### Table 1: The detection error rates of the time-domain source detection technique

"... In PAGE 3: ... The values were averaged over the neighboring 21 frames. The detection error rates are summarized in Table1 , where labeling a source as active when the source was inactive is defined as the false alarm rate, and labeling a source as inactive when the source was inactive is defined as the miss rate.... ..."

### Table 5. Computational complexity per symbol, time-domain filter length L, frequency- domain block length N.

2005

"... In PAGE 59: ... The time-domain equalizers are assumed to have length JL, which is the minimum lter length covering the receiver ISI window and consistent with the simulation con guration used for the results reported in the previous sections. An approximate operation count of the original recursive algorithm [36], the proposed square root recursion, the time-average approximation, the frequency-domain equalizer, and the frequency-domain matched lter approximation are tabulated in Table5 for a single equalization iteration. An approximate total op count as a function of the num- ber of separable channel multipaths for a 4 4 MIMO setup is also depicted in Fig.... ..."

### Table 4.1: Prediction gain vs. causal lter order (v) and anticausal lter order(w) thesize the speech. As described previously, in time domain speech coding systems the speech is reconstructed from the model by exciting the lter by the quantized excitation signal. For mixed-phase models, the impulse response of the lter is com- posed of decaying exponentials in positive and negative time, resulting in a noncausal system. This introduces possibly in nite delay into the reconstruction algorithm, and makes the traditional \frame-by-frame quot; closed loop excitation quantization algo- rithms impossible. Hence, although accurate models of the speech signal can be found using mixed-phase all-pole lters, problems occur in synthesis because the resulting lter model is used to lter the excitation signal. Extracting the appropriate phase information from the lter without the use of noncausal in nite impulse responses is di cult.

1994

Cited by 1

### Table 1.2: Data sequences in time domain

1997

Cited by 2

### TABLE II. Bandwidth requirements of time domain multiple access methods.

### Table 3: Advantages and disadvantages for separation in the time domain or separation in the frequency domain.

### Table 1. Time-domain features (averaged) Class Energy (x10-4) Energy Range Zero-Crossing

### Table 2. Results using only the time-domain features with different number of classes and different window lengths using Neural Network classifier.

in ABSTRACT

### Table 4. Results using only the time-domain features with different number of classes and different window lengths using kNN classifier.

in ABSTRACT