Convolve with me
I've been interested in the idea of digital room correction for some time. After following a thread about a year ago on the AVS forum, then a subsequent more recent thread, I decided to give it a try. The first of these threads was started by Jones Rush and they deal with using a PC and some low cost additional hardware. This allows you to get your feet wet with digital room correction without a vast outlay.
The only disadvantages to doing this on a PC are as follows:
1. You will have to playback your music via the PC
2. Playback will be limited to standard redbook CD or files at 44.1khz
3. It follows from the above that only digital sources will benefit.
As far as the first restriction is concerned I had the idea that I could record the output from the software player to my hard drive and subsequently burn the resulting file to a disc. That disc could then be played on a conventional Transport/CD player that plays CDR's. It will only make sense to play it back in the same room where you've applied DRC, of course. Otherwise the results will be unpredictable. I haven't tried this yet; but will give it a go this weekend.
Nothing can be done about the second restriction. There's nothing to stop you playing back a file at a sample rate greater than 44.1khz; but you will get some weird results. In my case, in some ways, the sound seemed processed and certain aspects could be exaggerated. e.g exaggerated, boomy bass, and part of the reason to apply DRC in the first place is to get rid of that sort of problem. There's nothing to stop you from resampling the file after DRC is applied, though. Using Foobar 2K you must make sure the resampler is applied after the convolver.
The guy who started the first thread on AVS forum has written very clear and precise instructions on how to go about this and there's no point in my reproducing that here. I have provided a synopsis of what's involved, though, since I'd like to try and encourage as many people to try this as possible. I've found the results very beneficial.
Here is a link to download the full instructions in pdf format:
http://www.mooneyass.com/DRC/Here's the synopsis of what's involved:
Download Cooledit Pro plus the Auroroa plug-ins
Download the DRC program
The links are provided in the instructions from Jones Rush.
Use a microphone preamp and a microphone. I ordered the exact models in the instructions. They cost me Euros122. You will also need an spl meter and a multimeter.
I connected the mic to the mic preamp using xlr cable and the preamp to the sound card on my PC using the tape out via RCA to minijack.
Set the mic up at your listening position.
My sound card is the (much maligned by Audiophiles) Audigy 2. If you have a better sound card, you'll probably get better results.
In the instructions he boosts the mic preamps gain to just below clipping, then sets the mic input gain on the PC (Sound card, recording, Mic-in volume) to minimum. This is how he obtains minimal s/n ration when recording. With my sound card I had to do the opposite, otherwise I overloaded the mic input on the sound card. I had to set the mic input volume on the PC to max and just use a small amount of gain on the preamp. You'll need to make a few test recordings to determine this. You'll want to make sure the sound cards' mic OUTPUT volume is muted, otherwise you'll get feedback.
When you record from within Cooledit pro using the mic, you want to make sure it peaks at no greater than -3db.
The mic preamp will only be outputting to one channel, which could be either the left or right. You need to determine which it is in order to get the setting right in CoolEdit pro.
When playing the test sweeps you want to try to get your spl meter hovering around 85db, whilst at the same time making sure you don't go over the max volume your speaker can handle, as described in the instructions.
With the frequency sweep hovering around 85db (assuming you can get away with that without hurting the speakers) and the recording peaking at -3db with minimal s/n ratio, you're ready to start.
BTW, I had to stuff cotton wool in my ears while playing these frequency sweeps.
Next you create a full frequency sweep in CoolEdit and use your PC Sound Output balance control to move the sound entirely to one channel. Either the left or right channel, it doesn't matter which one you start with. Play back the sound and record it with CoolEdit as instructed.
These files are both treated with the Aurora plug-in as per the instructions.
Repeat for the other channel.
Now you save the resulting files of the frequency sweeps for each channel to your hard drive and run each one through the DRC software.
(The DRC software has different options. NOTE: The instructions use the 'normal' setting and I followed those. At the end of the exercise I repeated from this step to produce strong files. This gives me the option of using either normal or strong correction in my software player (see below).)
This results in two room correction files for each speaker. Only one of each of these files is used for each channel.
Each of these new files is then opened within CoolEdit Pro and combined into a single stereo .wav file as per the instructions.
The software player you'll need to use is Foobar2000. You'll need the special installer which includes the 'Convolver' DSP plug-in:
http://www.foobar2000.org/foobar_special.phpWhen you've got foobar2k installed, go to the Preferences->DSP manager and make sure Convolver is in the active set of plug-ins. Then click on Convolver on the left hand side. Within the Convolver set-up area there is a button which allows you to select a file. Select the stereo .wav file that you produced as your final result within CoolEdit Pro. Set the effect slider to the max.
There is also a clipping protection slider and Jones Rush included instructions for determining what the value should be. Set too far below zero and you will compress dynamics, set too high and you will get clipping. I found the best setting just by experimentation. -3.5 works well for me on most recordings.
Like I said, I found the benefits to be really worthwhile. At first, after all the work, I was expecting a huge dramatic difference, such as amazing soundstaging. It didn't seem to sound so different at the beginning. As I listened further though, it gradually dawned on me that the sound had improved in a number of important ways.
I can listen into recordings like never before and hear things going on in familiar recordings that had eluded me before. If this sounds like an ultra detailed sound, well, it is; but certainly not in the sense of losing the meaning of the music. On the contrary, the correct perspectives are brought out along with the detail. Subtle details can now be heard; but they are not brought forward in favor of the main instruments. The relative loudness and softness of the various instruments now being correct or closer to correct than before Of course, I can't know this for sure. All I can say is, it sounds right. This makes it easier to follow all that's going on in the music and at the same time respond naturally to the music. It's definitely not a case of detail swamping musical meaning. Rather subtle details are brought out because they are not being swamped by what was before a relatively muddy mid-bass.
In fact you may feel there is less bass at the beginning (depending on your room and what correction was necessary). This is a common reaction, since many rooms contribute to a bloating of the mid-bass. The bass will not be lacking, however. You should find it tighter and the timing may be better conveyed.
This was using the 'normal file. When I swapped this for the 'strong' room correction file. all of the above applied, perhaps to an even greater extent. Additionally, there is a kind of 'transparency' that I've never heard before. This is most apparent on vocals. A singer's voice is uncannily 'there' in my room, even when heavy eq has been used on the disc. It's as if the room correction has somehow separated the actual vocal from the eq added to it. What this means is that on certain studio recordings I can hear a phenomenon which I can't quite identify. It could be the eq or the recording booth or some combination of the two. Whichever it is, it does sound processed. It's slightly phasing and metallic sounding. That's the best way I can describe it Of course, I'm not ruling out that it could be that this method of digital room correction isn't perfect or I didn't do it perfectly and what I'm hearing is a synthetic effect of the room correction itself. If that's the case though, I can't hear it effecting the voice itself or any of the other instruments. Weird. Maybe someone else might have heard this and be able to explain.
In any case, overall, I find the results very positive.
It would be great if more people can try this approach and post your results here.
Have fun!
geoff
