![]() I would like to try this out on my own head unit, I just don't have the necessary gear. Did all this actually happen? I haven't a clue. The head unit really IS delaying the sound." So he concluded, "no phase shifts happening here. He went all the way to a 5-millisecond difference then he called it a day. He expected to see pulses on channel-1 perfectly overlapping pulses on channel-2, and this is exactly what the scope showed. Just to check, he delayed the RIGHT-front channel by 0.2-milliseconds and played all 21 tracks. He saw that all pulses coming from the left-front channel were now delayed by 0.2-millisecond again regardless of frequency. Then he increased the delay to 0.2-milliseconds. What he saw was that all pulses coming from the left-front channel were delayed by 0.1-millisecond compared to the right-front channel regardless of frequency. True enough, pulses from the left-front channel began exactly 0.1-millisecond after the pulses from the right-front channel. Then he played track-1 and carefully studied the scope. He used the head unit's time alignment feature to delay the left-front channel by one step. ![]() Same result on the scope, pulses seen by channel-1 perfectly overlapped pulses seen by channel-2. Then the next track and the next track until all 21 tracks were played. Looking at his scope, pulses seen by channel-1 perfectly overlapped pulses seen by channel-2. Volume set to "22" gave him roughly a 1-volt RMS output. Then he played the CD starting with track-1, the 100Hz pulses. He then hooked up channel-1 of his scope to the left-front RCA output and channel-2 to the right-front RCA output. Nothing fancy, just a portable Fluke unit. EQ set to "flat." Crossover on all channels set to "all-pass." Balance set to center, fader set to the middle. He set time alignment on all channels to zero. To ensure a good reading, he put his head unit on a test-bench, powered by a fully-charged car battery. Then the WAV files were all burned to a CD. Twenty-one sequential WAV files were created, starting at 100Hz going to 10kHz in one-third octave steps. The pulses repeated every half-second for 60 seconds. On his computer, he created WAV files containing quarter-second pulses from sine-wave test tones recorded at -10dBFS. I heard of this guy who wanted to SEE the actual time-alignment delay from the RCA outputs of his head unit, an old Alpine CDA-9815. Turns out, time alignment is really basically a frequency dependent phase shift. For a long time I thought that electronic time alignment was simply a buffered delay of the entire channel's audio. I think that helps define what I have understood to be true for some time but couldn't grasp why. Nevermind what happens when the recording is stereo and what happens in production before it even gets to your car. I think the one main concept that doesn't translate between the guitar mic concept and this is the fact that the mics are full range devices with no active or passive processing between them and the mixer. Not everyone is using the same system configuration, automobile, or DSP. As the time alignment is really frequency dependent, you may be short-changing the drivers alignment at the opposite ends of their frequency spectrums.ĭon't get me wrong, I like the concept behind this, I just think this may be why the some are reporting hits and some are reporting misses with this technique. In the end,assuming #1 doesn't happen in your case, your alignment will really only be great for the frequencies in the overlap. The phase alignment you created by performing this alignment technique may be erased when you change your filter settings back to where they should be.Ģ. If you change crossover settings to create some overlap, you are creating 2 problems:ġ. Each different slope and crossover point will have a different effect on the phase and response of the drivers in that channel. For one, my understanding is that most electronic filters induce a phase shift just like passive crossovers do. That's why I think it's a bit more complicated than the technique suggested here. ![]() ![]() Click to expand.Good article posted there. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |