Workflow? Anything wrong?

[Swurveman]

Hello,

I know there’s another work flow topic. So, I apologize if I should have posted there. I didn’t want to get lost in the clutter.

I’m new to Wavelab and mastering and am wondering if my workflow is unusual, or if there is a better way

Here’s my workflow:

1. Audio Mixdown from Cubase at 96/32 float. Set up a file folder called “Source Files”. Open files in the Audio File work space

2. Make edits in the Audio File work space (crop file lengths to audible sections, do fade ins and outs etc) and apply Mastering Plugins to the Master Section of the Audio File work space.

3 Use Peak Master then Crystal Resampler plugins in final slots, then render without dithering a High Bit Resolution (HBR) file that is at 44.1/32 bit float. Name each song as “SongM01_HBR”. Put these in a file folder called “Mastering HBR Files”.

4 . Bring all the HBR files into the Audio Montage. Align them in order. Name them, do all crossfades between songs. Use the Wizard to determine time between songs.

5. Burn CD or DDP, using a 16 bit dithering plugin on the Master Section. Save the Montage.

6 With the HBR files intact, I use them as source files to make MP3’s in the Audio File work space.

I am under the impression that the Crystal Resampler is superior to resampling in Cubase’s Audio Mixdown. Is this true?

Thanks to anyone with experience/comments!

[geezer]

It does seem like a reasonable workflow, as long as you trust Crystal Resampler more than an D-A/A-D cycle. The double conversion cycles is what most high end mastering houses use, but I have not tried using Crystal Resampler yet, nor made any comparisons.

If you do not have another delivery method that will make use of 96k, and unless you don’t want to have the 96k as an archive file, then one should consider how the end product best reflects your original mix sound. If Crystal is that good, perhaps the method you are using is the best way to achieve that.

In the past, many folks believed that all resampling software was inherently less reliable than a high quality D-A/A-D cycle, which is why many folks tend to try to work at final sample rate from the beginning, but at a high bit depth…or to work in parallel sampling rates at the same time, fed by an analogue signal into separate converters.

It’s all in what your ears tell you. Phillipe has seemed to express a lot of confidence in Crystal Resampler, so maybe this is worth investigating for some of us old guys.

Beyond that, your workflow makes a lot of sense with how Wavelab works…at least to me.

[Swurveman]

@geezer

It does seem like a reasonable workflow, as long as you trust Crystal Resampler more than an D-A/A-D cycle. The double conversion cycles is what most high end mastering houses use, but I have not tried using Crystal Resampler yet, nor made any comparisons.

If you do not have another delivery method that will make use of 96k, and unless you don’t want to have the 96k as an archive file, then one should consider how the end product best reflects your original mix sound. If Crystal is that good, perhaps the method you are using is the best way to achieve that.

In the past, many folks believed that all resampling software was inherently less reliable than a high quality D-A/A-D cycle, which is why many folks tend to try to work at final sample rate from the beginning, but at a high bit depth…or to work in parallel sampling rates at the same time, fed by an analogue signal into separate converters.

It’s all in what your ears tell you. Phillipe has seemed to express a lot of confidence in Crystal Resampler, so maybe this is worth investigating for some of us old guys.

Beyond that, your workflow makes a lot of sense with how Wavelab works…at least to me.

Thanks for your reply geezer.

I am recording at 96 K based on this quote from Dan Lavry:

“It is possible that 1-2KHz will be more accurate when using 44.1KHz sample rate. In fact, one can do a great job for 1-2KHz with an 5KHz sampling rate. But the converter designer can not “go there”. We usually need to look at the whole audio range. 44.1KHz can be a somewhat tight squeeze, especially when we keep “piling” attenuation on that 20KHz range – most mics have 3dB loss at around 20KHz, then there is the AD with 3dB at 20K, then the speaker, the processor… Pretty soon the accumulated impact is such that there is not much 20KHz left… Moving the sampling a little higher (be it 60Kh, 88.2 or 96K) takes some of the “offenders” out of the picture”.

I did not notice a sound degradation using the Crystal Resamapler. However, I had never heard the idea that resampling software was inherently less reliable than a high quality D-A/A-D cycle. I have a pretty good Aurora 16 converter. So, I need to test this idea. I’ll record and mix some songs in Cubase at 44.1 32bit float and see what it sounds like.

Am I correct in staying in 32 bit float and saving the dithering to 16 bit as a final step in Wavelab?

Thanks again for your reply.

[geezer]

You have to realize that Lavrey is one of the principle makers of converters (i.e. analogue to digital, and digital to analogue), and he is talking about the value of converters, not about SRC (software Sampling Rate Conversion)…but I generally agree that higher sampling rates are better for archival purposes…at least up to 96k. There is some debate about the usefulness of sampling rates above that…

Your ears will have to be your guide about degradation from SRC. I have not yet heard any that I felt was anything like transparent, but I have not, as I said, used the Crystal Resampler. Converters are not really “transparent” necessarily either, but some of them sound really, really “good”. There is always some sort of tradeoff…but some folks really like some versions of SRC…so use your ears.

And yes, you want to reserve using dither until the very last stage.

Realize that anytime you do an edit, level change, fade in Wavelab, the result is automatically 32bit float (within the level change, fade, edit, etc.), even if the file you are working with is at 16bit. Therefore, if you convert files to 16bit at any point, you will lose your final dither back to 32bit float if you do any additional editing past this. (This will display as 24bit when you engage the bit meter.)…So if you are printing your whole project or individual tracks to 16bit files, you then don’t want to do anything more to them if you are then using them to create the CD master. Markers, etc., will not affect the bit depth, but any editing will.

[Vocalpoint]

@Swurveman

3 Use Peak Master then Crystal Resampler plugins in final slots, then render without dithering a High Bit Resolution (HBR) file that is at 44.1/32 bit float. Name each song as “SongM01_HBR”. Put these in a file folder called “Mastering HBR Files”.

Thanks to anyone with experience/comments!

Last time I touched base with PG on Crystal Resampler (I was building a Batch Processor to convert high res wavs) he suggested Crystal Resampler followed by Peak Master (in that order) VS Peak Master first.

I also tack a dither on as the third element in my chain. Have switched to the iZotope MBIT+ with WL8 now in my current work. Sounds excellent to my ears.

Cheers,

VP

[Arjan_P]

@Swurveman

2. Make edits in the Audio File work space (crop file lengths to audible sections, do fade ins and outs etc) and apply Mastering Plugins to the Master Section of the Audio File work space.

4 . Bring all the HBR files into the Audio Montage. Align them in order. Name them, do all crossfades between songs. Use the Wizard to determine time between songs.

This could be quirky, when doing a full album. I wouldn’t say the workflow is wrong, but essentially you are mastering each track on its own (in step 2) and then only crossfade in the montage. With some music this makes perfect sense, but I have done albums in the past where up tempo and ballads sit together after another, and in such cases I prefer to have control over the mastering as a whole, soundwise and dynamics wise. So then I’d start out with the montage and do everything there, without intermediate 44k1/32f files.

[Swurveman]

@Arjan P

@Swurveman

2. Make edits in the Audio File work space (crop file lengths to audible sections, do fade ins and outs etc) and apply Mastering Plugins to the Master Section of the Audio File work space.

4 . Bring all the HBR files into the Audio Montage. Align them in order. Name them, do all crossfades between songs. Use the Wizard to determine time between songs.

This could be quirky, when doing a full album. I wouldn’t say the workflow is wrong, but essentially you are mastering each track on its own (in step 2) and then only crossfade in the montage. With some music this makes perfect sense, but I have done albums in the past where up tempo and ballads sit together after another, and in such cases I prefer to have control over the mastering as a whole, soundwise and dynamics wise. So then I’d start out with the montage and do everything there, without intermediate 44k1/32f files.

Thanks for your reply Arjan P.

To be clear, are you saying there’s no added value in the Crystal Resampler and that I should render to 44.1/32 bit float files in Cubase? Or, are you saying apply the Crystal Resampler in the Audio Montage Master Section?

Secondly, I am having some difficulty understanding the difference between tracks and clips-and when to use one or the other in the Audio Montage. I bought the Ask Video Wavelab tutorial, which is where I learned the work process I used, but they didn’t address tracks and clips in the Audio Montage.They focused mainly on the Audio Workspace. Is there a good place that clearly explains their difference between tracks and clips, and how/why/when to use them in an Audio Montage workflow.

[Swurveman]

@Vocalpoint

@Swurveman

3 Use Peak Master then Crystal Resampler plugins in final slots, then render without dithering a High Bit Resolution (HBR) file that is at 44.1/32 bit float. Name each song as “SongM01_HBR”. Put these in a file folder called “Mastering HBR Files”.

Thanks to anyone with experience/comments!

Last time I touched base with PG on Crystal Resampler (I was building a Batch Processor to convert high res wavs) he suggested Crystal Resampler followed by Peak Master (in that order) VS Peak Master first.

I also tack a dither on as the third element in my chain. Have switched to the iZotope MBIT+ with WL8 now in my current work. Sounds excellent to my ears.

Cheers,

VP

Thanks for your reply VP.

What is the advantage of using the Peak Master after the Crystal Resampler? The audio tutorial I watched by Ask recommended the opposite.

I am not familiar with Batch Processing conversion yet. Did the high resolution files already have mastering plugins applied?

[Swurveman]

@geezer

You have to realize that Lavrey is one of the principle makers of converters (i.e. analogue to digital, and digital to analogue), and he is talking about the value of converters, not about SRC (software Sampling Rate Conversion)…but I generally agree that higher sampling rates are better for archival purposes…at least up to 96k. There is some debate about the usefulness of sampling rates above that…

Your ears will have to be your guide about degradation from SRC. I have not yet heard any that I felt was anything like transparent, but I have not, as I said, used the Crystal Resampler. Converters are not really “transparent” necessarily either, but some of them sound really, really “good”. There is always some sort of tradeoff…but some folks really like some versions of SRC…so use your ears.

And yes, you want to reserve using dither until the very last stage.

Realize that anytime you do an edit, level change, fade in Wavelab, the result is automatically 32bit float (within the level change, fade, edit, etc.), even if the file you are working with is at 16bit. Therefore, if you convert files to 16bit at any point, you will lose your final dither back to 32bit float if you do any additional editing past this. (This will display as 24bit when you engage the bit meter.)…So if you are printing your whole project or individual tracks to 16bit files, you then don’t want to do anything more to them if you are then using them to create the CD master. Markers, etc., will not affect the bit depth, but any editing will.

Thanks again Jim.

I did find that there was more high end clarity when mixing at 96K. It would be nice to have high resolution CD option. But you also make an interesting case for the superiority of A/D-D/A at 44.1 when using hardware plugins vs. doing it at 96 K and then relying on the resampling. I did read an interesting paper that did a listening test which showed that most people can’t hear the difference between 96 K and 44.1.

I did not know that editing converts the bit rate to 32 bitfloat in Wavelab. So, thanks for that. I intend to only dither at the final CD/DDP creation stage, but it’s good to know how Wavelab works internally.

[Arjan_P]

@Swurveman

To be clear, are you saying there’s no added value in the Crystal Resampler and that I should render to 44.1/32 bit float files in Cubase?

No, I’m saying: by doing it all in the montage, you’d simply be using the 96k/32f as source files in the montage. After all is said and done there, as a final step you’d go to the delivery format of CD or whatever, including SRC and bit reduction. That way, you have no intermediate files, and you can simply adapt the rendering stage to accomodate DDP, CD, MP3 or whatever delivery.

@Swurveman

Secondly, I am having some difficulty understanding the difference between tracks and clips-and when to use one or the other in the Audio Montage. I bought the Ask Video Wavelab tutorial, which is where I learned the work process I used, but they didn’t address tracks and clips in the Audio Montage.They focused mainly on the Audio Workspace. Is there a good place that clearly explains their difference between tracks and clips, and how/why/when to use them in an Audio Montage workflow.

I don’t know that video, but here’s how I work in cases as mentioned above. I put all raw mixes in the correct order in the montage, then work on each song (=clip) independently for EQ and other FX (clip effects). Also fades are done here. Then work on dynamics for each clip, comparing them between each other. Now I only use track effects if all clips on a track need to have the same treatment, for instance cut all below 30 Hz. In WL8 we now have the montage master section effects, which I haven’t used yet, but normally it is then followed by the regular master section where I may put in a limiter, Chrystal Resampler and then dither. Render or burn, and that’s it.

You could also use tracks in a different way; put all clips that need the same treatment in the same track, and apply track effects there i.s.o. on each individual clip.

[Swurveman]

@Arjan P

I don’t know that video, but here’s how I work in cases as mentioned above. I put all raw mixes in the correct order in the montage, then work on each song (=clip) independently for EQ and other FX (clip effects). Also fades are done here. Then work on dynamics for each clip, comparing them between each other. Now I only use track effects if all clips on a track need to have the same treatment, for instance cut all below 30 Hz. In WL8 we now have the montage master section effects, which I haven’t used yet, but normally it is then followed by the regular master section where I may put in a limiter, Chrystal Resampler and then dither. Render or burn, and that’s it.

You could also use tracks in a different way; put all clips that need the same treatment in the same track, and apply track effects there i.s.o. on each individual clip.

Thanks Arjan for explaining how you use of the clips and tracks function.

I have one last question: I did some trimming of songs/clips in the Audio Montage to cut out gaps between the beginning of the file and the start of the music. When I saved the Montage, the trimming edits were not saved. Did I do something wrong in the “save” process. Or, will the Montage not save time based cuts to songs/clips.

Thanks again for your information.

[Arjan_P]

@Swurveman

I have one last question: I did some trimming of songs/clips in the Audio Montage to cut out gaps between the beginning of the file and the start of the music. When I saved the Montage, the trimming edits were not saved.

I guess you mean the actual audio files weren’t trimmed. That is perfectly normal, since clips are only a representation of the actual audio. All editing done in the montage is non-destructive, which is one of the main advantages IMO. If you want to edit something permanently in a clip, double click it, and the audio file will open in the wave editor and you can destructively edit there. When these edits are saved, the clip in the montage will be updated automatically.

[Swurveman]

@Arjan P

I guess you mean the actual audio files weren’t trimmed. That is perfectly normal, since clips are only a representation of the actual audio. All editing done in the montage is non-destructive, which is one of the main advantages IMO. If you want to edit something permanently in a clip, double click it, and the audio file will open in the wave editor and you can destructively edit there. When these edits are saved, the clip in the montage will be updated automatically.

Got it. Thanks again Arjan.

[geezer]

A bit more clarity about tracks and clips in The Montage (at least from my perspective):

Clips are, of course, just areas of files that you place in the project. You can copy and/or move them anywhere (within one track or many), as well as mute them, split them, do level adjustments ad infinitum, add effects only within the clip, etc., etc…As stated above, this is all non-destructive (to the original file being referred to), but you can also create permanently altered new files from any clip through a save or render operation.

Tracks, unlike with many multitrack software packages, are simply organizational tools…unless you are working in a multichannel (i.e., surround) project, where they can suddenly have more meaning. In a normal stereo project, however, tracks can be created and deleted at will simply for organizing whatever you are doing at the moment. If you are not using tracks for applying effects (track effects) to a group of clips on the track, deleting or creating a track has no overall effect on the sound of the project unless you are deleting or creating clips along with it.

Being able to mute and solo tracks is one of the organizational tools I frequently use. I tend to do several live albums or live radio shows every year, often compiled and edited from multiple performances, requiring me to fake audience reactions and applause transitions. In Wavelab, I will keep one track for holding applause and laughter clips (actually whole, small files) I have on file. This track will remain muted most of the time, but will be soloed when I need to find something to make a new transition work, when I find the right clip, I will either copy and paste it to a track above that is adjacent to the transition, or simply move it up there), then remute the audience response holding track until I need it again.

I will often create tracks temporarily as I record a new segment or work with an audience transition. This makes it easy to see what I am doing and slide things around, and solo things as I fine tune it. When I am happy with everything, I will often slide all the new clips onto the same track while maintaining their place in time (shift/click and drag up or down on a PC) and delete the temporary one should I wish to unclutter my workspace.

While it matters where the clips are in time, it never matters how many tracks you have them splayed over. Your mix will be the same…unless you have track specific effects going on, or are, as stated above, working with multi-channel projects with track outputs assigned. But a clip that overlaps a clip on a track above it or below it will mix the same as when it overlaps a clip on the same track.

You do have to pay attention to all your tracks, especially since you can fold the individual tracks out of site and/or reduce the number you are looking at in your workspace (reduce the height of the view). When things get complex, using the Navigator bar can help you keep from making mistakes.

I find that in the end for these heavily edited live albums with hundreds of clips, I rarely wind up having to have more than 4 tracks most of the time, even with the audience response “storage track” I described above. If I want to copy and move a clip a half hour down the line, I may temporarily create a fifth or sixth track, but the final Montage will usually only have 4, with the “storage” track staying muted during the final render/DDP creation.

It is actual a very fluid working setup, and can be used in many different ways.