SOUND EDITOR - timing marks

[KP:]

Hi, all!

 

Perhaps I am just ignorant about the digital sound business these days. My use of Sound Editor is limited to recording analog audio to digital from my 4-track reel-to-reel system, then breaking up each tape into its individual cuts and storing them (as .wav files, right now) on my hard drive. When I'm done, I'll spend more time working on organizing them, and probably convert them to mp3 files for smaller space requirements. My hearing just isn't that good, anyway!

 

When I'm cleaning up the recorded raw tape file (also a .wav), I clean up the crackles and pops in the between-bands from the original phono LPs (Pre-rec tapes are absolutely awesome). I also standardize on exactly 2.25 seconds between cuts, and start each track in .25 seconds silence. Recently I realized that if you zoom in far enough, SE will break a second into pieces on the display. but the seconds are broken into thirds! it will display as 10:22:00, 10:22:25, 10:22:50, 10:23:00, 10:23:25, etc. If you drag a select cursor, it will show the length at the lower right of the screen, like 10:22:72, 10:22:73, 10:22:74, 10:23:00, 10:23:01, etc.

 

I don't understand this time format! Is it related to video timing or something? I'd rather see mm:ss.hh where hh = hundredths of a second. then 00:02.50 would be 2-1/2 seconds, not 2-2/3 seconds as it appears to be now. What's up?

[Brendon]

Hello KP,

 

I can't give definitive answers to all your questions because I haven't had enough practice at it. I've sent a message to THE Audio Guru, in the hope he can get time to drop in and help.

 

Regards,

Brendon

[d_deweywright]

Also, I just read the Wiki version of this information. It says, in part, "On the disc, the data are stored in sectors of 2352 bytes each, read at 75 sectors per second." (hence, 1/75th -second = 1 sector). I don't inderstand why the physical storage format on the CD has any impact on the person editing. Is the same "rate" implemented when playing music from a hard disk? Does the hard disk .wav or .mp3 file have the same framing issues? KP:

Heh heh... that indeed is interesting that Sound Editor displays the time in minutes:seconds:frames. I obviously haven't done enough with it to notice that, having used other editors much more.

 

KP, you're right, there's really no sense in imposing the frame characteristic on the editor, as you pointed out, there's no such thing as a "frame" or "block" until the audio data is actually put onto the CD, just bytes. Deleting a fraction of a "frame" just shifts the boundary of all the other "frames" as the bytes fill in. So, it is a bit of an oddity. The only place it gets really useful is where you have tracks that don't have any silence between them. Then you can look at the last block/frame and make sure that you're splitting the track on what will be a block boundary on the CD. If you don't, then the rest of that block is filled with zeroes (silence), which can cause an audible dropout between tracks. In your case, where you have distinct tracks that you're inserting 2.25 seconds of silence between, it's a non-issue. Many have complained of splitting tracks that aren't supposed to have any silence, and getting a short dropout between songs. It's surprising how audible a short bit of silence (1/75th second or even less) can be.

 

A quick example, two tracks are split and the first track ends 50 bytes into a 2352 byte "frame". The rest of that frame is filled with silence because a new track must start on a new frame. (That's how the directory structure on an Audio CD works.) So, you've got nearly a full frame of silence between the tracks. If the end of the first song was instead within a few bytes of the end of that frame, that bit of silence probably wouldn't be heard.

 

I hadn't realized that Sound Editor had given us the ability to see where the frames started/ended. Guess I'll have to pay more attention.

 

Audio really is just a byte stream. The frame characteristic is only really meaningful as far as actual storage on a CD, and then how it can affect transitions between tracks, again, on a CD.

[jeanrosenfeld]

Thanks to Brendon for explaining that the cause is the counting in frames. I did not know that.

 

However, I agree with d_deweywright that it makes no sense to do so in a sound editor. Other sound editors, such as Goldwave or Exact Audio Copy show their timeline in decimals well below 1 second (e.g. 1.00, 1.01 ...1.99, 2.00)

 

I have very often used Sound Editor to successfully remove a typical click signature (~ 5 to 10 msec) from an LP recorded track, with no bad effects on the sound (indeed improvement as the click is removed) either when played directly or after burning to Audio CD.

[KP:]

Heh heh... that indeed is interesting that Sound Editor displays the time in minutes:seconds:frames. I obviously haven't done enough with it to notice that, having used other editors much more.

 

KP, you're right, there's really no sense in imposing the frame characteristic on the editor, as you pointed out, there's no such thing as a "frame" or "block" until the audio data is actually put onto the CD, just bytes. Deleting a fraction of a "frame" just shifts the boundary of all the other "frames" as the bytes fill in. So, it is a bit of an oddity. The only place it gets really useful is where you have tracks that don't have any silence between them. Then you can look at the last block/frame and make sure that you're splitting the track on what will be a block boundary on the CD. If you don't, then the rest of that block is filled with zeroes (silence), which can cause an audible dropout between tracks. In your case, where you have distinct tracks that you're inserting 2.25 seconds of silence between, it's a non-issue. Many have complained of splitting tracks that aren't supposed to have any silence, and getting a short dropout between songs. It's surprising how audible a short bit of silence (1/75th second or even less) can be.

 

A quick example, two tracks are split and the first track ends 50 bytes into a 2352 byte "frame". The rest of that frame is filled with silence because a new track must start on a new frame. (That's how the directory structure on an Audio CD works.) So, you've got nearly a full frame of silence between the tracks. If the end of the first song was instead within a few bytes of the end of that frame, that bit of silence probably wouldn't be heard.

 

I hadn't realized that Sound Editor had given us the ability to see where the frames started/ended. Guess I'll have to pay more attention.

 

Audio really is just a byte stream. The frame characteristic is only really meaningful as far as actual storage on a CD, and then how it can affect transitions between tracks, again, on a CD.

 

What a GREAT resource this Community is! Where else but a sound studio or university could you have gone to find the people who KNOW what's up and get some serious information???

 

I have yet to read a justification for the 1/75 timing grid. Indeed, most agree that it's not appropriate for the editing task. Again, since I'm new here, how does one go about submitting a "change request" to Roxio that gets to the right people? Do they watch these threads? Do they respond to them? Is software modified based on some of the things discussed here? Is there a formal submission process that allows the originator to track the response, including status, schedule for review, response, and possible software updates?

 

I'm from a bit further up the time stream, and support organizations answered to the user community. That's not been my observation with the PC revolution. The support groups operate in isolation from their users, and the result is, well, just like this issue. Users recognize the need for change, but the people who can implement the change are not connected, don't want to be. Instead, the next release contains all kinds of whiz-bang new features that the marketing department hopes will attract new buyers. The old bugs are still there.

 

I pointed out that the "Undo" edit feature in SE (earlier versions) acts more like "Revert" in Adobe products, where ALL changes made are undone. Actually, it's not ALL, but enough to royally mess up the work you've done. "Undo" should keep track of each individual change made chronologically, and back up one step for each time Undo is invoked.

 

When one does a Selection and a Copy, the content of the copied block should exactly match what was visually selected. Instead, it appears to be a selection of the original clip. If you've Cut or Pasted into that clip, the clip length changes, but SE uses the original offsets into the audio stream to select, not one that matched the visual selection.

 

Maybe I'm asking too much of SE. I'm breaking a single captured digital audio stream into its constituent audio tracks, one-to-n albums, one-to-n tracks per album. I've learned how to accomplish this using SE, and it's still a WHOLE LOT EASIER than the same task would be in, say, Adobe Sound Booth. But it's not an intuitive, reliably predictable visual editor. Sure would be nice if the hundreds of dollars I've spent over the years for "Updated" versions would have some changes made that would make it easier to use.

 

Thanks again to all who have responded. I'm amazed each time I mess with one of these technological products and find how many have intimate knowledge not only of the product but the science behind it. Gotta LOVE IT!

 

KP:

 

Heh heh... [snipped out by KP:]

 

Dave:

 

It's clear you understand the part frames play in the audio CD. Again, when the audio editing person is working, are any of these issues important to him/her? I would think that the editor would ignore frame boundaries until burn time, and the CD burning process would align track markers during the burn process. Of course, I'm thinking silence precedes the track marker, and I know that's a myopic view. As an editor, though, even with spoken word, I'd probably put a track seperator BETWEEN words, not between syllables. Probably at a chapter break. Is this too "type-A"?

 

A second question: Do these 1/75th-second-size frames exist when the recording medium is NOT a CD? like a .wav file on my hard drive? When the player(s) read the file, are they finding an emulation of a CD? or are the frames non-existant outside the CD environment?

 

Thanks for your responses! KP:

[KP:]

The strange time line has been like that since, I think, EMC 8. I have complained, to no avail, at each new release.

 

See, for example,

 

http://forums.support.roxio.com/index.php?...hl=Sound+Editor

 

I also posted these pics, though the post in which they were seems to have disappeared:

 

Can't tell you how glad I am that this is not unique to my system, and that it makes no sense to anyone else. I just checked Wiki (surely the ULTIMATE AUTHORITY!) on the subject of time, and found no references to a division of a second into 75 parts. Maybe it was a joke from a disgruntled employee at Roxio?

 

Here's the TIME link, if you're interested. not a bad read.

 

I'm not sure how this user board is viewed by Roxio management and support. Do they monitor it? Do they respond. Have they ever taken action based on something reported here? If not, it's a help board, but not a feedback board to Roxio. That means a separate path to Support must be available, somewhere, yes? What's the history of Roxio Support response to user problems?

 

KP:

[d_deweywright]

What a GREAT resource this Community is! Where else but a sound studio or university could you have gone to find the people who KNOW what's up and get some serious information???

 

I have yet to read a justification for the 1/75 timing grid. Indeed, most agree that it's not appropriate for the editing task. Again, since I'm new here, how does one go about submitting a "change request" to Roxio that gets to the right people? Do they watch these threads? Do they respond to them? Is software modified based on some of the things discussed here? Is there a formal submission process that allows the originator to track the response, including status, schedule for review, response, and possible software updates?

Dave:

 

It's clear you understand the part frames play in the audio CD. Again, when the audio editing person is working, are any of these issues important to him/her? I would think that the editor would ignore frame boundaries until burn time, and the CD burning process would align track markers during the burn process. Of course, I'm thinking silence precedes the track marker, and I know that's a myopic view. As an editor, though, even with spoken word, I'd probably put a track seperator BETWEEN words, not between syllables. Probably at a chapter break. Is this too "type-A"?

 

A second question: Do these 1/75th-second-size frames exist when the recording medium is NOT a CD? like a .wav file on my hard drive? When the player(s) read the file, are they finding an emulation of a CD? or are the frames non-existant outside the CD environment?

 

Thanks for your responses! KP:

Why 1/75th second? Well, it appears to me that the idea of the Data CD was in the works at the time Audio CDs originally came out (several years before CD-ROM was introduced). So the idea of the frame on the CD was to hold 2048 (2K) bytes of data, with 304 bytes of Error Correction Code (ECC) for a Data CD. For the sampling rates chosen for CD Audio (44,100 samples per second, 16 bits per sample, 2 channels) that works out to an even 2352 bytes per 1/75th second (13.333mS). (Audio CDs don't have ECC on them.) So, I don't really think that Sound Editor is doing anything other than showing you where the frame boundaries will be when the particular bit of audio your editing is put onto CD. You're not being restricted to anything on those frame markers. The only really useful thing about that is as I said previously, to see how the last frame is filled up, otherwise, it can be considered just a random unit of time, unimportant to the editor.

 

I'll have to try it since I'm not at my home PC, but as Jean said, you can work with smaller than 1/75th second bits of your audio file in Sound Editor. The frame markers are merely displayed to show where you are relative to the beginning of the track, and really do nothing other than that. They're just indicators and the data is aligned "to itself," not the frame markers. As in, you clip out a 5mS noise click, and the data is shifted up to fill in, the frame markers stay where they are.

 

The whole "frame" thing only exists once you write an Audio CD. Data isn't stored that way on HD, it isn't played that way from your HD, it's all just a stream of data. As for an emulation of a CD? I'm not sure since I never played with an emulated Audio CD. (Data, yes, Audio, no.) I know Roxio/Sonic has their own proprietary format for an Audio CD "image", and that's because there is no .ISO format for an Audio CD. So, I've never played with it.

 

Does that help?

[Jim_Hardin]

Can't tell you how glad I am that this is not unique to my system, and that it makes no sense to anyone else. I just checked Wiki (surely the ULTIMATE AUTHORITY!) on the subject of time, and found no references to a division of a second into 75 parts. Maybe it was a joke from a disgruntled employee at Roxio?

 

Here's the TIME link, if you're interested. not a bad read.

 

I'm not sure how this user board is viewed by Roxio management and support. Do they monitor it? Do they respond. Have they ever taken action based on something reported here? If not, it's a help board, but not a feedback board to Roxio. That means a separate path to Support must be available, somewhere, yes? What's the history of Roxio Support response to user problems?

 

KP:

It has been in there like that for 5 versions – 4+ years.

 

The boards get maybe 1 question about it every 2 years…

 

Does that sound like anything that would warrant moving a major re-write of the graphics to the front of the line???

[ggrussell]

We've asked the same question many times and so far no response.

[jeanrosenfeld]

Does that sound like anything that would warrant moving a major re-write of the graphics to the front of the line???

 

How do you know it would require a major re-write? To me, it seems it's just a minor thing, resetting some counter somwhere in the code so it counts intervals from 00 to 99 instead of from 00 to 75 or whatever it does now.

 

And yes, I do think they could and should have fixed it in the past several years and versions or so that it's been reported.

[jeanrosenfeld]

The strange time line has been like that since, I think, EMC 8. I have complained, to no avail, at each new release.

 

See, for example,

 

http://forums.support.roxio.com/index.php?...hl=Sound+Editor

 

I also posted these pics, though the post in which they were seems to have disappeared:

[Brendon]

Have you folks not heard of FRAMES PER SECOND in the "Red Book" CD audio standard?

 

Data on a CD-DA is organized into sectors (the smallest possible separately addressable block) of information. The audio information is stored in frames of 1/75 second length. 44,100 16-bit samples per second are stored, and there are two channels (left and right). This gives a sector size of 2,352 bytes per frame, which is the total size of a physical block on a CD.

 

When editing a track, the boundary of the track should be on the frame [sector] boundary or else you will get sectors which are only partly filled with data, and you'll get pops and crackles. That is why Sound Editor counts down to the FRAME level.

 

Dave D-W has mentioned this many times. Who wasn't listening??

[Jim_Hardin]

Dave D-W has mentioned this many times. Who wasn't listening??

It would appear no one

 

Jean: after you sift through a 1000 lines or more of C++ and all that graphics that go into a waveform view with graphs, let me know

 

Edit: (that 'sounds' a little snippy... I sure did not mean to write it that way - sorry Jean)

[KP:]

Why 1/75th second? ...

Does that help?

 

Your explanation helped me understand where th 1/75th timing marks came from, but it does not explain their presence in the user interface for an audio stream editor. Who works in 1/75th-seconds? If I want 2.5 seconds silence, I must insert 00:02:375 (which, by the way, is not available on the interface, only 00:02:37 and 00:02:38). Plus, the tick marks and counter at the lower right can easily mislead one into thinking 00:02:25 is 2-1/4 seconds, when in fact, it's 2-1/3 seconds. 00:02:50 is not 2-1/2 seconds, it's 2-2/3.

 

Is there a path for change request submission that might actually get attention at Roxio? Just because it's been that way for a number of releases does not mean it must stay that way.

 

Thanks for the information, it really opened my eyes to some of the CD technology that I was ignorant on.

 

KP:

[d_deweywright]

Your explanation helped me understand where th 1/75th timing marks came from, but it does not explain their presence in the user interface for an audio stream editor. Who works in 1/75th-seconds? If I want 2.5 seconds silence, I must insert 00:02:375 (which, by the way, is not available on the interface, only 00:02:37 and 00:02:38). Plus, the tick marks and counter at the lower right can easily mislead one into thinking 00:02:25 is 2-1/4 seconds, when in fact, it's 2-1/3 seconds. 00:02:50 is not 2-1/2 seconds, it's 2-2/3.

 

Is there a path for change request submission that might actually get attention at Roxio? Just because it's been that way for a number of releases does not mean it must stay that way.

 

Thanks for the information, it really opened my eyes to some of the CD technology that I was ignorant on.

 

KP:

You're right, there's no real sense in the seconds:frames time marking in Sound Editor except to look at the end of the track. I don't think a "suggestions" section has been opened up for the "next" version of Creator, but we'll see if we can pass it along. Actually, under the "View" menu, there's an option for "Time Format". I would expect that to change the internal markings that are displayed, instead it just changes how the time is displayed in the title bar. So, in my mind, I'd have that option change it in both places so you could have it displayed as you wanted.

 

Then again, this is the first second time anyone's brought it up. (Sorry Jean.)

[KP:]

You're right, there's [snipped!] ...brought it up.

 

I found several other user interface discrepancies and outright bugs, some mentioned above. Can you see that they are passed along, as well? If there's no significant improvement scheduled for the next upgrade, I can't see any point in purchasing it. I don't use much of the video tools, just Sound Editor. Thanks for all your help! KP:

[Brendon]

Hi KP,

 

Anything they might decide to change [and this is the only complaint on this topic made so far] is very unlikely to make it through by the next version. Even the critical changes do take some time to make.

[KP:]

Have you folks not heard of FRAMES PER SECOND in the "Red Book" CD audio standard?

 

Data on a CD-DA is organized into sectors (the smallest possible separately addressable block) of information. The audio information is stored in frames of 1/75 second length. 44,100 16-bit samples per second are stored, and there are two channels (left and right). This gives a sector size of 2,352 bytes per frame, which is the total size of a physical block on a CD.

 

When editing a track, the boundary of the track should be on the frame [sector] boundary or else you will get sectors which are only partly filled with data, and you'll get pops and crackles. That is why Sound Editor counts down to the FRAME level.

 

Dave D-W has mentioned this many times. Who wasn't listening??

 

Brendon: WOW! I did NOT know that. It raises more questions, tho. If I highlight a block inside a 1/75th-second frame, and delete it (like a loud "click" in the recording), what happens to the data stream? While in the editor, I would expect a compression of the data and re-alignment of the timing marks. So why does the person editing see those intervals? Is the 1/75th-second interval important to the person doing an audio-stream edit, or is it just the physical characteristics for the recording medium. I would think the stream would be broken into its 1/75th-second chunks only when the data was getting recorded onto a CD. When I'm editing, I'm working in seconds. Does this 1/75th-second interval interleave with video formats the same way? Do they play ANY part in the editing process, and how should I, as an editor, deal with them?

 

This reminds me of the old dot-matrix printer standards of 80 and 132 character line lengths, which were set from the early 1950's when punched cards (80 column) and chain printers (132 characters) were the standard, and it just stuck. No good reason, it just stuck, still does! This is a bit different, tho, in that the editor is displaying a device-dependent data boundary that is probably of no importance to the actual audio stream edit process, unless one is dealing with the medium directly.

 

Is this in any way analogous to the 'frame' of video and/or movie? Off the cuff, I'd expect audio to be less frame-oriented, but is there such a thing as an audio frame, or is it strictly a bit stream? I believe that digital video is in frames, and if you messed with cutting up the data in an individual frame, it would probably mess up a portion of the video playback. Is digital audio the same way?

 

Sorry for all the questions, but I'd really like to understand HOW this all works, instead of just editing and listening. I wrote machine-language and Assembler computer code decades ago, and it gave me great insight into the working of the computer. I taught classes in analog sound reproduction at my university (also decades ago) to people who had no idea how a phonograph or tape machine worked. I'm just not fully "up" with this new digital technology.

 

I know how a light switch works, tho!

 

KP:

 

Have you folks not heard of FRAMES PER SECOND in the "Red Book" CD audio standard?

 

Data on a CD-DA is organized into sectors (the smallest possible separately addressable block) of information. The audio information is stored in frames of 1/75 second length. 44,100 16-bit samples per second are stored, and there are two channels (left and right). This gives a sector size of 2,352 bytes per frame, which is the total size of a physical block on a CD.

 

When editing a track, the boundary of the track should be on the frame [sector] boundary or else you will get sectors which are only partly filled with data, and you'll get pops and crackles. That is why Sound Editor counts down to the FRAME level.

 

Dave D-W has mentioned this many times. Who wasn't listening??

 

Also, I just read the Wiki version of this information. It says, in part, "On the disc, the data are stored in sectors of 2352 bytes each, read at 75 sectors per second." (hence, 1/75th -second = 1 sector). I don't inderstand why the physical storage format on the CD has any impact on the person editing. Is the same "rate" implemented when playing music from a hard disk? Does the hard disk .wav or .mp3 file have the same framing issues? KP: