Splitting album sides into tracks for DVD-A/V burning

QuadraphonicQuad

Help Support QuadraphonicQuad:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Colin Dunn

Active Member
Joined
Mar 1, 2012
Messages
74
I'm trying to figure out everything I need to author a DVD-A/V disc. One question that's come up:

I take an entire album side and record it into my computer, then normalize, declick, decode, and save off mono .WAV files.

The result is 4 mono .WAVs that are the entire album side. Encoders for AC-3, DTS, and MLP merge those into a multichannel sound file, the length of the entire album side.

Is there a way to set track numbers on the DVD-A/V based on time positions within these long sound files? Or do I have to split the .WAV into one for each track before being able to author onto a DVD?
 
I have not dealt with this sort of thing for a while, but I think the idea is to go ahead and split it into seperate files, but then, with the DVD authoring software, set gapless tracks for those tracks which are supposed to transition without pause.
 
I guess creating individual wav files (for each track) as Q8 says is probably the easiest.
 
I create a multi-channel FLAC file with AudioMuxer (easier to keep the files together and saves space) and then use a CUE file to set the track positions. Another advantage of FLAC is that the CUE file can be embedded in the file so you don't have to keep track of a separate CUE file. Various utilities are available to do this. I use foobar2000: Load the FLAC file in foobar2000, right click on the file and select "Utilities" and then select "Edit cuesheet". Unfortunately, the Cirlinca software doesn't recognize the embedded CUE file, so you have to extract it when you load the FLAC file in Cirlinca. In Cirlinca, load the CUE file (that points to the FLAC file) and the tracks will be listed as defined in the CUE file. I have to do this when creating a disc to play in my car, but when playing at home I just use foobar2000; so the embedded CUE file works great.
 
I'm trying to figure out everything I need to author a DVD-A/V disc. One question that's come up:

I take an entire album side and record it into my computer, then normalize, declick, decode, and save off mono .WAV files.

The result is 4 mono .WAVs that are the entire album side. Encoders for AC-3, DTS, and MLP merge those into a multichannel sound file, the length of the entire album side.

Is there a way to set track numbers on the DVD-A/V based on time positions within these long sound files? Or do I have to split the .WAV into one for each track before being able to author onto a DVD?

This depends on what authoring software you are using, but in general you have no trouble (or should not, anyway) using long files. In all seriousness with Live albums and/or segued material, it's a must.
There are issues though, and the biggest ones will be with the lossy streams as you are extremely limited on the apps you can use - I know of exactly 4 that are capable of outputting a DVD-A legal Video_TS and they all have problems or downsides (depending on your point of view):
1 - Scenarist SD - discontinued, requires MPEG-2 video files to accompany segued tracks.
2 - Media Chance Labs DVD-Lab Pro (Abstraction Layer MUST be turned off at compile, which gives you a whole heap of headaches) - also discontinued, track marker placement restricted to the nearest second in slideshow modes or to the frame with MPEG-2 files, limited to no more than 6 or 7 VTS or it will not compile, no preview.
3 - Spruce Maestro - discontinued
4 - Sonic DVD-Creator - discontinued, Mac OS9 only, probably the best one but requires a very moody hardware decoder unit to get any preview functionality.

There are 2 ways to go about preparing the Video_TS - and the first thing I absolutely must point out to you is to forget all about normalizing anything, ever - it gains you nothing at all except potential problems. Normalizing to peak levels will make very little difference to the percieved volume & normalizing to RMS levels will simply raise the noise floor along with the audio. Both versions also add a layer of quantization distortion & needless processing errors for absolutely no gain whatsoever. Please, I beg you, do not do this - all normalization plugins should, to my mind, be withdrawn/uninstalled/deleted & or scrapped. You will almost certainly find you need a lot less declicking as well. All digital audio will sound a hell of a lot nicer if you monitor your levels to VU where -18dBFS = 0dB VU, and allow yourself to go to +3dB VU. As long as the VU meter is bouncing around the red line, you will be fine. Forget all about digital peak metering as it is hopelessly inadequate because it tells you nothing at all about the loudness levels and only what the hottest transient peak is.
If you stick to this, you will find that as you want more volume you can turn up the amplifier (which is, after all, what it is designed to do) a lot more than you will ever manage with peak normalized files that are up around 0dBFS.

For the Video_TS (getting back on track) you have 2 options available, depending on application & space constraints.
1 - Slideshow mode. This gives you a single image per track and is the most space efficient method.
2 - Low Bitrate MPEG-2 method. This still looks like a single image per track, but is actually a very low bitrate movie. Depending on the complexity & quality of the original you may get away with going as low as 1500kbps.

Each of these methods has their pros & cons.
Slideshows are the most efficient but at the same time the most problematic as you are pretty much restricted to DVD-Lab Pro these days - although I am currently experimenting to see if MuxMan can create a legal Video_TS folder. DLP has it's own issues. The biggest one is that in slideshow mode chapter placement is to the nearest second and with tight segues this is awkward as you have to weigh up what the best playback will be - do you set it to make sense played as an album, or so that you can start on any track & not clip the start and end up with a second or two of incorrect slide display during full playback, or set it to make perfect sense played as an album & be slighty off-beam if accessing a single track. With segues, you cannot have it both ways in slideshow mode. Same basic problem in DLP's Audio Title mode as well, but with the added problem of timing drift - the longer the sequence, the further out of whack it gets towards the end. Oh - almost forgot. In DLP Slideshow mode, you have to create the SS, sort out the images and manually add the second stream - and this is never remembered when you need to recompile or if you reload a saved project - it will need to be checked every time. The big advantage to MPEG-2 or "movie" mode is that you get frame accurate placement - IF the required point also coincides with a block point in both audio streams - with AC3 and DTS you need to have 2 blocks aligned for a chapter marker to "take".

Scenarist is very hard to use, but it does give you the most compliant streams that will work reliably on more systems.
MPEG-2 mode is still the easiest to use, as the Slideshow option will *only* work where every track is a separate item. This is because it adds slides at the VOBU level instead of the CELL level, which is where we need it to be for a chapter marker within a stream. Basically put, if you use a slideshow in Scenarist you cannot have chapter markers in the slideshow itself - every song must be separate, and this can create issues when authoring.

Audio_TS you have either discWelder Chrome or Sonic's DVD-Audio Creator. Nothing else supports MLP.
Both will let you add track points - Chrome in actual authoring - you import the files, add them to a track & then create gapless track points, DAC you add these when importing the Audio file by creating them in the IMPORT AUDIO dialogue.
Chrome is easy to use but badly flawed & very buggy - it's an abstraction layer tool & suffers as a result, which is why it was discontinued as the way it works made the problems insurmountable. DAC is seriously stable as it only works with XML files, and never the actual assets.

For slides, remember these will be encoded to MPEG-2 anyway during multiplexing, so start with the best quality images you can get - ideally bitmap or uncompressed TIFF files. NEVER use a JPEG file - especially in Chrome as this will look ugly, as you will see the comprssion artefacts clearly.

I hope this helps a little - please just ask if there is anything else you need to know - and it would help if we knew what apps you have access to as well.

I know this looks like a very complex thing, but once you ge it all squared in your head it's actually not that hard to do, and we are all here to help
 
Neil -

The normalization I did in this process was not during DVD authoring, but in Audition immediately after recording (but before de-clicking). I did think of the issue of raising the noise floor with the audio level, as well as a small quality loss as every sample gets changed. After de-clicking, the noise levels were not objectionable (though not a "black background" like a good SACD or DVD-A). I contemplated the idea of de-noising with DeNoiseLF, but found that it took out bass drums; other LF noise (hum/rumble) was negligible and not worth trying to remove.

I'll see if I can avoid the need for normalization by getting peaks to -6dB. Getting this exact is difficult given the varying levels / dynamics from one LP to the next. I'm leaving the 6dB buffer because of the Audition scripts, which allegedly need this much headroom to avoid digital clipping as they process the stereo .WAV into 4 channels. I also am not sure my phono stage can drive my sound interface all the way up to a digital 0dB level, even with the input gain maxed out. After processing in Audition, I don't do any more normalization.

I put my DVD-A/V questions in the other thread about developing an open quadraphonic decode process.
 
Back
Top