請教各位
小弟有購買PCI 4461和sound & vibration ,想說先用no hardware方式,先產生20~20kHz掃頻來模擬頻率響應及產生1k Tone測試THD,SNR,IMD,Dynamic range,cross talk. 但遇到如下問題.
1.svxmpl相關範例的frequency response只能做一個聲道的(L or R),一個channel(Ai0拿來用stimulus)另一個則(Ai1拿來用response),那這樣4461的input都用光了,我有什麼方法可同時測兩個聲道的frequency response?
2.THD或SNR為什麼使用不同物件算出來會差這麼多,是我哪邊參數沒給好嗎?(如圖a,sg gen1).
3.為什麼frequency response 約在1k之後就會掉下來,而不是20~20kHz 維持在0dB?(如sound test1.vi)
[QUOTE=JerryLiao]
1.svxmpl相關範例的frequency response只能做一個聲道的(L or R),一個channel(Ai0拿來用stimulus)另一個則(Ai1拿來用response),那這樣4461的input都用光了,我有什麼方法可同時測兩個聲道的frequency response?
2.THD或SNR為什麼使用不同物件算出來會差這麼多,是我哪邊參數沒給好嗎?(如圖a,sg gen1).
3.為什麼frequency response 約在1k之後就會掉下來,而不是20~20kHz 維持在0dB?(如sound test1.vi)
[/QUOTE]
問題1：就你提供的硬體來看，因為輸入的通道數只有2個，如果要同時抓L與R的資料，
一定要有2個通道的資料作輸入才行，不知道你是否可以先量測L與R的數據
(使用別的硬體作Stinulus與記錄後)，再將Stinulus的資料丟至Labview裡面，
最後將L,R與Stinulus同時作處理
此方式的效果如果不行，我想還是建議你更新輸入通道硬體較佳吧
問題2：我使用軟體為 Labview 8.2 版，你的檔案我無法讀取外
某些vi的功能在8.2版也沒有(Tone Measurements,Noise Measurements)
因此沒有辦法給你建議，可能需要其他人的建議與參考
但是若比較一些內建vi，我試過Distortion Measurements 與 Analyzer vi的結果
是一致的，如下圖
問題3：你可以將vi轉成8.2版or 貼圖嗎？？
因為frequency response這個東西我不太了解，所以沒有試出來
可能需要其他人的建議了
I use LV version 7.1 so I can't open you VIs, sorry.
Years ago. I designed an acoustic test system with sound processor, which used LINE IN as inputs (DAQ) and SPEAKER as output(signal generater), all the test items contained: Power level, SNR, THD, FR, Cross talk, and so on. I think two inputs are enough.
For frequency response test, I didn't use chirp signal for source, but a serial of signal with different frequency and same duration(about 80 to 200ms), the frequenc points were chosen according to the curves from instrument like AP(audio precision) to make the final test curve smooth. Specially, the first frequency point should have a large amplitude at the output to affirm the start of the record signal.
In addition, when deal with each duration of the record signal, devide the each duration to 4 sections or more. because the record signal may not have a consistent phase which lead to curve falling at some points. discard the least FR result of the sections and use mean value of the other section's FR results.
Nnnn, for your refference :)