### Table 2: SNR gains of order and co-phasing algo- rithm when N = 3 in Ricean fading environment.

"... In PAGE 5: ... Then there holds c2;2 = Z 1 0 Q1 p2K; p2(1 + K)r 2dr: Closed{form solution for this integral is not avail- able, and therefore we have computed it numeri- cally. The SNR gains corresponding to order and co-phasing algorithm when Ricean fading is as- sumed, are given in Table2 when N = 3.... ..."

### Table 1. SNR gains in [dB] for =30 o (top) and =60 o (bottom).

2001

Cited by 6

### Table 1: SNR gains of order and co{phasing algo- rithm when N = 3 in Nakagami fading environ- ment.

### Table 1: SNR gains for the training and test sweeping sinu- soids with input SNR 10 dB and 0 dB. Signal NN1 NN2 NN3 NN4

1997

"... In PAGE 2: ... In Table 1, the achieved SNR gains are shown when the trained NNs are presented with their training signals, and test signals which are the same sweeping sinusoids with new AWGN sequences added. From Table1 it is seen that NN4 has not been able to learn its training signal, whereas NN1 oper- ates seemingly well in differing noise conditions. The NN1 output for predicting a test signal with SNR of 10 dB is shown in Figure 2.... ..."

Cited by 1

### TABLE I. PEAK SNR VS OPAMP DC GAIN COMPARISON

### Table 1: Source vector SNR gain performance of our algorithm as a function of microphone noise corruption level and the number sources and microphones. All gains are calculated in the time

"... In PAGE 5: ... Looking at the re- sults, we can see that our variational inference algorithm is able to yield a dramatic improvement over the norm- constrained inversion based estimate, and recovers a high fidelity estimate of the underlying source, despite the fact that there are two more sources than microphones, and the sources have strongly overlapping spectral-temporal fea- ture content. Table1 summarizes the SNR gain results obtained (relative... ..."

### Table 1: Source vector SNR gain performance of our algorithm as a function of microphone noise corruption level and the number sources and microphones. All gains are calculated in the time

"... In PAGE 5: ... Looking at the re- sults, we can see that our variational inference algorithm is able to yield a dramatic improvement over the norm- constrained inversion based estimate, and recovers a high fidelity estimate of the underlying source, despite the fact that there are two more sources than microphones, and the sources have strongly overlapping spectral-temporal fea- ture content. Table1 summarizes the SNR gain results obtained (relative... ..."

### Table 1: Source vector SNR gain performance of our algorithm as a function of microphone noise corruption level and the number sources and microphones. All gains are calculated in the time

"... In PAGE 5: ... Looking at the re- sults, we can see that our variational inference algorithm is able to yield a dramatic improvement over the norm- constrained inversion based estimate, and recovers a high fidelity estimate of the underlying source, despite the fact that there are two more sources than microphones, and the sources have strongly overlapping spectral-temporal fea- ture content. Table1 summarizes the SNR gain results obtained (relative... ..."

### Table 1 below summarizes the various Trellis coding schemes. The 512 state Feedforward Convolutional encoder is the clear winner giving the highest SNR gain Coder Type States Coding Gain Latency Modulation

1998

"... In PAGE 5: ... Table1 : Published Data on the Performance of Various Trellis Schemes (5.1 dB) with lower latency than other schemes with the same gain.... ..."