Matching Sound Timings

[HaroldB]

Al,

Just to add some more fuel to the fire, according to Cool Edit 2000, the music for my Dempster show starts at 1.7 seconds, but the first slide doesn't start until 3.1 seconds. I changed it to a "cut", and same thing - the time is off by about 1.4 seconds.

That makes no sense to me (and I'm sure to you either).

Tell me more.

1. What version of PTE are you using? Try using 4.0 so that we don't get into this beta area.

2. What do you mean that the "music starts" at 1.7 seconds in Cool Edit? do you mean to say that there is simply silence before that?

3. What do you mean when you say that the "first slide doesn't start until 3.1 seconds"? I assume that you are talking about PTE... IMX, the first slide always starts with the very beginning of the music file. Where does the number 3.1 come from?

4. Check the total length of the music. In PTE, that's shown at the bottom of the synch window. In Cool Edit, it will be in the lower right hand corner of th window in the View/Length when you first open the file.

Harold

[alrobin]

Harold,

You're right. No sense at all. I used to attribute it to PTE just taking time to "warm-up".

In answer to your questions:

1. What version of PTE are you using? Try using 4.0 so that we don't get into this beta area.

- It does the same thing in both (see my reply to Igor in another thread).

2. What do you mean that the "music starts" at 1.7 seconds in Cool Edit? do you mean to say that there is simply silence before that?

- Yes, simply silence.

3. What do you mean when you say that the "first slide doesn't start until 3.1 seconds"? I assume that you are talking about PTE... IMX, the first slide always starts with the very beginning of the music file. Where does the number 3.1 come from?

- Yes, PTE. However, I should have said "the first slide we have control over on the timeline", which is slide #2. Sorry for the confusion. (I use a dark slide as #1, so wasn't counting it. ) By the way, it too is a "cut", so processing time is minimal. 3.100 sec is where I placed my first transition (not counting the one at the beginning of the show for slide #1).

4. Check the total length of the music. In PTE, that's shown at the bottom of the synch window. In Cool Edit, it will be in the lower right hand corner of th window in the View/Length when you first open the file.

- CoolEdit shows it as 10:43.395, and PTE 4.10, b8, at 10:44. The timeline itself is 10:44.6 in length.

Please let me know if you have any clues or suggestions.

[Alan Lyons]

Harold,

Do you have a long "display slide for" time on your dark slide. As this is only ment to be a marker to allow you to control your real slide 1, I usually set it to display for 1 sec or less. Using the traditonal numbering method for slide trays I number this slide 00.

Alan

[alrobin]

Hi, Alan,

I presume your last post was directed to me. I consider slide 1 to be just a "dummy" slide, as the time it appears could vary from system to system. So, to be safe, I always use a dark slide and give it a transition time of 20 ms (in other words, a "cut" or in PTE parlance, a "quick" transition).

However, in the old PTE, I found I had to leave slide 1 "on-screen" for at least 3 seconds to avoid early glitches. This is apparently not a problem in the new beta as much as it used to be.

[HaroldB]

Please let me know if you have any clues or suggestions.

Frankly, your post left me mystified. And the stuff you posted to Igor was even more bizarre.

Let's see if we can concentrate on exactly one problem: the difference between Cool Edit 2000 and PTE 4.0 using MP3 files. I use that combination for all my shows, and have never had the problems that you describe.

Let me first paraphrase what I think you are saying so that you can see if I have it right. You have an MP3 file that shows as duration 10:43.395 in Cool Edit, but duration 10:44.6 in PTE. Furthermore, the first 1.7 seconds of that file are silence in Cool Edit (and the music begins at that point), yet in PTE 4.0, the music begins at 3.1 seconds on the PTE timeline.

So here are my new questions:

1. This is a silly question, but are you 100% sure that you are using the same file in Cool Edit and PTE? I have always seen them show exactly the same total duration.

2. Is this a VBR file? If not, what bit rate is it? I use CBR, 128KB, 44100 Hz.

3. What created this file? If it was not Cool Edit, could you save the file in Cool Edit and see if that synchs up with PTE?

4. What "options" do you use in Cool Edit to save the MP3? I use CBR, 128kb, 44HZ, Stereo, Fast Codec, Allow mid-side joint stero, Allow intensity joint stereo, Write CRC checksums, and ISO padding. I discovered during my early testing that the ISO padding was particularly important to get the file to match to PTE. If you use different options, could you try it with those?

Remember that when you save an MP3 with Cool Edit, it adds .027 seconds of silence to the front of the file. So after you save it, close it in Cool Edit, and reopen it to look at it.

I hope this finds the source of the discrepency. FWIW, I am running Win/ME, but this certainly doesn't sound like an operating system type of problem to me. OTOH, who knows?

Harold

[alrobin]

Harold,

I am converting from .wav to .mp3 using the following options: CBR, 96 Kbps, 44.1 kHz, Stereo (14.7:1), fast codec, ISO Padding, and writing CRC. No "Joint Stereo", but will try that to see if it makes a difference (not sure what it does, exactly, but it seems that if you did use it, you would only need one, not both).

And, it is definitely the same file in both Cool Edit and PTE. I have repeated the process several times, with different files, and consistent results.

The sound files I am using were created from MIDI files, played back from an external sound module through my sound card into Cool Edit where they were mixed down, and saved as mp3 files.

I don't have the same problem in Win Me, either - just Win XP and Win 2000. I suspect there is some incompatibility between Cool Edit and these other OS's. Maybe when Adobe take over, they will make everything compatible???

All I know is that right now in order to use your method of marking music peaks I would have to test for the difference between Cool Edit and PTE first, and add in a fudge factor, or else go back to the original session and measure from there. (not that this would be that big a deal, but a bit of a nuisance). It's a mystery, though!

[alrobin]

Harold,

I have just run a quick test on the Win 2K system using 128 Kbps, with both "mid-side joint stero", and "intensity joint stereo", and lo and behold, I get the same addition of silence as you do, instead of a truncation as with 96 kbps. Go figure!

I will try a few more tests tomorrow on both Win 2K and on Win XP to see if they react the same way.

Thanks for your help!

[HaroldB]

I have just run a quick test on the Win 2K system using 128 Kbps, with both "mid-side joint stero", and "intensity joint stereo", and lo and behold, I get the same addition of silence as you do, instead of a truncation as with 96 kbps.    Go figure!

Al,

Keep in mind that my "addition" of .027 seconds of silence is really miniscule, and it's an actual ADDITION to the MP3 file. In contrast, your "truncation" seems to be something that is between PTE and Cool Edit, no?

FWIW, I have no idea what "joint stereo" might mean... those boxes came checked by default and I left them that way .

I remember spending more than a bit of time experimenting with Padding, thinking that my extra .027 seconds of silence might be some kind of padding I could get rid of. But in my experimentation, I discovered that you really NEED ISO padding on MP3 files .

It's interesting that you said you had no truncation effect with Win/ME. That obviously explains how you and I can both be sane .

I have done no sound programming on Windows, but I hope the fact that this is only a problem with Win/XP and Win/2000 (which share a common code base) might give Igor a Clue as to what might be happening here. How could Cool Edit possibly display a waveform differently depending on the OPERATING SYSTEM it is running under?

Harold The Puzzled

[HaroldB]

Al,

Another thought. I don't have the time to do this myself, but maybe a little reading on the Cool Edit forum might explain some of these mysteries.

Let me know if you find anything good!

Harold

[alrobin]

In contrast, your "truncation" seems to be something that is between PTE and Cool Edit, no?

Hi, Harold,

No, it seems to be completely within Cool Edit. When I load back the mp3 file into Cool Edit, it shows up truncated. That's what twigged me to it in the first place - I kept adding 3-seconds of silence (in Win XP), and part of it kept disappearing.

So the next time I made a change to the waveform in Cool Edit, I would add it back in, thinking that I needed it for PTE, when all along, PTE was seeing it correctly. I didn't know that it made no difference in PTE as I assumed that I needed to add the silence back on before using the modified file in PTE again.

It does seem to be a "Win NT"-based problem, though.

It certainly is strange - I will try to do some more experimenting today and let you know the outcome. Thanks again for all your suggestions and patience in helping me sort through this.

[alrobin]

Well, I did some more digging this AM, and at first found I couldn't reproduce the problem.

Then I realized that I had sampled the previous problem files at only 22.05 KHz, not 44.1. So, I re-saved my test ".wav" file (with test 30-ms tone bursts at the end of 3 seconds of silence, and at 1-second intervals after that) as an mp3 22.05 KHz file and loaded it back into Cool Edit.

Now, instead of truncation to 1.711 sec, the silence becomes truncated to 2.428 sec., for both Win XP and Win 2000. For 44.1 KHz, the silence is augmented by 0.027 sec. as it is supposed to be. It doesn't seem to matter whether a CBR of 96 or 128 is used, or whether "Mid-Side" and "Intensity" Joint Stereo are checked or not.

At least I know now what to watch out for when creating mp3's for use in PTE.

[HaroldB]

At least I know now what to watch out for when creating mp3's for use in PTE.

Al,

That's exactly how I felt when I finally figured out the .027 second addition. What I now routinely do is create my MP3 file in CoolEdit, close it, and then re-open it in Cool Edit. At that point, it is .027 seconds longer than it was before the save, but it matches PTE 4.0 exactly.

BTW, you typed 0.27 seconds in your last message. IMX, it is .027 seconds. Yes?

Harold

[alrobin]

Oops! Yes, I should have said .027 sec.!

[Igor]

Dear Al,

I just checked your dempster3_beta7.mp3 file which you've sent me a month ago.

PTE v4.01, v4.10 beta, WinAmp, Nero WaveEditor play it exactly similar.

This music file has a short (100 ms.) sound "click" at 1,7 second. Main sound appears at 3,1 second. All this waves well displayed in the graphical area in the Nero WaveEditor.

[alrobin]

Thanks, Igor,

Could you please email me this file as I may have deleted the version I sent you. I have one, but it does not seem to match the parameters you mentioned (e.g. the music doesn't start until 4.73 sec., and I don't see a "click" at 1.7 sec.). It's recorded as 16-bit, CBR 96 kbps, sample-rate 22050 Hz, by the way.

Meanwhile, I'll email you a better file for measuring timing. I'll also send you one recorded as a 44100 Hz mp3 file for comparison. If you have time, could you please tell me where the first tone appears in both Nero and both old and new PTE versions? The first tone will be at 3 sec in the wave file, but at slightly different times in the mp3 files due to ISO padding. However, they should both look the same in PTE. If not, please let me know. Thanks very much.

By the way, I've found that Cool Edit does truncate the silence in Win Me, after all, if you save a wave file to .mp3 as CBR 96 kbps, and waveform 22050 Hz. For example, 3 sec of silence at the beginning of a piece ends up as 1.775 sec.

[alrobin]

Igor, I have sent you some files I have been using lately in testing Cool Edit, and the different PTE versions.

For anyone else interested in making some tests of your own, or seeing the results of my own tests on PTE timing, and the effects of converting to MP3 in Cool Edit, I have included them as downloads on my Web Page.

The first zipped file includes two mp3 files from the same wave file, but one using 44100 Hz sampling and the other 22050 Hz sampling. The original wave file has short tone bursts at 3.000 sec, and then every 1.0 second afterward. However, the MP3 files show different times, as silence is added to the 44100 Hz file and deducted from the 22050 Hz file.

The original wave file is included in the second zipped file on the web page for anyone who wishes to explore things further. It is around 2.6 Mb in size, though.

Also included are two PTE files using each of the above-mentioned mp3 music files to test whether the apparent difference in timing affects the way they play back in PTE. If you unzip and load everything into the same folder, the PTE files should play OK, except that you may have to change the address for the music files.

I would appreciate any feedback you may have, and your thoughts on whether there is any difference in the synchronization results in PTE when previewed in different versions.

[Igor]

So this problem is also solved by beta #9

[alrobin]

Yes, Igor. Thanks very much! I left the files up in case someone wants to do some experimenting for themselves.