• Get involved.
    We want your input!
    Apply for Membership and join the conversations about everything related to broadcasting.

    After we receive your registration, a moderator will review it. After your registration is approved, you will be permitted to post.
    If you use a disposable or false email address, your registration will be rejected.

    After your membership is approved, please take a minute to tell us a little bit about yourself.
    https://www.radiodiscussions.com/forums/introduce-yourself.1088/

    Thanks in advance and have fun!
    RadioDiscussions Administrators

Playback speed arbitrary?

T

Tom Wells

Two laptops, A and B.... A is an HP 8230 (this one), B is an HP 8430 ( running Zara and Breakaway Broadcast on the air).

Finally I put the music files on this computer in the same file structure as the other machine, so playlists can be
shared or transferred.

Just as experiment, I saved the running Zara playlist of what's on the air currently.
Dropped it in on this machine, and started the file so as run the audio in synch with what's playing on the air.
I was able to pause-start and perfectly synch up after a few brief pauses.

This computer soon lags in time, losing maybe a second in five minutes.

A lot more difference than I expected.
Made a lot more sense when there were machined surfaces in tape and phono drives.
Plus/minus errors and slippage, why is it that computers seem apparently no more "speed correct" than any consumer grade
tape recorder or phono of yore?
Where's the dang speed calibration pot on this machine?
Is it really the 60 hz time base? Both machines on same 60 hz.
Is it the single core vs dual core laptops?

Where Does the time go when it gets gained or lost? :-X
 
Is one of those an AMD processor by chance? One of our Adobe Audition machines exhibits similar behavior, and it happens to be the only one with an AMD processor.
 
Tom Wells said:
Two laptops, A and B.... A is an HP 8230 (this one), B is an HP 8430 ( running Zara and Breakaway Broadcast on the air).

Finally I put the music files on this computer in the same file structure as the other machine, so playlists can be
shared or transferred.

Just  as  experiment, I saved the running Zara playlist of what's on the air currently.
Dropped it in on this machine, and started the file so as run the audio in synch with what's playing on the air.
I was able to pause-start and perfectly synch up after a few brief pauses.

This computer soon lags in time, losing maybe a second in five minutes.

Sync up a 10 minute song on each with no segues and see if they stay in sync.

A lot more difference than I expected.
Made a lot more sense when there were machined surfaces in tape and phono drives.
Plus/minus errors and slippage, why is it that computers seem apparently no more "speed correct" than any consumer grade
tape recorder or phono of yore?
Where's the dang speed calibration pot on this machine?
Is it really the 60 hz time base?  Both machines on same 60 hz.
Is it the single core vs dual core laptops?   

Where Does  the time go when it gets gained or lost?   :-X
Click to expand...

It could be a threading issue.  The dual core may process the segues more correctly, as it doesn't get bogged down as easily.  It only takes 4 segues with 250ms of latency to get you to an entire second of difference.

Sync up a 10 minute song and see if they are still on time by the end. The segues are where I can see a difference and this test would take that out of the loop.
 
Oh, it's well detectable within one minute of play within a single song.
It quickly becomes a phase shift, then a real time difference that grows steadily.
Within a minute or so it's double-slap audio affect, etc.

I'll have to check the 8430 in the morning, this 8230 is an Intel Pentium at 1.73 Ghz.
It's not up to running both Zara and Breakaway, but the other with a dual core loafs at 40% in mono running both.
I seem to think the 8430 was Intel based, hmmm.

Now comes the real question. Which one is "more right"?
Maybe I'll just have to get another 8430 and put this one to pasture.
 
The exact answer probably has more to do with codecs and sound card clocks... if the file format is a variable bitrate file of any type... doubly so...

If the two machines have the exact same codec (brand/author/version) then theoretically they should play the same. A VBR MP3 file, for example WILL play differently on a Frauhofer codec, than say, a LAME codec, if it was encoded using the Fraunhoufer codec. That is where the funky phase watery phase distortions come from when decoding an MP3... a highest bit rate, CBR file will probably do the best. I've been writing automation software and for years have fought with this problem. Different sound cards, on-board clocks, jitter on the cards, AES locking, etc, all have a part to play. Very rarely have I ever had the OS or motherboard hardware be a real factor, as these run at thousands of times the re-fill rate of the data buffers that hold the data being ingested by the sound card. The data path is: Hard drive (pure data) > OS (pure data) > ring buffer > codec (PCM approximation of the original sound) > ring buffer (PCM) > sound card. the final ingest control is ultimately handled by the sound card and it's clock mechanism. Best method of eliminating the middle man is to play and record PCM wave files... even then, differences in sound hardware will make a huge difference. That's why pro sound interfaces for real recording studio work usually run from a master AES clock. Most consumer and broadcast grade sound cards run from a colorburst crystal, or an RC oscillator and are largely unlocked clocks (not synchronized to anything else in the world) and no precision by any stretch of the imagination. Your PC's master clock can vary several MHz. Realistically, crystals can only be manufactured with an operating range up to about 50MHZ. Meeting the requirement for a frequency above this is created by a frequency multiplier circuit. They come in 3 order (3X) and 2nd order (2X) varieties, without the use of inductive circuits. Consider a crystal oscillator for your CPU needs to be (hypothetically) 100MHz and it comes from a 33.33MHz crystal. A frequency tripler (3x) is used to synthesize 100 (or there about) MHz.... if the power supply that runs that oscillator varies by a few millivolts, the oscillator will vary about 0.5% of 33.33MHZ, or about 15,000 Hz... multiply this figure by 3 and it becomes 45Khz or more of error. The same thing applies to crystal oscillators and RC oscillators that run a typical DSP chip/subsection of a sound card.

Modern sound cards are actually sloppier than the pitch control on your old turntable in many ways. If two different codec are involved, one encoding one way and another decoding using a different algorithm, anything is realistically possible as these create thousands of iterations of error per second as perceived useless data is glossed over and removed from the data path.

Yes, I went to Valpo Tech too... I've heard of you Tom, but never met you, that I know of...
J.
 
Got another HP 8430 to replace the 8230. Much better on time coordination during elements, but
way different on segues/fades. Oh well :D.
 
Status
This thread has been closed due to inactivity. You can create a new thread to discuss this topic.


Back
Top Bottom