Printable View
Redacted from a support email:
[QUOTE]We want to replicate a study that used a voice activated relay to collect RT data. The participants speak their responses in a Stroop Task. Do you have a hardware set-up that could do this for me. I am NOT handy and do not think that I could construct this from components without really good directions.
[/QUOTE]
First of all, DirectRT comes with a voice rt stroop sample (see the samples folder) that you can try to test with your current systems. So long as you have decent soundcards, you shouldn’t need any extra hardware.
If you're really, really interested in capturing the voice responses in recording-studio quality, we can recommend several great microphones, both lavalier (clip-on) and stand mounted.
They'll do a fantastic job but will likely cost more than the computer you're connecting them to!
That being said, I'd head over to the computer section of your campus bookstore. They will probably have a large selection of microphones which plug directly into your computer's sound card. You can choose between clip-on, stand mounted and even headset mounted, similar to what telephone operators wear.
The sound quality will be just fine for what you're doing, and I'd be surprised if the cost is more than about $25 per unit.
Let me know if you don't find anything at the bookstore and I'll find you some online candidates.
Cheers,
John
----
John Chapman
Empirisoft
I would add to that simply that you will want a good quality sound card for optimal results--[I]Creative[/I] makes good stuff (see their [I]Audigy[/I] and [I]Live[/I] lines) both internal (for standard PCs) and external (e.g., for laptops via USB). They also sell headsets with both earphones and microphone in one unit. I would avoid using a wireless (e.g., bluetooth) microphone if RT precision is an issue. See [URL="http://www.creative.com/"]www.creative.com[/URL]
-Blair
A related exchange:
[quote]
I need to record voice responses from my participants in 30 second intervals at 9 different occasions during the study. I found the setup for the voice recording in the DirectRT help manual, applied it, and it is recording perfectly. The only problem is that the QUALITY of the recordings is terrible. You can barely tell that a person is talking in the .wav file that is generated. I need coders to be able to listen to and code each audio file, so this is a problem. I have the "RT threshold" down at 1 (on the soundcheck option in DirectRT), and the microphone seems to be registering, it is just the .wav file that is very static-y and inaudible.
The professor that I am working with on this project suggested that I email you and see if you had any suggestions for improving the quality of the sound file. Any insight that you have to offer would be greatly appreciated!!
[/quote]
What version of DirectRT are you using (i.e., 200x.x.x)?
[quote]
Thanks for getting back to me! I actually fixed my problem...the original mic i was using to record was faulty, so once i switched microphones, the sound quality was fine. :) Thanks!
[/quote]
Just curious, what kind of mic was your old one and your new one? Just in case anybody else has similar trouble in the future.
[quote]
The old microphone was borrowed from a friend. It is a Sony electret condenser microphone (as you can probably tell, I am not an audio guru, so I have no idea what that means). It runs off of an internal battery.
The current microphone is a headset/microphone that plugs into the computer via the sound card (output) and the microphone (input) jacks in the back of the unit. It was purchased at a RadioShack. Hope that helps!
[/quote]
Note that as of v2006, we increased the sampling rate of voice recordings from 11,000hz to 22,000hz making for better sounding files while keeping overall file size reasonable.
I'm running a study that measures response times on a pronunciation task. I'm wondering how I can check my sound card to see if the computer I'm planning to use is measuring precisely (as much as possible) to give me accurate data. Also, is there a way to record the response (at the onset to make sure it was indeed the response of the word on the screen and not some background noise that triggered the response)? I'm guessing a standard microphone will be fine for my needs, but again, I'm wanting the RT to be very accurate. Is there a difference in the quality of microphones in terms of their ability to measure accurate RTs?
As far as microphones go, see above. Doesn't usually make a big difference but you can get significant variability due to extraneous noise if your sensitivity is set too high. And you can miss valid responses if you have it set too low. The best way to validate your RTs is to:
1. Check out this link and the section in the the users guide on voice RTs and setting up the sensitivity: see [URL]http://www.empirisoft.com/support/showthread.php?p=269[/URL]
2. Collect some sample voice RTs and save the responses as wave files (see user's guide). Open the samples up in a freeware audio editor like Audacity and compare the RTs reported by DirectRT with the onset of the voice responses as indicated in the audio file editor.