What kind of clips are these e.g. DV, HDV or whatever.
Vegas by default applies tiny fades to all audio at cuts to avoid this problem. You can disable this in the Preferences, maybe you've disabled this which could explain your problem except:
When you say "roll in/out" do you mean fade in/out?

Bob.

Sorry - one always seems to think one has stated all relevant material, only to find some things missing, like "HDV"...;-(
Since one clip has had its gain reduced to "0" ("- infinity", in timeline "speak") and the adjacent clip has a fade in/out also to "0", one would think the issue of DC offset would be nonexistent(?). One cannot totally disable the auto clip end fades in Preferences, but I have minimized the value to avoid a sudden audible audio level drop to zero with two clips butted together (no other editing software has either offered the auto fade length option or has had problems like the one I described earlier). This problem is odd, but with some rendering and additional cutting (a pain), or some added timeline audio level editing, I can generally get around it - but I think it should not exist as a problem...

If you cut the clip at an amplitude point anywhere above or below zero baseline (thus the term "zero crossing"), there will be a sound or "click." All that little fades can do is minimize the perception. That explains the physics of your question.

If you are unable to make your audio cuts exactly at zero crossing points, your other option is to crossfade the audio by dragging the end of one (ungrouped) audio event over the end of the other. This blends the fades into each other and will eliminate your problem. Start with 50ms crossfades and work your way up until you are happy.

"One cannot totally disable the auto clip end fades in Preferences, but I have minimized the value to avoid a sudden audible audio level drop to zero with two clips butted together"

You can easily disable the audio auto-fades in Options, but doing so or minimizing the value is exactly what is causing your clicks. Again, I don't know how your "other" programs handle audio edges, but the way Vegas does it is not a "problem" because you can change the default behavior. Since the autofade is not to your liking, using crossfades is probably going to be your solution. After you have decided on a crossfade length and type you like, learn how to use the numeric "/" shortcut.

Agree with all of this. I've yet to find any audio with a DC offset after around 1,000 hours of it, some of it of very dubious origin.
The default way Vegas applies a small fade simply works. Cutting at zero crossing is no panacea either, you need to be fairly careful what you do or you can still get a glitch. I have at times noticed the default way Vegas does things can cause a just audible hiccup in the audio during timeline playback however the rendered output is clean.
All that said of course one man's minor hiccup can be anothers monsterous thump so we shouldn't be too dismissive as there could be something seriously wrong. I'd hate to send anyone off thinking we find that acceptable. If what's above said doesn't heal the problem I think we need the OP to very specifically relate how serious this glitch he's hearing is.

Bob.

"Since one clip has had its gain reduced to "0" ("- infinity", in timeline "speak") and the adjacent clip has a fade in/out also to "0"

Well... if you're getting a "tick" then SOMETHING is not at 0db at the point of the cut. A tick is a voltage being applied to the speakers (or one being abruptly taken away) And yes... it is possible to have a DC bias when all has been reduced to 0

Granted a DC bias is rare these days and is usually the result of failing recording equipment.... something to check anyway.

"Well... if you're getting a "tick" then SOMETHING is not at 0db at the point of the cut."

Not quite true. If the adjoining fade is short enough you'll still get quite a "tick". This seems to be what the OP has done, reduced Vegas's default fade length.

"And yes... it is possible to have a DC bias when all has been reduced to 0"

I can't speak for anything else but if Vegas is doing this then it'd have to be seriously broken. One of the audio gurus would have picked this up by now surely.
Note that realworld audio is very likely to have a DC bias, check content in SF and you'll find almost nothing averages to 0 over time. Speech very commonly has a DC bias or to put it another way is asymmetric.

Bob.

"Not quite true. If the adjoining fade is short enough you'll still get quite a "tick". This seems to be what the OP has done, reduced Vegas's default fade length."

Regardless to how long (or short) the fade is Bob, a voltage either ac or dc MUST be present to cause a ticking sound

As for a dc bias it's quite possible to have a 0 volt ac wave (in other words NO ac signal... which is what sound is) riding on a 10 volt dc line. In SF or any other editor what you usually see is a sound wave going up/down over the 0 crossing mark but if there is a dc bias then what you see is a sound wave riding up/down using that dc voltage as the 0 crossing mark.

Cheap single supply amplifiers work on this concept. Instead of having a positive DC supply to handle the positive AC sound cycles and a negative DC supply to handle the negative cycles of the sound wave, they will use a DC supply to raise that 0 crossing mark so to speak to say... 5 volts (roughly half the total DC supply), so now you have the positive AC cycles functioning between 5 and 10 volts and the negative AC cycles functioning between 0 and 5 volts. In these amplifiers they will have a filter capacitor at the output blocking the DC voltage and allowing the amplified AC voltage to pass. When the filter capacitors get weak (which they do), the DC starts to pass. Now if this is the case then you can adjust the sound level to 0 and STILL have a voltage being applied to the speakers. You won't hear the DC voltage either... but you'll sure hear the click if you remove/cut it. It's easy enough to verify... take a double A battery or similar and connect it to a speaker. You hear a click, see the cone get pushed or pulled (depending on the polarity in which you connect). The cone will stay pushed/pulled until you remove the battery.... at which point you will hear a "tick".

Now I haven't a CLUE if SCS has worked all of this into Vegas or not... i would be surprised as you say if they didn't.... but then if you look in the AC3 pro encoder... you'll find a switch INCLUDED for filtering DC out.

"Regardless to how long (or short) the fade is Bob, a voltage either ac or dc MUST be present to cause a ticking sound"

We're talking about digital audio and at the sample level. There's no voltage as such at all. Also the sample rate is 48KHz, well beyond what we can hear. The sample rate is so high due to the Nyquist limit and gives a theoretical limit for DV of 24KHz.

Sound is produced by the changes is sound pressure, in turn that's represented by changes in voltage in the analog realm and changes in sample values in the digital world. So if you have a string of sample values that represent no voltage and then some that represent a voltage you have a "sound". You need to use Fourier analysis to know how it'll sound however any sudden changes will produce a lot of harmonics and sound most likely like a 'tick' if they're only a few samples long. This is why when you zoom in a LONG way at a cut you'll see Vegas by default applies a fade That's to attempt to avoid a sudden change in sample values which would become a fast dV/dT in the analog world and hence a tick or even a thump.
Now the trick that Vegas uses is good however it can fail if you put a cut in the middle of loud sound. Even though at the exact sample level of the cut there's no "voltage" it is only for a few microseconds. It'll be adjoining sample values that are not 0 and hence you'll create a new unwanted sound.


I don't know much about the tricks the ac3 enocder uses however I'd imagine given how it makes a considerable effort to optimise the bit budget I'd say the filter for cutting DC out is because of exactly what I said before. Many sounds are naturally asymmetric i.e. have a DC bias. Rather than encode that DC bias, by removing it less bits can be used by the encoder. This is probably the same as the trick used in AM transmitters known as asymmetic modulation. This is totally different to the DC bias problem is analogue amplifiers, don't confuse the two.


Bob.

Yes... it all makes perfect sense and I agree with you on everything bob.... BUT... in my original question I was not talking at sample levels.... maybe you misunderstood?


"Now the trick that Vegas uses is good however it can fail if you put a cut in the middle of loud sound. Even though at the exact sample level of the cut there's no "voltage" it is only for a few microseconds."

And this is exactly why in my original question (above as you will see) to the OP I asked: "Were the audio clips cut at the 0db crossing mark?"

"in my original question I was not talking at sample levels.... maybe you misunderstood?"

This is however a problem at or close to sample rate timeframes and the OP is manipulating his audio at around the sample level.

"One cannot totally disable the auto clip end fades in Preferences, but I have minimized the value to avoid a sudden audible audio level drop to zero with two clips butted together "


Even IF the cut is at 0 crossing you can and quite likely will get an unwanted 'new' sound. If you cut from 0 to a sound even at zero crossing you do create harmonics.

Keep in mind also that Vegas has no tools to assist cutting at zero crossing, for that you need SF unless you want to do a lot of eyeballing in Vegas.


Getting back to the OP's question though the full answer is fairly complex and I think he's probably worked out the answer for himself. You just cannot cut vision and expect audio to be OK, You might well find yourself using small J and L cuts, turning off Quantize To Frames or using asymmetic fades to get it sounding OK.There's considerable skill involved and entire books written on this subject.

On the other hand if this is HDV captured with scene detection it could be another matter entirely. I've NEVER captured HDV this way however I do know as the cameras go in/out of record you can get a bit of a glitch in the audio.

Bob.

"If you cut from 0 to a sound even at zero crossing you do create harmonics."

THAT bob I'll call you out on because it just isn't true. You can't get something from nothing. If you but 2 clips together both cut at the 0 crossing, there is no "harmonics"... or anything for that matter. Now what happens micro seconds before or after that joint is a different ball game.... but that's not what we're talking about.

"Now what happens micro seconds before or after that joint is a different ball game.... but that's not what we're talking about. "

Well I don't really know what you're talking about then and fail to see it's relevence to the OP's problem. What happens microseconds later is totally relevant to what we hear and to what the PPM meters in Vegas display. They do not display the value of every sample, they use a different and complex algorithm that can be fooled. If they didn't it would not be possible for them to display greater than 0dB FS and yet they do.

Bob.

yeah, blink,
it's all voodoo . . .

THAT bob I'll call you out on because it just isn't true. You can't get something from nothing. If you but 2 clips together both cut at the 0 crossing, there is no "harmonics"... or anything for that matter. Now what happens micro seconds before or after that joint is a different ball game.... but that's not what we're talking about.

I just dropped by and saw this post and still hope to find an answer to the original question (to the original poster: read my last paragraph below for another thing you should try).

However, in the debate between blink and Bob, I'm afraid that blink is not correct.

At the risk of sounding didactic, if you take a sine wave:



and then edit it in Vegas by cutting it at the zero crossings and eliminating the negative cycles, you create the equivalent of a rectified sine wave:



As you can see, this waveform (the solid line part -- forget the dotted lines) has the edits at the zero voltage (zero energy, zero sample) points, but if you do the Fourier analysis, you'll find that it contains infinite harmonics, with significant energy all the way up to the 20th harmonic.

The reason for this is that the zero voltage crossing really isn't the thing which controls the harmonics (and with them, the possibility of something which sounds like a pop). Instead, it is the discontinuity which causes the harmonics. Any waveform with a discontinuity or infinite slope (which the rectified sine wave has because it instantly goes from one direction to the other) will have very nasty sound.

However, I think I understand where blink is coming from because when you actually design an electronic circuit, like a triac light dimmer, you try to use something like the old 3059 zero voltage switch to turn the power on and off at the zero voltage crossing in order to avoid surge currents and the buzz that they can bring. If you look at the waveforms for triac dimmers that don't use zero voltage tricks (which most do not), you'll see HUGE discontinuities in the waveforms as they go almost instantly from zero to some huge voltage. This is why these darn things generate so much noise.

As to the original poster's problems, I would recommend changing the playback device. Go to Preferences -> Audio Device and change the Audio Device Type to something else. Before you make this change, make sure you have something that clicks every time. Then, make the change and do the playback again. If that doesn't work, keep trying until you've tried all the devices. Report back here and tell us if that works. I'll bet it will solve your problem.

Continuing on from John's post...



If it's cut at the foot of a slope on the in & out points you will retain a continuous sine wave. However, at the very least, this would be extrememly tedious to do if there are a lot of cuts with this issue in a large project. I think the Vegas approach works much better in practice (fade or cross-fade the cuts).

"and then edit it in Vegas by cutting it at the zero crossings and eliminating the negative cycles, you create the equivalent of a rectified sine wave:"

What???

Sorry... what you have shown John is what happens when you run a sign wave through a diode (without flipping the negative cycle of course)... which has absolutely NOTHING to do with cutting at the 0 crossing. Who said ANYTHING about eliminating negative cycles???

If you look at the sample Porpoise1954 has placed... you want to cut your wave at the 0 crossing marks (the hatch lines)... at 0 volts.... where there is no sound. If you cut anywhere else, and at worst case, at max (or min) voltage... in other words at the very top or bottom of the wave... THEN you stand a great chance of clicks and noise at the edit points.

And while we're at it... here's dc forward bias:



The wave is not riding at 0 volts DC... so even if you make your cuts at the 0 crossing mark (relative to the sine wave), you're STILL getting a DC voltage at the output. This is how single supply amplifiers work. If the dc current is not filtered out, it doesn't matter what you do or where you cut, you will get clicking and ticking sounds at all of your cuts.

That's exactly the same as what can happen if you make a cut at zero crossing, well apart from the roughly 0.7V forward bias in the diodes.

Try it in Sound Forge. Cut out half a cycle at zero crossing and join the ends and you get exactly a rectified sine wave.

What you need to do to avoid that is to match the slope. SF also has a way to help you do this.

Bob.

"What you need to do to avoid that is to match the slope. SF also has a way to help you do this."

That's a good and valid point. If you cut at the 0 crossing and butt 2 negative cycles together then bad things will happen. If at all possible you want cut at the 0 points and join a negative with a positive one so the flow stays natural. But then on the other hand Vegas's auto fades at the ends SHOULD be of long enough duration to blotz out any of these little errors.

"But then on the other hand Vegas's auto fades at the ends SHOULD be of long enough duration to blotz out any of these little errors. "

YES!
Exactly what we've been saying. Vegas has no mechanism to make a cut a zero crossing apart from eyeballing it and that's none too easy. The whole zero crossing idea that you introduced is irrelevant to audio editing in Vegas as it's default autofade will generally mask all the glitches. Like anything though there are times when it will fail. You (as I said before) might need to use a J or L cut to get it at the right point, like right up against the start of a word or a note. Note that choosing such a point is not zero crossing, you're just choosing a point that has very low signal amplitude.
What the OP has done is to reduce Vegas's autofade length to zero, hence he's dramatically increased the risk of getting a glitch.

Bob.

"What the OP has done is to reduce Vegas's autofade length to zero,"

And the op CAN do this.... as long as it is done correctly and there is no DC bias.... which is the point I'm trying to make.

Vegas" auto fade at the ends is a some what universal fix to a rather detailed issue. I don't know about SF because I use Audition but there is a switch that when enabled will cut at the closest available 0 crossing mark instead of EXACTLY where you want to cut. There is also a dc bias filter switch that will remove any dc bias that should one exist.

Vegas does not have these finite adjustments... but that doesn't mean it can't be done..... it just requires a lot of "eyeballing".

Even though you can do it in Vegas and you can easily do it in SF or Audition there's another problemo.
If I find a zero crossing point in clip A, cut there, find a matching zero crossing point in clip B and cut there and then slide them together it may sound OK except as I slipped clip B's audio I just took it out of sync with the vision! Worst case with say a 50Hz sound you could put it out of sync half a frame easily, do that a few times and depending on your luck it could all add up to a few frames, not good.

Bob.