Huge problem with the Tempo Map and audio following tempo changes

[Craig Anderton]

You're probably not interested in a workaround, but as someone who thinks tempo changes are important, I've found the best way to introduce them is in the two-track mix. That way you can record with tempo-synched effects, quantize parts if needed, but once mixed they will follow any tempo changes. The quality seems to be better as well, compared to trying to change project tempo when a bunch of clips all need to change at once. 

I use several DAWs and wrote up how to do this in Studio One if you need more details; the same technique works with Cakewalk, and works well. Bring the two-track mix into Cakewalk. Select the two-track mix, open AudioSnap, and select Clip Follows Project. The tempo track doesn't need to match the two-track master, because any tempo changes are relative, not absolute. Just move tempo markers around on the tempo track to change tempo.

This technique also makes it easy to add accelerandos, ritardandos, etc. 

If you want to hear what this sounds like, every song in my album The Singles (the link goes to YouTube) has tempo changes that I think breathe life into the songs. 

[Olaf]

On 2/28/2022 at 12:54 PM, Will_Kaydo said:

I think you have stretch on, thats why the information in side the clip is moving. It's trying to adjust to the tempo its stretched too.

Hey, Will! Normally, stretching the clip would proportionally stretch the wave content along with the edge you're stretching and in proportion to the clip length, not slide it inside inside the clip like that. And it would do it as you change the stretch value, not on trim. What's in the video it does even without stretch enabled, and with "follow project tempo" on.

If you want to check out the project, follow this link, I've aligned the clips via stretch for about half the project, around 3:55, it starts to be out of sync again, for the next two parts - I had those aligned, too, but slowed down the tempo again, by a bit. It's the same recordings, only I've bounced a few clips, after having them aligned, and it's a little less embarrassing ?. Same improvised vocals, recorded on the phone.

https://we.tl/t-d4aBu8cIEt

On 2/28/2022 at 2:36 PM, marled said:

Most of the time I can handle it by copying/moving the particular clips to a new working track (with only 1 take lane!) and then bouncing one of the clips before I do the crossfades. Sometimes a project close and reopen helps.

A lot of people say that CbB is much more bug-free than Sonar, maybe it's true for a lot of things (mixing, recording, MIDI). But in respect of clip editing I don't agree at all! There are so many bugs and hassles there.

Yeah, I agree with you. Ever since the new export dialog, I can't export a song in the same session I've been mixing - it only exports some of the audio tracks completely unprocessed, and ignores everything else. I need to reopen the project every time before export, and redo all the setting that it doesn't remember. Very frustrating. Every time I open a single take track - after comping, I like to keep everything on one lane - it arbitrarily creates a second lane and it automatically distributes the clips between them.  Another thing that bugs like hell is that when you record a new take it automatically splits all the takes in the parallel lanes, at the point where you stop recording. I have to manually remove the clips before it and resize them all, and if you're recording in a loop, the splits are very small and you have to zoom in, cause you can't grab them, and so on. Do it after every single take. It doesn't do it in all projects, and I have no idea what prompts this behavior. I've changed all the settings in the comping settings, no change. Very frustrating.

Absolutely agree with you.

 

Edited March 5, 2022 by Olaf

[Olaf]

On 3/2/2022 at 6:02 PM, Craig Anderton said:

Bring the two-track mix into Cakewalk. Select the two-track mix, open AudioSnap, and select Clip Follows Project. The tempo track doesn't need to match the two-track master, because any tempo changes are relative, not absolute. Just move tempo markers around on the tempo track to change tempo.

Hey Craig, thanks a lot for the advice! I really appreciate your audio knowledge, and have read a lot of your articles. I am interested in workarounds, but where I also need to move the parts around, to various structure variants, and adjust the tempo accordingly, very frequently, I think the process of exporting everything and reopening it separately might be very cumbersome. I've adopted manually stretching every clip for now - it's still a lot more cumbersome than just readjusting the tempo and have the clips follow along, but at least it solves the alignment issue, and it seems the simplest, under the circumstances.

I'd like to pick your brain on one issue, though. I use a lot of warmers in my mix - console channels on every track, a console bus emulator on the two-bus, for the summing, plus a transformer and crossover emulator, tapes on everything, and the end result is warm, and thick, with a good stereo image, the harshness rounded off, all good and well, but then I listen to Berlin - Marillion, which I have as a reference track for this mix, and everything seems to be so focused in their mix, no flab, with a lot more net, yet not harsh, transients, whereas mine seem pasty and diffuse, bloated, and grainy - a smooth grain, but still - it has a "grey" sound.

Would clippers help on every track? - I've got a clipper on the maser bus, cause it does have that type of effect - netter - more of a metallic brilliance, but without harshness - transients, and it sucks in the gut. You do seem to have that kind of transients in your Youtube project, do you mix through a console? Nice songs, I detect a little bit of a Mike Oldfield guitar in there, Bryan Ferry vocals - who's singing? -  and a little Pink Floyd mood combined with a Brit sound. I like the guitar sound at 10:10. Real amp, right? Did you play everything? My favorites are "My butterfly" and "Goodbye to you".

Edited March 7, 2022 by Olaf

[Craig Anderton]

On 3/5/2022 at 1:41 PM, Olaf said:

I'd like to pick your brain on one issue, though. I use a lot of warmers in my mix - console channels on every track, a console bus emulator on the two-bus, for the summing, plus a transformer and crossover emulator, tapes on everything, and the end result is warm, and thick, with a good stereo image, the harshness rounded off, all good and well, but then I listen to Berlin - Marillion, which I have as a reference track for this mix, and everything seems to be so focused in their mix, no flab, with a lot more net, yet not harsh, transients, whereas mine seem pasty and diffuse, bloated, and grainy - a smooth grain, but still - it has a "grey" sound.

Would clippers help on every track? - I've got a clipper on the maser bus, cause it does have that type of effect - netter - more of a metallic brilliance, but without harshness - transients, and it suck it the gut. You do seem to have that kind of transients in your Youtube project, do you mix through a console?

Everything is mixed in the box, no console. I don't think clippers would help, but it's hard to answer because we have quite different philosophies. I try to avoid effects as much as possible, and do a lot of my processing on tracks using destructive DSP, especially for the vocals. I think the focus you hear in my music is due to track EQ. I always start mixes in  mono, and use EQ so that each instrument owns a specific part of the frequency spectrum. Once all the parts are distinct in mono, then I start thinking about stereo. I also do a lot of multiband processing, which helps with imaging and focus (especially distorted guitar). 

I also never use processing on the master bus. I see mixing and mastering as very different processes: mixing is about obtaining a balance, mastering is about polishing that balance to perfection. After finishing a mix I let it sit for a few weeks, and then I master it. Perhaps some of the effects that you add to every track would be more effective if used while mastering? Also, don't overlook transient shapers. For masters, the best I've found is Waves Trans-X, because it's multiband. 

On 3/5/2022 at 1:41 PM, Olaf said:

Nice songs, I detect a little bit of a Mike Oldfield guitar in there, Bryan Ferry vocals - who's singing? -  and a little Pink Floyd mood combined with a Brit sound. I like the guitar sound at 10:10. Real amp, right? Did you play everything? My favorites are "My Butterfly" and "Goodbye to you".

I haven't used a real amp for recording in decades - even when I was doing concerts back in the 60s, I used clean FRFR keyboard amps, and got my sound before it hit the amp's input jack. And yes, that is me playing all the parts and doing all the singing. A comparison to Bryan Ferry is a first, though...usually people say I sound like David Byrne or Roger McGuinn. I'll take any of those  

The album The Singles was from 2019, when my wife was dying, so it was influenced by the turbulence of that year. You might want to check out some other projects. Take Me Back to Tomorrow is from 2020, and one listener called it "tropical pop" which I think is really accurate. It has an upbeat, island vibe. But for 2021, I got into a 60s retro mode with Anything Goes - lots of guitars, Chicago blues harmonica, etc. I never plan these "themes"...I don't have to worry about being "commercial," so I just do whatever feels good at the time.   

You might also like Neo-, which is short for "Neo-Psychedelic Music for the 21st Century." It was my take on what psychedelic music would sound like if it was invented today. It's the most layered, with the thickest mix and lots of guitar.  There are other projects on my YouTube channel, but I think those would interest you the most, and maybe give you some ideas

Let me know if you have any other questions, I'd be happy to help.

[marcL]

On 3/5/2022 at 8:18 PM, Olaf said:

Another thing that bugs like hell is that when you record a new take it automatically splits all the takes in the parallel lanes, at the point where you stop recording. I have to manually remove the clips before it and resize them all, and if you're recording in a loop, the splits are very small and you have to zoom in, cause you can't grab them, and so on. Do it after every single take. It doesn't do it in all projects, and I have no idea what prompts this behavior. I've changed all the settings in the comping settings, no change. Very frustrating.

Nowadays when I do loop recording I set an additional punch region that is smaller than the loop. Then I try to stop outside of the punch area. Like this I can avoid the annoying random splits!

[David Baay]

On 3/5/2022 at 12:18 PM, Olaf said:

Another thing that bugs like hell is that when you record a new take it automatically splits all the takes in the parallel lanes, at the point where you stop recording. I have to manually remove the clips before it and resize them all, and if you're recording in a loop, the splits are very small and you have to zoom in, cause you can't grab them, and so on. Do it after every single take. It doesn't do it in all projects, and I have no idea what prompts this behavior. I've changed all the settings in the comping settings, no change. Very frustrating.

This only happens in Comp recording mode and is a necessary feature. In order for Comping to work properly, every lane has to have clips with matching boundaries so that the content of only one lane is heard at any given point. Marled's solution of using a punch region to define the boundaries for all takes and stopping the transport outside it is a good one, Otherwise the usual procedure when you don't want any part of the last take it is to just delete the last take lane and then heal the split in one of the remaining takes, either by sweeping through the whole clip, starting from outside it, or by selecting the two clips and Ctrl+clicking one of them with the Comp tool.

[Olaf]

On 3/7/2022 at 7:03 AM, Craig Anderton said:

There are other projects on my YouTube channel, but I think those would interest you the most, and maybe give you some ideas

Let me know if you have any other questions, I'd be happy to help.

Thanks a lot, I'll check them all out.

On 3/7/2022 at 7:03 AM, Craig Anderton said:

Perhaps some of the effects that you add to every track would be more effective if used while mastering?

Weirdly enough, even though I have a ton of plugins in the project (about 90, including the sends and the synths), all of them almost exclusively serve one purpose, that of mirroring recording in a real room, through consoles, to tape. That's it. So I have room/studio reverbs, consoles, summers, transformers, tapes, etc. Plus the amps.

It's true to that you have to pay special attention to keep the levels constant when going into bus saturation, clipping and/or compression, so I try to do that. If not, a lot of the mud will come from exactly the master bus, and I found that out the hard way, several times. The misleading part is that what you think helps you, on a certain day, cause it thins out the honk, for instance, you realize adds a lot of bloatware lower down, that you weren't focusing on, when you listen to it the next day, with fresh ears. But I keep them on the mix bus, because I feel that without them, things don't have the necessary depth, roundness, glue and rich texture, and since they're gonna be there at the end, I think it's better to make all the EQ/compression/saturation moves through them.

On 3/7/2022 at 12:18 PM, pwalpwal said:

it did get a digital remaster in the 90s, could that be having an influence? like, you're piling on the analog emulators, but the remaster was probably trying to cut out the analog-ness at least to some degree? just a thought

That's an interesting thought. I know there are some plugins out there that can do de-saturation, but I don't think they work on complex sources, and if the remaster was done in the 90s, they wouldn't have been available. My version is plenty analog, and it sounds great - except for the cymbals, which sound like they've been recorded separately - it just has that focused sound, with smooth, but more metallic transients - that I believe comes from clipping. I've tried that on the master bus, and that's exactly the effect it had.

And I was thinking about consoles, and how people drive them, and how much gain they can take, whereas plugins seem to have a very narrow sweet spot, usually within a few fractions of a db - certainly smaller than what consoles seem to have. And I was thinking about why that was, and about that sound consoles put out, where, when driven, they seem to get these exact characteristics - punch, smooth, brilliant attack, less dynamics, and a cleaning up of the low mids, instead of swelling them, like plugins do - that clipping generates. And it dawned on me that maybe the reason for which they can get away with so much saturation, and so much drive, and have a wider sweet spot, is exactly that - that they push the signal into clipping. So I figured I'd give it a try.

I've tried an experiment, if anybody's interested. Listening test, to see if you can spot the differences, or it's just my self-suggestion. I've exported the same song snippet in three versions, one with no clipping added, the 2nd with clipping on the master bus, and the 3rd with clipping on each track - except the piano.

Note that I've gone VERY easy on the clipping, so the differences are very subtle. I'm curious if you can hear the differences among them, and what those differences are.

 

On 3/7/2022 at 5:14 PM, David Baay said:

Otherwise the usual procedure when you don't want any part of the last take it is to just delete the last take lane and then heal the split in one of the remaining takes, either by sweeping through the whole clip, starting from outside it, or by selecting the two clips and Ctrl+clicking one of them with the Comp tool.

I think the result would be the same, while recording in a loop, even with punch regions, cause you'd still go over the previous takes - because of the loop. I just wish it didn't split the previous takes automatically. On some projects it doesn't, but I haven't identified the setting. Do you know, by chance, how to set that up?

[Olaf]

@Noel Borthwick New finds, while working on the project. So it's all these things. I hope all these details are telling , so maybe I won't need to add different ones - they're all in the vein of the main phenomena described here.

These are the latest.

Resizing a clip inserts a portion of a different clip in the resize space, and moving a tempo envelope segment always creates new nodes where the initial node was.

The node values are displayed inverted in the tempo envelope.

Clip resizing becomes unresponsive, for some reason. It's happened several times in the past.

[David Baay]

3 hours ago, Olaf said:

Resizing a clip inserts a portion of a different clip in the resize space,

It's a bit hard to follow exactly what is happening, but the tool shown is for moving the split point not the slip-editing tool for moving a single clip boundary. Moving the split point will replace the waveform on the left side with the waveform on the right.

3 hours ago, Olaf said:

moving a tempo envelope segment always creates new nodes where the initial node was.

While the behavior you captured does seem a bit buggy, I would say the root of the problem is that you have two tempo nodes on the same time to begin with rather than having a single node with a 'jump' transition from the previous tempo. Moving the single node at a jump point will behave as expected.

[msmcleod]

7 minutes ago, David Baay said:

While the behavior you captured does seem a bit buggy, I would say the root of the problem is that you have two tempo nodes on the same time to begin with rather than having a single node with a 'jump' transition from the previous tempo. Moving the single node at a jump point will behave as expected.

There's nothing wrong with having the tempo nodes organised like that, and no need to make it into a jump point.

Something else is preventing the slip edit.

[David Baay]

1 minute ago, msmcleod said:

There's nothing wrong with having the tempo nodes organised like that, and no need to make it into a jump point.

Maybe there shouldn't be, but based on the video, it seems there can be...?  Jumps just seem a lot more logical and simpler to deal with - one node per change.  Seems kind of weird to see two different tempos on the same timestamp in the tempo list too.

[Olaf]

2 hours ago, David Baay said:

It's a bit hard to follow exactly what is happening, but the tool shown is for moving the split point not the slip-editing tool for moving a single clip boundary. Moving the split point will replace the waveform on the left side with the waveform on the right.

I know, that's what it's supposed to do, but it doesn't do it. If you notice in the video, it inserts an entirely different wave clip - I don't know how it chooses the particular one - in the slip space.

So, instead of having a boundary between two clips, you now have three.

2 hours ago, David Baay said:

While the behavior you captured does seem a bit buggy, I would say the root of the problem is that you have two tempo nodes on the same time to begin with rather than having a single node with a 'jump' transition from the previous tempo. Moving the single node at a jump point will behave as expected.

It's just normal envelope editing, same as with gain envelopes, pan, or anything else. It's a linear tempo change, according to the instructions: https://www.cakewalk.com/Documentation?product=Cakewalk&language=3&help=Tempo.07.html#1999891 and https://www.cakewalk.com/Documentation?product=Cakewalk&language=3&help=Tempo.04.html. You can see there're two nodes on a number of tempo changes, in the tutorial images, too. I don't believe it processes the two nodes as being exactly at the same timestamp, though, I think that's just the graphic representation, that would be the logical assumption. That how it should be, at least - adjacent, distanced by the smallest timeline resolution. In any case, it shouldn't create a third node.

2 hours ago, msmcleod said:

There's nothing wrong with having the tempo nodes organised like that, and no need to make it into a jump point.

Thanks for confirming.

2 hours ago, msmcleod said:

Something else is preventing the slip edit.

I've moved the clip on a different lane, exactly to show that there is no clip underneath it, or to either side - usually invisible clip fragments would prevent the edit, but it's not the case here. The weird thing is that if you leave it alone, and do some other operations, you come back to it, and the slip works again.

[David Baay]

14 hours ago, Olaf said:

It's just normal envelope editing, same as with gain envelopes, pan, or anything else. It's a linear tempo change, according to the instructions: https://www.cakewalk.com/Documentation?product=Cakewalk&language=3&help=Tempo.07.html#1999891 and https://www.cakewalk.com/Documentation?product=Cakewalk&language=3&help=Tempo.04.html. You can see there're two nodes on a number of tempo changes, in the tutorial images,

I actually reported as a defect a long time ago that automation editing should not create two nodes on the same timestamp. For consistency and ease of editing, it seems to me that an 'instantaneous' change should always be represented as a jump segment.

[Olaf]

3 hours ago, David Baay said:

I actually reported as a defect a long time ago that automation editing should not create two nodes on the same timestamp. For consistency and ease of editing, it seems to me that an 'instantaneous' change should always be represented as a jump segment.

To be honest I don't understand myself what the difference is between a two node linear change - which is my normal way of working with envelopes, so didn't think twice about it - and one node jump changes. At first I thought jump changes were a defect ?, and the node missing was due to a bug. It looks a little confusing.

[msmcleod]

2 hours ago, Olaf said:

To be honest I don't understand myself what the difference is between a two node linear change - which is my normal way of working with envelopes, so didn't think twice about it - and one node jump changes. At first I thought jump changes were a defect ?, and the node missing was due to a bug. It looks a little confusing.

The tempo track shares the same envelope editor as the automation lanes.  When using jumps in automation, the dotted lines literally mean no automation at all.  This means if you started a project in the middle of a dotted area, no automation would be sent out until it reached a node.

Obviously you can't have an area in the project with no tempo, so for tempos, there's pretty much no difference between a jump and a straight line.  It's really down to how you want to work.

Multiple nodes at the same position should in most cases end up being merged into one node.  If in doubt, refer to the Tempo Inspector to see if the list of tempos makes sense - and of course, there's nothing wrong with deleting nodes / changing them to jumps if you want to.  Obviously, if the rare case when you start seeing strange envelopes being drawn with nodes / lines in the wrong place, this is a bug - the easiest way around this is to simply delete the offending nodes.

FYI - this is how the tempo track works:

1. The tempo stream contains shape events in exactly the same way as an automation lane contains shapes.
2. When you change the tempo envelope, behind the scenes a linear list of tempo changes ( the tempo map ) is regenerated based on snapshots of the tempo envelope at regular intervals. By default this is every 1/16th note, although it can be changed in Cakewalk.ini to be a higher resolution using the TempoMapDecimationResolution setting.
3. The resultant tempo map is exactly the same tempo map used prior to the introduction of the tempo track. 
4. Saving a project saves both the tempo track envelope and the tempo map.

This approach has several advantages:
1.  The vast majority of the code didn't need to be changed, as it's referring to exactly the same tempo map as it did before.  In other words, the tempo track is used for only as an editor for the tempo map, and only the tempo map is used when playing, or whenever edit operations need to calculate the current tempo - just as it did before.
2.  Projects remain both forward and backward compatible with SONAR and earlier versions of CbB : Backwards compatible, because we can generate a tempo envelope from an existing tempo map if a tempo envelope doesn't exist;  forwards compatible because there's always a tempo map stored with the project - earlier versions will simply ignore the tempo envelope stream.
3.  We can use the same envelope editor for the tempo track as we do for automation envelopes.

For the most part the tempo track envelope is king, and the tempo map is only generated from the tempo track envelope.  There are a few exceptions where the opposite happens - i.e. the tempo map is manipulated directly and the tempo track envelope is then regenerated from the map:

• Loading older projects that don't have a tempo track envelope
• Set Measure / Beat at Now
• Tempo detection from audio - either through AudioSnap, or using Melodyne (drag audio track to time ruler) 

In these cases, separate Cakewalk.ini variables are used to determine the resolution the envelope should use.

The ini file settings are documented here:  http://www.cakewalk.com/Documentation?product=Cakewalk&language=3&help=Tempo.15.html

 

 

[Olaf]

On 3/14/2022 at 11:33 PM, msmcleod said:

The tempo track shares the same envelope editor as the automation lanes.  When using jumps in automation, the dotted lines literally mean no automation at all.  This means if you started a project in the middle of a dotted area, no automation would be sent out until it reached a node.

Obviously you can't have an area in the project with no tempo, so for tempos, there's pretty much no difference between a jump and a straight line.  It's really down to how you want to work.

Multiple nodes at the same position should in most cases end up being merged into one node.  If in doubt, refer to the Tempo Inspector to see if the list of tempos makes sense - and of course, there's nothing wrong with deleting nodes / changing them to jumps if you want to.  Obviously, if the rare case when you start seeing strange envelopes being drawn with nodes / lines in the wrong place, this is a bug - the easiest way around this is to simply delete the offending nodes.

Thanks, man, great explanation!

I hope the bugs get fixed over the next releases.

[Olaf]

Another problem, where slip resizing a clip comp slices another clip on a different lane. I guess it interprets it a comp slice selection, but only on that lane.

A problem where slip resizing a clip to one side actually resizes it into the opposite direction. And it doesn't allow the clip edge to be aligned to the gird line, always leaving a gap to it past which it cannot be resized.

 

[David Baay]

33 minutes ago, Olaf said:

Another problem, where slip resizing a clip comp slices another clip on a different lane.

Mouse cursor is showing the Crossfade/Split Move Tool. Try holding Ctrl to get the Slip-Edit Tool.

EDIT: The second issue may also be related to this. Hard to tell exactly what the goal is.

I would recommend bringing up one issue per thread unless they are clearly related. It's not clear the new issues are related to tempo changes.

Edited May 19, 2022 by David Baay

[Olaf]

On 5/19/2022 at 10:54 PM, David Baay said:

Mouse cursor is showing the Crossfade/Split Move Tool. Try holding Ctrl to get the Slip-Edit Tool.

The cursor is in the smart tool mode, and it's in comp split move in the example, as it should be, and as it is whenever you adjust comping points. it's not the comp slice mode, so it shouldn't cut/split any clips. And, if you look in the example, it only slices the lane above, none of the others. It looks like a bug, it's not the cursor mode that's the problem.

On 5/19/2022 at 10:54 PM, David Baay said:

EDIT: The second issue may also be related to this. Hard to tell exactly what the goal is.

I would recommend bringing up one issue per thread unless they are clearly related. It's not clear the new issues are related to tempo changes.

They're related, since this kind of behavior happens a lot with tempo changes - but not only - and I've noted these issues in that context initially, hence the post, but I've noted they're probably not only owed to tempo changes, and existed independently, but you don't need to flood the forum with a separate topic for each of these videos, particularly since they refer to the overall same kind of problems.

Which one is the better choice, making 20 new topics, one for each video, or group all similar problems together, since they're related anyway? Which one is better for referencing?

If it were up to me, I'd make a separate forum chapter for bugs only, no duplicates allowed, any various additional observations added as comments, likes on each topic from those who have the same problem. Gather all on a single list, grouped on chapters/topics, and tick all the boxes, one by one, as the problems get solved. And solve all bugs, starting with the big ones, before implementing any new feature - that is only gonna come with new ones and/or build on the old ones. Basics first. CW really needs some organization.

And there's tons of other problems - for instance notes being written in the wrong place on the grid, in the piano roll, at certain zoom factors, and, more importantly, all kinds of problems with the audio driver implementation, that are directly affecting audio performance - for instance varying the playback speed depending on the latency settings; worse performance at higher buffer values; different performance after activating and redeactivating a plugin vs. with it just deactivated, that resumes after playback stop; changing performance depending on the place you deactivate a plugin from - the macro frame vs the internal on/off - if I were to write them all down, I'd make 30 different topics off of those alone. This is trying to keep it light.

And if/when I write about all those - I would need to setup sound capture first - do you think they should go into a single topic, reflecting the larger overall problem, or do you think they should make up 30 different topics, just to fill the forum, and make it harder to follow? I don't get this. This is probably the kind of philosophy because of which these problems still exist decades later.

Edited May 23, 2022 by Olaf

[David Baay]

2 hours ago, Olaf said:

but you don't need to flood the forum with a separate topic for each of these videos, particularly since they refer to the overall same kind of problems.

Which one is the better choice, making 20 new topics, one for each video, or group all similar problems together, since they're related anyway? Which one is better for referencing?

From a development standpoint, they each have to be docuented and investigated individually; if they are found to have a common cause and solution, great, but that's for the Bakers to determine. From a forum standpoint, multi-issue posts are not as sensible as you might think; they tend to just become a confusing mess of random  responses. Videos can be helpful but they only show the actual result, not the expected result which is required to have a full understanding of the issue, (and to confirm whether there is even an issue or whether the expectation was faulty) and requires a verbal description.

2 hours ago, Olaf said:

I'd make a separate forum chapter for bugs only, no duplicates allowed, any various additional observations added as comments, likes on each topic from those who have the same problem.

The problem with this is that you're relying on users to correctly determine whether their issue is actually a  bug and whether it's the same bug as someone else reported. I can't tell you how many times someone has posted "I have the same issue", and it's not the same at all. Sometimes the same symptom, but a different cause (at least as often a  config or training issue as an actual defect), and sometimes not related at all.

2 hours ago, Olaf said:

And there's tons of other problems - for instance notes being written in the wrong place on the grid, in the piano roll, at certain zoom factors, and, more importantly, all kinds of problems with the audio driver implementation, that are directly affecting audio performance - for instance varying the playback speed depending on the latency settings; worse performance at higher buffer values; different performance after activating and redeactivating a plugin vs. with it just deactivated, that resumes after playback stop; changing performance depending on the place you deactivate a plugin from - the macro frame vs the internal on/off - if I were to write them all down, I'd make 30 different topics off of those alone. This is trying to keep it light.

I guarantee almost none of that will be reproducible by other users or the Bakers without a detailed description of the steps and much if it will turn out to he hardware/plugin/project-dependent. Personally, I have never encountered any of that or recall it being reported by anyone else in so many words with the possible exception of "worse performance at higher buffer settings" which can happen when your audio and disk buffers are not well-matched among other things.

Bottom line: if you seriously want to see anything 'fixed", it needs a separate post with reproducible steps and a copy of the project in case it's not readily reproducible in a new project started from the Basic template. 

I believe your expectation of what should happen in the last two cases is mistaken, but I can't make sense of what you're trying to accomplish or why you're talking about 'slip-resizing" a clip and expecting other clips not to be affected when you're using the split move tool instead of the single-clip crop tool. I was mistaken BTW;  It's Shift that changes this tool. not Ctrl.