Table 1. Cause and solution for the preecho problem in audio coding.. Cause Tradeoff between audio quality and compression efficiency Presence of an attack Insufficient power in preattack section to mask quantization noise Solution Adaptive block-size depending on the source signal Staircase gain to make the attack flat Staircase gain to pre- and deemphasis to adjust the power complementary gain successfully compresses the quantization noise introduced in encoding so that it is inaudible. Table 1 compares different views of the preFigure 5. Another solution to preecho: Application of staircase gain before MDCT as in (a) and complementary gain after IMDCT as in (c) to erase attack. Deemphasis in the silent section successfully compresses the quantization noise introduced in encoding. power in the preattack section to mask the quantization noise. A natural solution to this problem is emphasis of the source signal in the preattack silent section11. Either solution can be implemented as a staircase gain in the time domain as depicted in Figure 5(a). The gain is greater than unity to amplify the silent section whereas it is unity after the attack to keep the same magnitude. By this preemphasis process, the signal before MDCT looks like the one in Figure 5(b) where the distribution of signal magnitude is more even than in (a) due to the staircase gain. Shown in Figure 5(c) is the signal after quantization in the encoder, dequantization and inverse MDCT (IMDCT) in the decoder. Quantization noise is evenly distributed across the whole block. The antistaircase gain which is complementary to staircase in Figure 5(a) results in the decoded signal in (d). The deemphasis process with the May/June 2021 echo problem and the corresponding solutions. For the same problem, it is possible to identify different causes based on different views, each of which leads to different solutions thereto. -GaNCrystal Fabrication for Blue LaserDiode A similar example is found in the field of device technology rather than information technology. The 2014 Nobel Prize in Physics was awarded to three people who developed high quality crystal for blue laser diodes. One of the laureates is Dr. Shuji Nakamura. His contribution was development of the so-called two-flow method for mass production of GaN crystal for blue laser diodes. Figure 6 shows a problem which is encountered when a crystal film of GaN is fabricated on a sapphire substrate by metalorganic chemical vapor deposition (MOCVD) which is also known as metalorganic vapor-phase epitaxy (MOVPE). A reactant gas consisting of NH3 þH2 is provided with the substrate in a reactor to form a GaN film. A problem with this method is that the quality of the fabricated crystal is not sufficiently high. Amano et al. identified that reactant gas is undelivered to the substrate surface and introduced a long downward deliverytubewithaslanted substrate12. The tube successfully conveys the reactant gas, which also contains Trimethylaluminum (TMA) in their method, to the surface and the gas flow directly hits the substrate with an angle for efficient chemical 101