone. You can download them on MiniDSP's website. There you simply have to enter the serial number of your particular U-Mic. You can find your serial number on the mic and on the packaging. After typing in your serial number you download two files. One normal one and one for 90 degrees measurements. Ok, now you plug in the U-Mic One and open up RoomEQ Wizard. Make sure you have the newest version because it will automatically recognize the U-Mic One and will say U-Mic detected, do you want to use it for measurements? Click yes and it will ask you if you have a calibration file. Click yes again and then browse to the location where you saved the calibration files. Select the normal calibration file because for our measurements we will point the mic at the speakers. After you selected the calibration file you simply click ok. The U-Mic is now correctly calibrated. If REW didn't ask you for the calibration file, still follow the next steps and you'll be able to import the file later on. Before we can make any measurements we have to go into the preferences and setup everything correctly. For now we only have to look at this first tab called Soundcard. Here we will first look at the input device. The U-Mic One should already be selected. Now under input we go from default to microphone and if you use Windows tick the box which is labeled control input mixer slash volume. The input volume should be set to 0.53. Now we configure the output device. Here you select the device where the sound you want to measure is put out. In my case I have an NAD amp connected to my PC which drives my speakers. If you have active monitors connected to an audio interface you would select the interface here. The output can be left at the default options. The next step is only necessary if you didn't already load the calibration file. Click onto the mic slash meter tab and here you select mic or z-weighted SPL meter. Then hit browse and select the U-Mic calibration file. Now check if the settings in the Soundcard tab are still correct. Finally close the preferences menu. It is time to setup the U-Mic in the correct measurement position. Attach the U-Mic to a stand with the provided microphone mount. Point the mic in the middle of your speakers and place the mic at your listening position. The tip of the mic should be at ear height and pretty much exactly where your head would be when listening to music. Next we go back into REW and here you click on SPL meter and generator. This opens up two separate windows. In the SPL meter window you make sure that the button SPL is selected. Then check that the middle button C is selected and the button labeled S is pressed. After that you hit the red record button. This will activate the SPL meter and show you the current sound pressure level at the microphone. Let this run and now focus on the generator window. Here you select pink noise from the drop down menu and select speaker cal on the left. The RMS level DBFS should be set to minus 12. Now turn on the volume of your sound system. I control the volume directly on my amp. If you use an audio interface for example you would turn down the volume there. Hit the green play button and pink noise will start to play. Slowly turn up the volume on your audio system and watch the number in the SPL meter panel. Turn up the volume until you have reached about 75 dB. This doesn't have to be perfect, plus or minus 1 dB is totally fine. After that you can hit the play button again to stop the noise and close both windows. It's now time to do our first measurement. Click on the measurement icon in the top left corner. This opens up a new window. Here we choose 10 Hz as our start frequency and 20000 Hz as the end frequency. The level is set to minus 12 dBFS and length is set to 256k. To check that everything is set up correctly you can click the button check levels. But be warned it will put out pink noise like before and this might be pretty loud. Cover your ears and check the levels. The noise stops after a few seconds and it will say something like this. The exact number here is not important. It will typically be around minus 45 to minus 50 with the U-Mic 1 and it should say level OK. Now step away from the mic a bit so that you do not interfere with the sound of your speakers. Then hit the start measuring button. If you are in the way the moment you click this button you can set up a delay of a few seconds here. If you then hit start measuring it will count down which lets you move out of the way and will then start the measurement. Once you hit the start measuring button a sine sweep will be played by your speakers and recorded by the microphone and the frequency response of your audio system will then be displayed in a graph. First click the all SPL button and then the 20 to 20000 button in the right corner to see the full frequency response. The response looks a bit messy but that's because it's not smoothed yet. I recommend to click on graph and then select the VAR smoothing option. This smooths out the frequency response to more closely match how we as humans perceive it. And one thing you should do is to name your measurement before advancing further. So give the measurement a descriptive name. Now I recommend to make some more measurements at slightly different positions because your head will not always be 100% exactly where the measurement mic is placed and we want to get an average reading of your listening position. With the U-Mic 1 still set to ear height move the mic 4 inches or about 10 cm to the right and do another measurement. Then move the mic a bit more to the left of your center position. You can do one more measurement in the middle and slightly more closer to the speakers and one more further away. And I even go to the center position again and make one measurement slightly above and one below the listening position. This way you will end up with 7 measurements all in your listening area. Now in REW in the all SPL tab make sure all the boxes of the measurements are ticked and then hit the average the responses button. This creates a new frequency response which is an average of all your measured responses. Before going any further I recommend you save all the measurements. You do this by clicking on file and then save all measurements. Here you can add a note to your measurements which you will be able to see next time you open them up again. Now you can untick all the boxes to hide the graphs from the chart and only leave the average response enabled. This is the average frequency response of your listening position. Of course this will look a bit different than my example here. You might already have a flatter frequency response to start with or it might even be worse than my example. As you can see in my response I have a few peaks and troughs. This is a very common frequency response curve for an audio system in a small room and this is what we are going to correct with EQ. Before we can start to EQ anything we have to set up our house curve. What's a house curve? Well this is our target curve we want our frequency response at the listening position to look like. And this is where it gets a little bit controversial. Some people say to get the perfect sound you have to EQ everything flat so that in the end your frequency response resembles a flat line on the graph. But research has shown that when you place speakers with a flat frequency response in a room they will have more pronounced bass due to room modes and a slight drop-off in the higher frequencies. Harman did a few studies on that and in the end they came up with the preferred listening curve. I modeled this curve with a few data points, put them into a text file and can then load it into REW and use it as a house curve. I will write the data for the house curve into the description of this video so you can simply copy it and paste it into a text file to save it. After that you go into the preferences in REW. Here you can find the tab house curve. Click browse and select the text file you just created. We can finally start to EQ our frequency response to match our house curve. Exit out of the preferences and make sure you still have the average frequency response selected on the left. Then click onto the EQ button at the top and this opens up a new window. The first thing you have to do is to open the equalizer section. We are going to use equalizer APO later to EQ our system so in order to export the EQ filters correctly you have to select the point generic. Next go into your target settings. Turn the LF rise slope and the HF fall slope to zero. If you didn't load in a house curve you could do a high frequency roll off or a bass boost here. But I think a proper house curve is the way to go so you shouldn't need that here and have to make sure that they are both turned to zero to not mess with your house curve. After that you have to select the appropriate speaker type. You would select full range if you have very big speakers which are able to reproduce the whole frequency range like floor standing speakers, bass limited when you are EQing smaller bookshelf type speakers or subwoofer if you are only working with a subwoofer. The next point is crossover. If you have sealed speaker enclosures you would select 12 dB per octave and if your speakers have a bass reflex port you would select 24 dB per octave. Cutoff determines the point where your speaker starts to roll off in the lower frequencies and for bookshelf type speakers this is usually somewhere around 80 Hz. You can finesse this value a little more later on to make sure that the blue target line in the graph matches the roll off of your speaker. Next you have to set your target level. This shifts the whole target curve up or down. When you click set target the curve will usually be moved to a decent starting position. Generally speaking you don't want the target to be too high otherwise only the very tips of your frequency response will be EQed and when you set the target level too low the whole curve will be EQed and this pulls down the whole audio level and you will have to crank up your volume on your sound system very high to get a decent output. So it's a balance. You don't want to go too high otherwise the EQ correction doesn't do anything but you also don't want to go too low because you lose a lot of volume. So make sure the blue target curve is roughly in the middle of your frequency response. Now it is time to generate the EQ correction filters in the filter task section. First you have to specify the range the filters should be calculated for. Because a typical human can hear from 20 to about 20 000 Hz that's the default range you should use. Individual max boost describes how much each filter can boost a certain frequency and I would recommend setting this to about 6. The overall max boost should be zero. This means that no troughs in your frequency range will be boosted. This might seem like a small drawback but when you boost a trough it will reduce the overall volume and you have to use more volume on your audio system and use a lot more amplification. I don't think that's worth it so leave this set to zero. And the flatness target defines how closely the EQ curve will match the target curve. 2 dB is a good starting point but you can go down to 1 dB to get more precision. To generate the EQ correction filters click the match response to target button. REW will calculate the filters and this might take a short time. After that you're going to see the filters and a preview of how your frequency response is going to look like after correction. Okay now you have to apply these EQ filters to your system in order to correct the frequency response. For that click export filter settings as text. You can then add a note and hit save. Now you have to open up the configurator of equalizer APO. Check the box for the output you want to enable EQ APO. In my case I want to use it with my NAD amp which is connected to my PC. After that you have to restart your PC to get EQ APO to work correctly. After the restart you have to find the correction text file you just created with REW and copy it. Now open the configuration editor of EQ APO. Here you click file and open and this opens the location where the correction files are placed. Simply paste the correction file here. Then select the file and hit open. The correction file is now opened in a new tab and you can see the correction curve in the analysis panel. Due to all the filters it might happen that at some point in the spectrum your curve goes above 0 dB. This is indicated by the red peak gain in the bottom left corner. If this is the case you have to click the small green plus which says add filter. Now choose basic filter and then preamp. Here you have to turn it down so that the peak gain says 0.0 dB. If you change a value and the change is not directly reflected in the analysis panel you have to tick the instant mode box. If this mode is not enabled you have to save the file by pressing ctrl s and then the changes will be applied. Okay your room correction file is now correctly implemented into EQ APO but it still does not have any effect on your output because only the filters in the config tab are applied to the selected output. So go into the config file tab and here select all the filters. Now click add filter, ctrl and then include. Here you can click on the file icon and then select your correction file. Finally check if the device at the top is the device you want EQ APO to equalize and then save everything. And that's it. If you want you can make another measurement with your mic at the listening position and you should see a considerable improvement in your frequency response. If you're not satisfied with the results you can go back to rew, play around with the target settings and filter tasks, generate a new correction file and import it into EQ APO the same way as shown. In the config tab you can add another include and include the second correction file. With the power button for each correction file you will then be able to quickly toggle between the correction files and see if you made an improvement. Okay that's it for now. If this was useful to you please give me a thumbs up and subscribe if you didn't already do so. There are more videos to come so we'll see you in the next one.