Audacity Forum

MichloIW wrote:Greetings,

after needing to catch up on other things as of tonight and the coming days I shall be recording again.

Does anybody have any further advice on the minor changes ACX wanted me to make please? In particular, where I ought to add the compressor step to my current list? I'm hoping I won't need the EQ as I'm going to try harder not to have peaks as mentioned by ACX and here in this forum (also I'm not certain how to work the EQ range ACX and members here are talking about). Plus, any suggestions on where I ought to position the microphone and mudguard to lessen the hard surfaces I can't cover up yet please?

This is what ACX said: "I also recommend using a bit of EQ to take out a bit of a resonance. I found that taking a bit out around 600 Hz helped smooth out some of this issue. I also used a compressor with its threshold around -8dB, and ratio set at 4.5 to 1 to reduce the dynamic range a bit. This especially helped with the louder peaks of exclamation “Let’s Go!”

Hello Ian,

It is a pity that I can't see the pictures of your home recording studio.
Can you give me the dimensions--length, width, height?
I can at least calculate the room modes, i.e. the frequencies that resonate, from those.
No matter if you don't want.
There are certain frequencies that are boosted due to the room dimensions. Especially those under 400 Hz. Now, it seems that your cabin is fairly small (above 2500 cubic feet is ideal) and thus some frequencies in the mid-range are boosted. This can't be avoided, unless you tear down the walls and make a free field studio.
The trick is now to find the sweet spot from where to speak from. Near a wall is bad, it boosts the most. In the middle is also bad, it will prefer odd harmonics and damp even ones.
Strange enough, the ideal point is the golden ratio (1:0.618034), approx 38 % of the dimension.
So, if the room has 10' X 5' X 8', you would start by moving 3.8' from a shorter wall and facing the other short wall, 1.9' parallel to the long wall.
The same is true for the height, but is scarcely applicable.
The microphone would now be about 1' in front of you.
I can't tell how your mud guards etc. must be placed, I think they have some instructions attached.
In general, bass traps are placed at all possible corners, where the walls meet, the walls the floor and the walls the ceiling.
Foam absorbers, diffusors and similar are placed:
- at head-height behind and in front.
- at head-height, on the long wall, between you and the microphone (in our example 3.95' from the short wall)
Very important is to place one at the ceiling too and perhaps at your feet, if the carpet doesn't absorb enough. However, I think you've already covered all walls.

Besides the standing waves, there are also the comb effects we have to include. They are created when your speech bounces off a surface and arrives at the microphone with a min delay. This is especially noticeable in small rooms.

You might see the dilemma now:
You can try to find the mentioned sweet spot by some amount of trial and error or you can use my submitted eq-curve to correct the resonance in post production.
The first option is preferable since it is always better to correct on the input side.
However, if you start to move around, the spectrum will be changed and my eq-curve won't be valid anymore which is bad because you can't avoid equalization anyway (due to the room properties).
In other words, I would have to find a new profile fitting to the recordings you are going to submit next.

I actually find that the Cattle1 eq-curve isn't too bad. We can include the roll-off for speech and perhaps remove the boosts above 3150 Hz (they become piercing at high volume playback).

By the way, don't meddle equalization with compression.
The EQ is not against the mentioned peaks "Let's go" etc.
You can't control the resonances by your performance (unless you whisper ).

The peaks (which are actually just certain words that do have too much energy) I will try to treat in another post.

I can't see Koz' point that the boominess is scarcely noticeable and not important therefore. In my opinion, it makes the difference between a cheap YouTube video and a ambitious production that should be sold in the end.
I usually do not bother to give feed back on projects that I don't really care for.
It hurts to see your children not to go to any length to make the most of themselves.

Rob,

I did put up a link to the pictures a few posts back but you must have missed it. The room is 5x5x8.

Here you are:

https://onedrive.live.com/?cid=ff725192 ... 8Nf1t1nwlY

I'm reading this just as I go to bed so I shall re-read it tomorrow but at least you have more to work with now.

I truly appreciate yours and everybody's time, even your problem child.

Cheers.

MichloIW wrote:Rob,

I did put up a link to the pictures a few posts back but you must have missed it. The room is 5x5x8.

I didn't miss it. I meant it literally because I'm blind.
I gather from the text that there are some gaps between the tiles. You don't have to buy a bunch of new ones, it is enough to cluster them to the left, right, front, back and ceiling. This will lower the so called axial modes.

Here they are for your room:
(sorry for the wrong layout) This gives us a rough idea which frequencies are boosted.
You don't have to fully appreciate those modes, it isn't that important. It is rather a tool to design a room from scratch.
However, one thing gets clear at a glance, if the length and width of a room are equal, the same frequencies are boosted which leads to a fairly high peak.
Some posts back, we've mentioned 450 Hz and that's exactly where two such axial modes overlap.
There are also some of the less strong tangial (comparable to a billards shot involving all 4 edges and stopping at the starting point) and oblique modes that do reside nearby. Those frequencies pile up and produce the boominess.
Moving the mic and yourself around will change the balance between the modes. Thus the resonance at 450 Hz could disappear and one at 900 Hz build up, but there the tiles will show more and more effect at higher frequencies.

We know why the room sounds like that. It's not fully soundproofed.

Each square soundproofing panel is about 1/3M on a side, but there are not enough of them to cover each surface, so they are spread out: panel, bare wall, panel, bare wall in both directions. On average, only about 2/3 or less of each wall or surface is covered. The echoes and enclosure ringing are only reduced by 2/3.

Koz

That was my semi-funny joke of putting the heavy winter coats back in the room with you to get additional damping. Koz

kozikowski wrote:

/snip

There is, as usual, a trick. Crumple a sheet of LA Times in front of you in the studio — roughly where your head normally is. That's as close to "white noise" as you can get with a tool you can buy at 7-Eleven/Tesco. I have samples of this test (somewhere) and I use it to indicate differences between microphones. That's how I certified the asian knock-off of the Shure SM58. You may be able to use it to see where your honk peak is.

Koz

Koz and Rob,

I'm sorry for the delay. I have to admit that whilst having a lot going on trying to earn money so that I can actually do this I was also a little disheartened. Koz put that into perspective for me and it really helped.

Rob, thank you very much for all the calculating you've been doing. A lot of it is over my head but I appreciate the time and I know understand why you have the golden ear.

I just did the paper trick finally. I used a brown paper bag so I hope that is sufficient.

Here are the raw files:

https://onedrive.live.com/redir?resid=F ... =folder%2c

It looks like I may be able to acquire more tiles soon but not in time for this first book. I need to get it done. I'm going to try to hang some clothes as suggested and then start over. I hope your patience will last a little longer as you hear chapter 1 yet again.

Cheers.

Any heavy clothing and you don't have to cover all the surfaces. Nodal Ringing depends on two parallel live surfaces, so you really only have to kill one wall (each direction). Particularly since you already have most of the sound damped with the foam panels.

But yes, I agree, you should not change anything in the middle of a book. Do we have a sample of the room in which you're going to perform this latest book? I'm betting between all the elves we can come up with a relatively simple equalization curve to get rid of most of the "small room honk." We can make this curve while you're recording since corrections can happen in post production.

There is provision to send you a curve. Did you know that? You don't have to do this second hand in narrative or pictures. Effect > Equalization can import a pre-baked curve for your use. We just have to get the elves across ten time zones to agree on which one to send.

I'll make a fresh cuppa....

Koz

https://onedrive.live.com/redir?resid=F ... =folder%2c

Thank you, but Kraft Paper (brown paper bags) doesn't have the crisp snap that a fresh newspaper does. It produces a muddy sound.

We'll see how this works.

Koz

Download...

I can't tell what you did. AUP files don't work without the associated _DATA folder and I don't see any _DATA folders. The WAV files are the only ones we can play.

You're going to have to describe what those test tones are. The description in the file block gets cut off.

Koz

Koz,

thanks for the reply.

The data folder is in there. I saved directly to my OneDrive.

And good, that is what I did already, suit on one wall, sweater on the other and the comforter on the door (which has the most free surface and is behind me).

As for the spectral tests, they were requested of me earlier in the thread (http://forum.audacityteam.org/viewtopic ... &start=190). The descriptions are in the wav format files.

Cheers.