BPM issue

[Dj Real]

I have been using audacity to edit tracks. I export as an MP3. Most songs are okay. but every now and then, I come across a song that normally is 70 BPM and when I analyze it in Torq, it comes up higher. This instance is a KRS-One song in which Mixmeister says that it is 89.12. The site where I purchased the track says the same thing. I export it into Torq and it comes up 102 bpm. I have everything in Audacity set correctly. Is there something I need to do in Torq?

(Torq is my DVS program)

[kozikowski]

It's not unusual for software to get the BPM wrong. The ones that count bass notes die a flaming death if the song double-times in the middle of a phrase, has a back-beat that comes and goes, or, in the case of Trance/Dance, stops dead for six or eight seconds.

The only sure way is headphones and a stopwatch. You are much more likely to ignore those errors than the software is.

There is one more painful problem. If you have a crappy sound card, the speed of the playback may be slightly off. A number of people composing music on modest systems find they can't sing to themselves because the record and playback parts of their sound cards are so far off calibration. "I sing exactly in rhythm, but the composite is always off." Or: "The two parts of the harmony only match at the beginning, not at the end."

Koz

[Fip]

I have the same problem : recorded track of the output signal is slightly faster than the original track.
The simplest multi-track recording is therefor impossible.....for example id like to play the drums to a click track to practice my timing.... actually quite a funny thing that i´m now able to say: "I was in time, its the machines fault"

I have tried different setups:
- zoom h2 handy recorder as a usb-mic and line out, playback from a track in audacity (overdub), recording of the outputsignal of the headphones (with audacity).
- same but with the onboard soundcard and mic (Realtek HD Audio)
- same but output from onboard soundcard , recording with H2
- playing a song (tool- forty six and 2) on my mp3, recording the headphones with the H2- mic (as standalone recorder onto SC-Card), then importing original and recording into audacity
- etc.
...all the same....

how does that make any sense?
I understand, that every standalone device has to have a clock in it, otherwise there couldnt be a sample rate...
And I guess that the problem lies in the diffrence of the position of the samples...meaning that when lets say the H2 has a slower clock than my computer/soundcard/whatever, then the H2 records 44100 samples per what it thinks is a second and the computer puts the samples neatly into position, so that there are 44100 of them in what IT thinks is a second....thus making the recorded track faster..... ???????????????
so where does the time signal come from in the above set-ups, where the H2 is an interface, or only the Realtek is involved ? I cant believe, that the H2 as an audio interface would be THAT shitty, and the whole internet wouldn´t care exept for me... cause i didn´t find anything on that topic.. I would believe that the Realtek was "Killefit" though.
help...heeeeelp

[whomper]

I...
so where does the time signal come from in the above set-ups, where the H2 is an interface, or only the Realtek is involved ? I cant believe, that the H2 as an audio interface would be THAT shitty, and the whole internet wouldn´t care exept for me... cause i didn´t find anything on that topic.. I would believe that the Realtek was "Killefit" though.
help...heeeeelp

you can't ?
why not ??
its a portable device with a lot of features built to a cost point

ditto realtek - low end build to a cheap cost for budget pc item

but they should be closer than you seem to think you are getting

[kozikowski]

You can almost always track failures like this back to the $4.98 USD sound card.

There is a painful work-around. Play a click track into your headphones and record a new instrument to the click track. Three minutes later re-rack the whole thing to the beginning and perform the next instrument to the click track. Mute the first instrument playback.

Anything you record to the click track will be in perfect sync to each other not the click track. You can never play to your other instruments. The click track error is common to all the instruments and since sound cards rarely change over time, all the instruments will come out in perfect sync.

No, I probably couldn't do it this way, either.

Koz

[steve]

FWIW I've tested recording and playback through my H2 and the timing is extremely close - I've only recorded a couple of minutes, but the "tempo" seems to be spot on and the playback length is exact as accurate as I'm able to time. Are you playing back through the Realtek sound card or through the Zoom?

[Fip]

@koz:
yeah that would work i guess... i could record my Metronome alone, then play to the Metronome (not the recording) and then the both recordings should fit.... except if there´s a difference between recording with the H2 mics or its line in... well, I´m a bit lazy for that method anyway.

@steve:
I tried pretty much every possibility...just for fun i just tested recording the click track of audacity again:

disabled Realtek in Windows´System settings
prefence settings: Windows Direct Sound, Playback "H2 Wave Out", Record "H2 Wave In"
Headphones plugged into the Headphone Jack of the H2, Headphones held before the H2- mics

I recorded for 5 minutes, moved a (recorded) click to the exactly "second 2.00".
If everything was ok, now every click of both tracks should be perfectly in sync.
at second 8.5: recorded click is 6ms behind
14.5 : 12ms behind
20.0 : 16 ms behind
20.5 : 5 ms behind
21.00: recorded click comes 15ms too early....man

generally the recording is slower than the original

I ordered a M-Audio Fast Track Pro.... but Im afraid something will not work anyway... I guess it will depend on where the time signal comes from, or if there are diffrences in the rec and play - time signals.
does anyone know where audacity gets the time from ? should always be the current audio device right?
I read with USB 1.1, the computer derives the time from the USB frequency, so it depends on the USB drivers.
If thats the case, i´d need a USB 2.0 interface cause as they said, those have their own clock in it...
Im a bit confused right now.
I plan to use the H2 as a normal mic (using the direct monitoring from its headphone jack).But it would be nice if i could use the H2 as a standalone and later mix it in with other recordings, for the room-sound for example or as an overhead and thereby not having to occupy the two mic-ins.
I like the H2 a lot, and i still hope, that its all just my soundcards fault... man the H2 lets you set its metronome to tenth of a beats/minute (like 120.6 for example). Should that really all be a charade?

Thanks for your answers
Philipp

[kozikowski]

It's not the H2. Too many people are using those for production -- barring that yours may be broken.

The place of analog to digital conversion is the magic place where the sample frequency is established. If you connect the H2 USB to the computer, then the H2 determines the frequencies. If you send an analog signal to the built-in sound card in your computer, then the little clock oscillator in your sound card is doing the heavy lifting. As you noticed, if you try to compare multiple pathways, the characteristics change. Even if you try the same performance multiple times through a ratty sound card, you may get different show durations.

Another fuzzy rule is the computer speed. If the computer can't keep up, you get holes in the show and Audacity may crash. That will probably not cause show timing problems.

Audacity doesn't use any clock for signal generation. It makes up a bag of digital data and slaps tags on it with the expected sample rates and speeds. That happens in a very small fraction of a second. It's when you play the work back to your ears that the computer goes looking for a time standard and that's usually the sound card or external USB (or FireWire) device.

Try it. Open a fresh Audacity and generate 10 seconds of tone. It doesn't take ten seconds to make it, does it? It probably takes longer to draw the blue waves on the screen than to actually make the tone data.

So no, the computer itself isn't the shortcoming here although USB1 isn't helping. USB1 is very slow and sloppy and may cause instabilities later when you try to do serious multi-tracking and overdubbing.

But probably not show timing errors.

The real way to do this is use an external timing device or service to compare show times. I wonder about making a Music CD with the Audacity click track on it. The CD burning service doesn't go through the sound card. If you create a 10 second CD, it should take 10 seconds to play. Play the CD back in a separate CD player and use that sound as a standard.

Those of us with shortwave radios in the US can tune into the National Bureau of Standards time signals from Boulder, Colorado and use that. Canada has one, too. In Canadian French, thankyouverymuch.

Koz