I'm so excited about all the new timing features in DP 4.52 - I just can't seem to get them to help me make my recordings tighter.
For years, I have been using DP's tempo information to tighten up tracks, and it has worked miracles. I wouldn't consider any other platform - ever. Since DP 4.5, however, my old tricks don't seem to work anymore. I have read every thread I can find, and have found a lot of fantastic information explaining the intricacies of beat detection, but I can't seem to apply any of it without really strange behavior.
I have posted in a few threads, trying desparately to describe the strangeness I have been getting, but I still haven't found out what's wrong. So, I offer the following for Human Error Detection, which is where I'm really hoping the problem lies.
Why won't this work?:
I record a hand drum into a stereo microphone, to a click being generated by DP. What I get is a performance that is pretty tight, but I want it to follow the click more closely than I played it. So, I select the stereo track with the unedited take of drumming in it, as well as the conductor track, and Copy Selection to New Sequence from the Tracks Overview window.
Then I switch to the new sequence I just created, and go to Modify Conductor Track -> Adjust Beats. Dragging each quarter note in the timeline ever so slightly, I make it all the way to the end of the song, and define a tempo map that makes the metronome play exactly to my drum playing. Of course, this tempo map is only accurate to the quarter note, but that's close enough for me. Now, when I play the sequence, I hear my original playing, unchanged, only now, the metronome in DP plays perfectly in time with my drumming. So far, so good.
Next, I record the audio of the metronome back into a mono audio track. Then I tell DP to detect beats on this recording of the quarter note clicks. It does a perfect job, putting one faint line on each click. Then I Copy Beats from this audio click track to my stereo drumming track. (*incidentally, before BDE, I used the Copy Sequence Tempo to Soundbite command in place of this step in the past.)
Finally, I copy the drumming audio, switch back to my main sequence, and paste the drumming into a new stereo audio track. At last, I groove quantize the drumming audio to a straight groove.
What happens is truly strange! Not only does the audio not play back super tightly as it used to in every other version of DP (*before BDE), but the left and right tracks are now different from each other!! That's right, the timing of the left recording is different to that of the right!!! The pahasing that is created is kind of neat, but totally unusable. Neither side is accurate to the session tempo, and if I undo this action, delete the Analysis Files folder, and immediately repeat the Groove Quantize, I get different results, though still wildly unusable.
Where have I gone wrong? Or, can anyone tell me what the recommended way of accomplishing this is, or at least a way that works?
Thanks for being here,
-James
Timing audio - where am I going wrong?
Moderator: James Steele
Forum rules
This forum is for seeking solutions to technical problems involving Digital Performer and/or plug-ins on MacOS, as well as feature requests, criticisms, comparison to other DAWs.
This forum is for seeking solutions to technical problems involving Digital Performer and/or plug-ins on MacOS, as well as feature requests, criticisms, comparison to other DAWs.
-
musicarteca
- Posts: 705
- Joined: Fri Oct 15, 2004 10:01 pm
- Primary DAW OS: MacOS
- Location: Montreal
- Contact:
Re: Timing audio - where am I going wrong?
I truly don't understand what are you trying to accomplish and why are you using this procedure, which is totally wrong. No wonder why you get strange results.
What I am figuring out is that you want to "straighten" or quantize your performance to the click. Well, there are much simpler ways to do it:
First of all duplicate the first chunk just is case you want to go back, and work on the copied sequence.
Go to the wave editor, Select the soundbite, from the beats mini menu select "Find beat is Selection". If you press the beats tab, you should see the beats.
Now you have two ways to go:
First Possibility:
Go to the sequence editor, select the soundbite, go to the region menu and select quantize. On the menu from what to quantize select "beats within soundbite". Choose the quantization resolutions that suits the music. Press apply. that is all, the hand drum performance should be in perfect time with the metronome.
this procedure induces some time stretching, therefore there might be some audio deterioration depending on how far the beats have to move.
Second possibility:
Splice the soundbite in separate beats, this works better for short release percussion sounds. There are many ways to splice the soundbite, you can use strip silence, the scissors tool or choose create soundbites from beats from the audio menu.
Once the initial soundbite is separated select all the soundbites and choose quantize, from what to quantize choose: soundbites, select the resolution required and apply. That is all.
Groove quantize is an entirely different subject.
What I am figuring out is that you want to "straighten" or quantize your performance to the click. Well, there are much simpler ways to do it:
First of all duplicate the first chunk just is case you want to go back, and work on the copied sequence.
Go to the wave editor, Select the soundbite, from the beats mini menu select "Find beat is Selection". If you press the beats tab, you should see the beats.
Now you have two ways to go:
First Possibility:
Go to the sequence editor, select the soundbite, go to the region menu and select quantize. On the menu from what to quantize select "beats within soundbite". Choose the quantization resolutions that suits the music. Press apply. that is all, the hand drum performance should be in perfect time with the metronome.
this procedure induces some time stretching, therefore there might be some audio deterioration depending on how far the beats have to move.
Second possibility:
Splice the soundbite in separate beats, this works better for short release percussion sounds. There are many ways to splice the soundbite, you can use strip silence, the scissors tool or choose create soundbites from beats from the audio menu.
Once the initial soundbite is separated select all the soundbites and choose quantize, from what to quantize choose: soundbites, select the resolution required and apply. That is all.
Groove quantize is an entirely different subject.
Re: Timing audio - where am I going wrong?
I know the procedure I outlined sounds totally strange, but it has been a culmination of much advice I have gotten here, on other BB's, and from tech support at MOTU.
I agree 100% that what you have outlined is perfectly sensible, and should work like a dream. For that matter, the procedure I described in the first post in this thread, although convoluted, should also work.
I've tried several versions of all these methods, and I always get the same kind of trouble. I have been very seriously trying to find any way whatsoever to get audio tracks to change timing/speed/feel/tightness. You know what I mean.
I'm trying to point out that for me at least, and maybe someone else out there, that which should be happening is not. Maybe there is something about my speific setup that is causing a strangeness other users in this forum are experiencing. Maybe it is not happening to you, maybe it is, and you haven't noticed. What I'm telling you, is that for some of us, there is a definite issue, beyond operator error.
I appreciate that it is annoying to read posts from people who are having trouble, and seem like idiots. I remind you that I am not bashing or ranting, and appreciate very much the power of forums of this type, whose sole purpose is to help.
I urge anyone who has had this problem to continue to post here, so that maybe we can try to get to the bottom of this issue.
To those of you that are not having any problems of this kind, congrats! I envy you terribly, and can't tell you how much I were in your shoes. Maybe it's not possible to get to the bottom of this issue here in this forum, and I'm beginning to think that, until it happens to someone with more clout than me, people such as me will continue to be chalked into the "user error" category, and written off.
For all our sake, please look closely at any audio that you have stretched ,in any way, to make sure that your confidence in your understanding of how it is _supposed_ to work isn't clouding the fact that is actually not working.
Thanks for all the help, I'm going to reinstall everything, again, and see if that helps. I just can't think of anything else to do, which is why I came here in the first place.
-James
I agree 100% that what you have outlined is perfectly sensible, and should work like a dream. For that matter, the procedure I described in the first post in this thread, although convoluted, should also work.
I've tried several versions of all these methods, and I always get the same kind of trouble. I have been very seriously trying to find any way whatsoever to get audio tracks to change timing/speed/feel/tightness. You know what I mean.
I'm trying to point out that for me at least, and maybe someone else out there, that which should be happening is not. Maybe there is something about my speific setup that is causing a strangeness other users in this forum are experiencing. Maybe it is not happening to you, maybe it is, and you haven't noticed. What I'm telling you, is that for some of us, there is a definite issue, beyond operator error.
I appreciate that it is annoying to read posts from people who are having trouble, and seem like idiots. I remind you that I am not bashing or ranting, and appreciate very much the power of forums of this type, whose sole purpose is to help.
I urge anyone who has had this problem to continue to post here, so that maybe we can try to get to the bottom of this issue.
To those of you that are not having any problems of this kind, congrats! I envy you terribly, and can't tell you how much I were in your shoes. Maybe it's not possible to get to the bottom of this issue here in this forum, and I'm beginning to think that, until it happens to someone with more clout than me, people such as me will continue to be chalked into the "user error" category, and written off.
For all our sake, please look closely at any audio that you have stretched ,in any way, to make sure that your confidence in your understanding of how it is _supposed_ to work isn't clouding the fact that is actually not working.
Thanks for all the help, I'm going to reinstall everything, again, and see if that helps. I just can't think of anything else to do, which is why I came here in the first place.
-James
-
musicarteca
- Posts: 705
- Joined: Fri Oct 15, 2004 10:01 pm
- Primary DAW OS: MacOS
- Location: Montreal
- Contact:
Re: Timing audio - where am I going wrong?
Hi James,
I can see your concern, and believe I am only trying to help, and by no means I have intended to put you down or diminish your opinion.
If you have doubts whether you might have some configuration problems that affect the way BD works for you, follow step by step my previous post. Does it work for you? Is this the result that you expect? If you get the correct answer, then there is a good chance that your system is working fine.
I have re-read your first post and I am convinced that your problem has to do with user error, there are several ways to destroy a soundfile, what you are doing is one of them. Lets see in detail:
First of all, if you want to experiment with BD, tempo mappings, quantizing, groove quantize, etc. you only need one chunk, it is useless to duplicate it and copy modified soundfiles back and forward like you suggested, there are tremendous possibilities to get unexpected results this way. Try to experiment using one sequence only, all the functions in BD are available to be used within one sequence.
You then create a tempo map from the performance by Adjusting beats. This is fine and beneficial IF your goal is to structure a tempo map where the conductor track follows the performance so you can add additional instruments to the adjusted click. But your goal is totally different, you want the opposite, you want your performance to follow a straight click, therefore this step is unnecessary.
Later you copy the beats from the click whose tempo has been derived from the hand drum and copy them into the hand drum????. It is just the same as finding the beats from the hand drum to begin with.
You get the point, or at least I hope so. I am sure that you have read about some of these procedures in different BD threads, but I think that you have mixed the information and different techniques. There are many ways to get awkward results with BD, it is as simple as not following a technique to the smallest detail, but this does not represent a fault in the program, just erroneus experimentation.
Stick with me here, I am willing to help you all the way, but you have to open your mind, accept the probability of human error, and try to solve the problems one step at a time. If you want to experiment set yourself a goal, for example: quantize the performance to the click, or create a tempo map from the performance, or groove quantize the bass according to the drums, etc., and follow the technique. Try to find the fastest and sharpest way to solve a problem, and not the most convoluted one.
Not so long ago, on another BD thread, I took the time to post a sequence with a step by step BD mini tutorial that described and solved the related problem. Did you try it out?
I can see your concern, and believe I am only trying to help, and by no means I have intended to put you down or diminish your opinion.
If you have doubts whether you might have some configuration problems that affect the way BD works for you, follow step by step my previous post. Does it work for you? Is this the result that you expect? If you get the correct answer, then there is a good chance that your system is working fine.
I have re-read your first post and I am convinced that your problem has to do with user error, there are several ways to destroy a soundfile, what you are doing is one of them. Lets see in detail:
First of all, if you want to experiment with BD, tempo mappings, quantizing, groove quantize, etc. you only need one chunk, it is useless to duplicate it and copy modified soundfiles back and forward like you suggested, there are tremendous possibilities to get unexpected results this way. Try to experiment using one sequence only, all the functions in BD are available to be used within one sequence.
You then create a tempo map from the performance by Adjusting beats. This is fine and beneficial IF your goal is to structure a tempo map where the conductor track follows the performance so you can add additional instruments to the adjusted click. But your goal is totally different, you want the opposite, you want your performance to follow a straight click, therefore this step is unnecessary.
Later you copy the beats from the click whose tempo has been derived from the hand drum and copy them into the hand drum????. It is just the same as finding the beats from the hand drum to begin with.
You get the point, or at least I hope so. I am sure that you have read about some of these procedures in different BD threads, but I think that you have mixed the information and different techniques. There are many ways to get awkward results with BD, it is as simple as not following a technique to the smallest detail, but this does not represent a fault in the program, just erroneus experimentation.
Stick with me here, I am willing to help you all the way, but you have to open your mind, accept the probability of human error, and try to solve the problems one step at a time. If you want to experiment set yourself a goal, for example: quantize the performance to the click, or create a tempo map from the performance, or groove quantize the bass according to the drums, etc., and follow the technique. Try to find the fastest and sharpest way to solve a problem, and not the most convoluted one.
Not so long ago, on another BD thread, I took the time to post a sequence with a step by step BD mini tutorial that described and solved the related problem. Did you try it out?
Re: Timing audio - where am I going wrong?
Alex,
I made the downlaod from your website, and am happy to report three things.
1) This set of sequences I downloaded from your site illustrates, very excellently, how to determine average tempo ranges within soundbites, using BDE, and how to get tose averages up into the conductor track.
2) Even though the point of the downloaded tutorial is different from the point of this thread, it is very useful, and shoud be considered on of the most useful posts on ths topic (right along some from Magic Dave.)
3) The part at the end where you recommend "Adjust Soundbites to Sequence Tempo", in fact, does not do this accurately - which is the point of this thread.
I'm not trying to cause a stink here at all, and I do, very much appreciate everyone's time and help. I think we need to spend a little time defining the problem, which is the first step in the scientific process.
Due to the fact that communication is limited to typing here, I think it may suit everyone to first define the goal, which leads to the problem:
Goal:
Get a recording of a drummer that played to a click, to go more closely with that click s/he tried to play to.
Necessary guidlines;
1) Since there is more than one track (the recording is not monaural), any group of phase coherent tracks must stay that way after the process. I.e., what ever stretching/compressing that happens to one track must happen in precisely the same way to all the other tracks, or phase is lost.
2) Since the tempo of what the drummer actually played will vary slightly and differently shot to shot, any Analyzed Tempo (if necessary) must reflect the change in tempo between each and every time any drum in the entire kit is played. I mention this due to the fact that the Analyze Soundbite Tempo command tends to find short runs of averages accross detected/adjusted beats.
Real world scenario, in layman's terms:
I wrote a song on my guitar, and recorded a rough version of it in DP, playing to the Metronome, at a tempo I chose to fit the vibe of my song, 100 BPM. for awhile, I use some loops I have, that are 100 BPM, to hold place while I record a vocal track and a bass track. Now, I get the drummer in the studio. In his headphones, he hears the mix of my tracks including the drum loops, locked to the grid, playing at 100 BPM, while he plays along.
After I record him, I zoom up on his tracks, and notice that he is human. Overall, he did a great job, but compared to the precision accuracy of the grid (exactly 100 BPM every musical beat of the song, he is ahead and behind each beat - a little bit.
What I'd like DP to help me with is to get his tracks more tightly onto the 100 BPM grid, without sacrificing my flat 100 BPM tempo for the song.
The Acceptable Solution:
i.e.,The Only Solution, will allow me to manipulate the recordings as a batch - whatever happens to one track will be exactly what happens to all of the others. It is not acceptable to slice the files into little tiny soundbites, and then crossfade them back together. The resultant tracks must retain phase coherency against each other for obvious reasons.
If there is a way to get DP to do this, I'd love to hear it. I have exhausted everything I know trying, and have yet to read a post in any thread on any forum which leads to anything other than the exact same problems originally indicated.
This post is getting long, but I swear on my life, a greater understanding of a real problem will be had, and a greater understanding of its solution as a result. I am not the only person having this problem, as proved by many other dangling threads out there from people who are certainly experincing this very buglike behavior within DP.
When I get to the studio today, I will try to snap a bunch of screenshots and make a website of my own, where these pictures, and a description of what DP actions led to them can be had. Would this be worthwhile?
In the mean time, anyone experiencing Time Scaling wierdness, please post your experiences here, with the hopes that we can all learn where this bug lives and squash it! Also, any recommendations as to how one might go about accomplishing the above task, will be tried thourouhly, and the outcomes posted here.
Thanks again,
-James
P.S., I am the first to admit error, and enjoy nothing more than finding that the bug is in me, and not my gear. I really don't wish to come off as angry, or holier-than-thou, I merely want to get back to working happily in what I know to be the best available platform - DP. After all, some of the fun is in the hunt!
I made the downlaod from your website, and am happy to report three things.
1) This set of sequences I downloaded from your site illustrates, very excellently, how to determine average tempo ranges within soundbites, using BDE, and how to get tose averages up into the conductor track.
2) Even though the point of the downloaded tutorial is different from the point of this thread, it is very useful, and shoud be considered on of the most useful posts on ths topic (right along some from Magic Dave.)
3) The part at the end where you recommend "Adjust Soundbites to Sequence Tempo", in fact, does not do this accurately - which is the point of this thread.
I'm not trying to cause a stink here at all, and I do, very much appreciate everyone's time and help. I think we need to spend a little time defining the problem, which is the first step in the scientific process.
Due to the fact that communication is limited to typing here, I think it may suit everyone to first define the goal, which leads to the problem:
Goal:
Get a recording of a drummer that played to a click, to go more closely with that click s/he tried to play to.
Necessary guidlines;
1) Since there is more than one track (the recording is not monaural), any group of phase coherent tracks must stay that way after the process. I.e., what ever stretching/compressing that happens to one track must happen in precisely the same way to all the other tracks, or phase is lost.
2) Since the tempo of what the drummer actually played will vary slightly and differently shot to shot, any Analyzed Tempo (if necessary) must reflect the change in tempo between each and every time any drum in the entire kit is played. I mention this due to the fact that the Analyze Soundbite Tempo command tends to find short runs of averages accross detected/adjusted beats.
Real world scenario, in layman's terms:
I wrote a song on my guitar, and recorded a rough version of it in DP, playing to the Metronome, at a tempo I chose to fit the vibe of my song, 100 BPM. for awhile, I use some loops I have, that are 100 BPM, to hold place while I record a vocal track and a bass track. Now, I get the drummer in the studio. In his headphones, he hears the mix of my tracks including the drum loops, locked to the grid, playing at 100 BPM, while he plays along.
After I record him, I zoom up on his tracks, and notice that he is human. Overall, he did a great job, but compared to the precision accuracy of the grid (exactly 100 BPM every musical beat of the song, he is ahead and behind each beat - a little bit.
What I'd like DP to help me with is to get his tracks more tightly onto the 100 BPM grid, without sacrificing my flat 100 BPM tempo for the song.
The Acceptable Solution:
i.e.,The Only Solution, will allow me to manipulate the recordings as a batch - whatever happens to one track will be exactly what happens to all of the others. It is not acceptable to slice the files into little tiny soundbites, and then crossfade them back together. The resultant tracks must retain phase coherency against each other for obvious reasons.
If there is a way to get DP to do this, I'd love to hear it. I have exhausted everything I know trying, and have yet to read a post in any thread on any forum which leads to anything other than the exact same problems originally indicated.
This post is getting long, but I swear on my life, a greater understanding of a real problem will be had, and a greater understanding of its solution as a result. I am not the only person having this problem, as proved by many other dangling threads out there from people who are certainly experincing this very buglike behavior within DP.
When I get to the studio today, I will try to snap a bunch of screenshots and make a website of my own, where these pictures, and a description of what DP actions led to them can be had. Would this be worthwhile?
In the mean time, anyone experiencing Time Scaling wierdness, please post your experiences here, with the hopes that we can all learn where this bug lives and squash it! Also, any recommendations as to how one might go about accomplishing the above task, will be tried thourouhly, and the outcomes posted here.
Thanks again,
-James
P.S., I am the first to admit error, and enjoy nothing more than finding that the bug is in me, and not my gear. I really don't wish to come off as angry, or holier-than-thou, I merely want to get back to working happily in what I know to be the best available platform - DP. After all, some of the fun is in the hunt!
-
jimjacobsen
Re: Timing audio - where am I going wrong?
James,
I know it's off topic, but you write very well.
Jim
I know it's off topic, but you write very well.
Jim
Re: Timing audio - where am I going wrong?
jimj,
Thanks! I noticed that you only have 4 posts here, and while that may not mean much by way of your experience on bulletin boards such as these, I feel obliged to tell you that you have, in one small post, hit the nail on the head in terms of the positive, helpful vibe of these places.
This site, as I'm sure you're aware, is not funded, moderated, or in any way affiliated with the company, MOTU, which it serves so dutifully. You little post serves very strongly as a reminder to folks like me who have haunted these threads for years, that it's the kindness of others which backbones success.
I hope that you find what what you came here to find, and that the rants/etc. here don't scare you off. The people who come here typically are seasoned professionals, whose immeasureable experience can only be quatified by its presence throughout the threads in these forums.
The value of this website to a company like MOTU is largely underplayed - even unrecognized. That said, their flagship, Digital Performer, is still the best platform out there, at any budget. There ARE solutions, and the bleeding edge is approachable thanks to folks who post here.
I'm off to have a look at your other posts to see if there's anything I can lend by way of help.
Thanks again, I was perhaps more delighted to find your post than the evasive solution I thought it might be when I logged on.
Cheers,
-James
Thanks! I noticed that you only have 4 posts here, and while that may not mean much by way of your experience on bulletin boards such as these, I feel obliged to tell you that you have, in one small post, hit the nail on the head in terms of the positive, helpful vibe of these places.
This site, as I'm sure you're aware, is not funded, moderated, or in any way affiliated with the company, MOTU, which it serves so dutifully. You little post serves very strongly as a reminder to folks like me who have haunted these threads for years, that it's the kindness of others which backbones success.
I hope that you find what what you came here to find, and that the rants/etc. here don't scare you off. The people who come here typically are seasoned professionals, whose immeasureable experience can only be quatified by its presence throughout the threads in these forums.
The value of this website to a company like MOTU is largely underplayed - even unrecognized. That said, their flagship, Digital Performer, is still the best platform out there, at any budget. There ARE solutions, and the bleeding edge is approachable thanks to folks who post here.
I'm off to have a look at your other posts to see if there's anything I can lend by way of help.
Thanks again, I was perhaps more delighted to find your post than the evasive solution I thought it might be when I logged on.
Cheers,
-James
Re: Timing audio - where am I going wrong?
Interesting topic/problem. And James you are a good writer!
Now this is only my opinion so: Producing takes a lot of different forms. That said It seems like it's a bit backwards to try to get a live drummer to sync to a guitar performance that is recorded to a loop and then have to work the now live drum performance into the quantized feel of the loop driven recording.
The beauty of the live player IS the human part of it. For what it's worth I have always begun the recording with the drummer and bass player [if it's going to be live or keybord. Have them and any other rhy section players [guitars whatever ] play/sing along as scratch or final tracks depending on what you get. Whatever the case I always let the drummer drive the push and pull of the groove so the piece sounds . . .well. . like there's a live drummer on it!
If it's a loop-driven-quantized tight kind of piece then you are having the loop[s] be the primary drum part of the final track. Any live percussion against that is used for the reason that it adds a LIVE feel to those moments that it is in the track. It seems like you are defeating the beauty of what the 2 worlds accomplish in these recordings.
Beyond that, it sufices to say that if the drummer you are having play your parts is so far off the click that you are putting all this energy into fixing their parts then I would have to recommend trying some other players. [ Not trying to be nasty or anything but some guys are great in the club but not under cans in the studio. ]
Now this is only my opinion so: Producing takes a lot of different forms. That said It seems like it's a bit backwards to try to get a live drummer to sync to a guitar performance that is recorded to a loop and then have to work the now live drum performance into the quantized feel of the loop driven recording.
The beauty of the live player IS the human part of it. For what it's worth I have always begun the recording with the drummer and bass player [if it's going to be live or keybord. Have them and any other rhy section players [guitars whatever ] play/sing along as scratch or final tracks depending on what you get. Whatever the case I always let the drummer drive the push and pull of the groove so the piece sounds . . .well. . like there's a live drummer on it!
If it's a loop-driven-quantized tight kind of piece then you are having the loop[s] be the primary drum part of the final track. Any live percussion against that is used for the reason that it adds a LIVE feel to those moments that it is in the track. It seems like you are defeating the beauty of what the 2 worlds accomplish in these recordings.
Beyond that, it sufices to say that if the drummer you are having play your parts is so far off the click that you are putting all this energy into fixing their parts then I would have to recommend trying some other players. [ Not trying to be nasty or anything but some guys are great in the club but not under cans in the studio. ]
Dual 2G-G5/2G ram/Digidesign HD2 Accel/Sony DMX R-100/MeyerHD-1&NS-10s - Hafler Ars Nova/DP5.12/PT7.3.1/2.0+TByte/GigaS3
Re: Timing audio - where am I going wrong?
Now M2 has some great points, I'm not trying to take sides, but he's right,... the best thing to do is to start with the drums and a few basic scratch tracks, record a few times till you get the drums the way you and the drummer like,... then everyone else plays to the beat of "his" drums.
I think sometimes users expect too much from the features in dp like BDE. Users often expect magical results to a sloppy track and that's just not the case in the real world, even in dp which I love dearly!
Just my thoughts.... good topic, oh btw James, you do write very well.
dpDan
<small>[ April 13, 2005, 06:46 PM: Message edited by: dpdan ]</small>
I think sometimes users expect too much from the features in dp like BDE. Users often expect magical results to a sloppy track and that's just not the case in the real world, even in dp which I love dearly!
Just my thoughts.... good topic, oh btw James, you do write very well.
dpDan
<small>[ April 13, 2005, 06:46 PM: Message edited by: dpdan ]</small>
Re: Timing audio - where am I going wrong?
Dear god, I totally I agree with you! The problem is that Beat Detective is an enormous part of so many tracks on the radio today - we the people of the Independent Movement rely on companies like MOTU to level the playing field for us a bit.
I conjugated the post above so that people reading could understand where there is an actual bug in DP (maybe two bugs actually). I have been having the hardest time explaing where these bugs live, and I think this is entirely due to the fact the DP's new BDE is not the norm, Pro Tools has had it for years, and main stream music is addicted to it. DP has stepped up to the plate with its new 4.5 release, and it does not yet provide the opportunity to break free from the other, more expensive, competition.
What if Vinny Caliuta comes in and lays down the ultimate groove - wouldn't you want to be the bass player that's "glued to the kick"? What if you alone have an idea for a tempo map to a song that noone has played on yet - perhaps dictated by actions on a film, but you want _everyone_ to play to this groove _extremely accurately_.
This is the promise of many of the new features in DP 4.5x, and the fact is, the reality is currently different. Of course there is the new consolidated window business, but its chargeable value is not quite as high it seems...
This is not to say that the best players in the world don't also play on mainstream recordings - they do indeed. In fact, I have recorded a number of them, and they TOTALLY KICK ASS!!!! It's amazing to have a look at an L.A. studio players tracks after he leaves (you know "under the microscope" / at intense horizontal zoom). What you see isn't always crazy robot like tightness to a square grid, but, when you get used to what you're looking at, you can actually see an artform that should not, and could not be duplicated by all this crazy crap we dream up and get addicted to nowadays - like Beat Detection. When you "get used to what you're looking at", to quote myself, you find that you want accurate, dependable control over the timing of all of the audio you record. After all, you do have control over every single other aspect - pitch, volume, EQ, compression, reverberatoin - everything. That's why it's called "Pro Audio".
The point I will one day make in this thread, is that tightness, to any groove, human, square, or otherwise, can not be increased by the methods that MOTU has prescribed with its new technologies, that, incidentally, are currently for sale. BDE does, very cleverly, detect beats - what's not working is the manipulation of the audio that contains those beats once they have been detected and or manipulated. It just isn't working - yet.
Look at it this way, if you go and buy this years new Ford Explorer, and it happens to have a faulty alternator by design, you can get them to replace it, due to the warrantee - in fact, they will probably make you aware of the recall effort already underway, or you could hear about the recall on Oprah.
On the other hand, if the new Ford Grandson that's recently taken over, lends you a new prototype of next year's model explorer for a weekend, or better yet sells you one, then you probably will gladly understand that there is no warantee expressed or implied. Does this make the ride not worth taking? This is very much the dilemma with software today.
You probably have encountered many other posts which describe us, the consumers, as paying beta testers. It's only the underlying bitching that bothers me about these posts. Complaining should be reserved for the ears and time of Tech Support professionals whose income is provided by the creators of the product that causes it. All of us here are here by choice, to help and be helped. As such, maybe we are beta testers to a degree.
Would we have it any other way?
-James
P.S., My drummer does kick ass, and I'd secretly like to make him in charge of the grooves on our album. What I love about beat detection in general is that it allows me lean on those with the gifts they got that I (the bass player) may not have gotten in such abundance! Multitrack audio, most commonly, live drum set nowadays, seems to be one of the last bastions of a multitrack, phase coherent, "microscopable" example that proves that this technology is not yet a reality in MOTU software.
However, I'm not going to bring this up in this thread until I get support from some other "realized sufferers". Please post!
I conjugated the post above so that people reading could understand where there is an actual bug in DP (maybe two bugs actually). I have been having the hardest time explaing where these bugs live, and I think this is entirely due to the fact the DP's new BDE is not the norm, Pro Tools has had it for years, and main stream music is addicted to it. DP has stepped up to the plate with its new 4.5 release, and it does not yet provide the opportunity to break free from the other, more expensive, competition.
What if Vinny Caliuta comes in and lays down the ultimate groove - wouldn't you want to be the bass player that's "glued to the kick"? What if you alone have an idea for a tempo map to a song that noone has played on yet - perhaps dictated by actions on a film, but you want _everyone_ to play to this groove _extremely accurately_.
This is the promise of many of the new features in DP 4.5x, and the fact is, the reality is currently different. Of course there is the new consolidated window business, but its chargeable value is not quite as high it seems...
This is not to say that the best players in the world don't also play on mainstream recordings - they do indeed. In fact, I have recorded a number of them, and they TOTALLY KICK ASS!!!! It's amazing to have a look at an L.A. studio players tracks after he leaves (you know "under the microscope" / at intense horizontal zoom). What you see isn't always crazy robot like tightness to a square grid, but, when you get used to what you're looking at, you can actually see an artform that should not, and could not be duplicated by all this crazy crap we dream up and get addicted to nowadays - like Beat Detection. When you "get used to what you're looking at", to quote myself, you find that you want accurate, dependable control over the timing of all of the audio you record. After all, you do have control over every single other aspect - pitch, volume, EQ, compression, reverberatoin - everything. That's why it's called "Pro Audio".
The point I will one day make in this thread, is that tightness, to any groove, human, square, or otherwise, can not be increased by the methods that MOTU has prescribed with its new technologies, that, incidentally, are currently for sale. BDE does, very cleverly, detect beats - what's not working is the manipulation of the audio that contains those beats once they have been detected and or manipulated. It just isn't working - yet.
Look at it this way, if you go and buy this years new Ford Explorer, and it happens to have a faulty alternator by design, you can get them to replace it, due to the warrantee - in fact, they will probably make you aware of the recall effort already underway, or you could hear about the recall on Oprah.
On the other hand, if the new Ford Grandson that's recently taken over, lends you a new prototype of next year's model explorer for a weekend, or better yet sells you one, then you probably will gladly understand that there is no warantee expressed or implied. Does this make the ride not worth taking? This is very much the dilemma with software today.
You probably have encountered many other posts which describe us, the consumers, as paying beta testers. It's only the underlying bitching that bothers me about these posts. Complaining should be reserved for the ears and time of Tech Support professionals whose income is provided by the creators of the product that causes it. All of us here are here by choice, to help and be helped. As such, maybe we are beta testers to a degree.
Would we have it any other way?
-James
P.S., My drummer does kick ass, and I'd secretly like to make him in charge of the grooves on our album. What I love about beat detection in general is that it allows me lean on those with the gifts they got that I (the bass player) may not have gotten in such abundance! Multitrack audio, most commonly, live drum set nowadays, seems to be one of the last bastions of a multitrack, phase coherent, "microscopable" example that proves that this technology is not yet a reality in MOTU software.
However, I'm not going to bring this up in this thread until I get support from some other "realized sufferers". Please post!
-
TheCoalman
- Posts: 88
- Joined: Thu Oct 28, 2004 10:01 pm
- Primary DAW OS: MacOS
- Location: Nashville TN
Re: Timing audio - where am I going wrong?
My question is how do you get rid of all the pops and clicks that happen after you create soundbites from beats and then quantize.........Smooth Audio Edits doesn't do it, and it's very very time consuming to go into each little soundbite and crossfade........
TheCoalman
G5 2.7 DP/10.4.11/3G RAM/DP 4.61/5.12/Mach 5 1.2/828/Express XT/UAD-1 4.2/Spectrasonics/EWSQ Gold/Sampletank 2.9
G5 2.7 DP/10.4.11/3G RAM/DP 4.61/5.12/Mach 5 1.2/828/Express XT/UAD-1 4.2/Spectrasonics/EWSQ Gold/Sampletank 2.9
-
musicarteca
- Posts: 705
- Joined: Fri Oct 15, 2004 10:01 pm
- Primary DAW OS: MacOS
- Location: Montreal
- Contact:
Re: Timing audio - where am I going wrong?
Hi James, nice thread, I hope something beneficial will come out of it. There was a post a few months ago by Magic Dave that dealt with the problem that you are presenting, that is adjusting a multitrack drum to a click, unfortunately I don't seem to find it.
We all agree and is a known fact that time compression/expansion induces audio artifacts, how serious are these artifacts will depend on the situation, and if they are usable or not will depend on your judgment.
I have not tried to correct a situation like yours with beat detection. Actually I have experimented a lot with BD because I want to get as sharp with the technology as possible, but in real world situations I have only seldom used it. I will avoid to work with time compressed/expanded audio at all times, unless it is unavoidable. I am a producer, and always records several takes for every instrument, until the musical material is completely error safe covered, then I would prepare a comp. take for each instrument, including multitrack drums, editing the best pieces for each part.
If you look a few years back you will probably remember that technologies like pitch shifting were almost unusable because of the amount of artifacts that it produced, it was not considered a bug back then. Technology has advanced very quickly is the last few years, but still it is not the answer to every problem. With your example you are pushing BD possibilities to the limit. Once again, and I think that this is the main point to the thread, I am trying to convince you that this is not a bug.
I know that you are very concern about time integrity, that is why I don't think BD will totally solve your problem. timing and phasing problems are there almost all the time. Think about it for a minute. You do a multitrack drum recording, probably on the overhead and HH channels you will have some likage from the snare drum, right there from the beginnign that represents a phase incoherence determined by de distance from the snare mic to the snare, from the HH mic to the snare and from the OH mics to the snare, all this translates to phase and timing incoherence, but guess what, the ear and human brain are able to tolerate them, a similar kind of phase cancelation will be experienced if you go to a concert hall and hear an orchestra, you will hear the direct sound comming from the stage mixed will the sound bouncing off the roof, the back wall and the side walls. Try to record an orchestra with two mics on the stage and to ambient mics at the back of the hall, then zoom the audio signals in your computer, you probably won't like what you see, but like what you hear. When in doubt close your eyes and listen, try to zoom in with your ears, let them be the final judge. Don't worry to much on what you see on the screen if you are satisfied with what you hear.
Imagine that you are recording an acoustic group that have to play close togheter at the same time. What do you do if one musician makes a mistake? You have to repeat the complete segment, you can't punch in the individual instrument if there is bleed between the mics. I think that the same situation applies to you multitrack drum example.
Going back to your example. Adjusting a multitrack drum is a very complex problem spcecially because of the leakage between the microphones. Supose an ideal situation where you had the snare and kick so well isolated (or gated), that you could splice them and quantize them into the beat, then you will have a serious problem with the overheadas where the sanre likage will be disaligned from the quantized one. One way to cope with this is to find the beats on the quantize snare and copy them to the overhead tracks and then quantize the OH. (it will only quantize the snare hits on the OH). You will have to repeat the same thing with the HH, toms or anything else that has bleed from other tracks. I will never attempt to do this but only experimenting for fun to see how far technoly gets. The only way that you can have efective multitrack quantization is if you use a drumer playing MIDI pads, where every part is separated.
If you have a great drumer, then it would be better to construct a tempo map based on his perfromance and try to adjust everyone else to it. Have you tried creating a groove from the drums performance and groove quantize the bass for example?
By the way you referenced PT and beat detective, if you think there are no timing issues there, I suggest that you take a closser look.
We all agree and is a known fact that time compression/expansion induces audio artifacts, how serious are these artifacts will depend on the situation, and if they are usable or not will depend on your judgment.
I have not tried to correct a situation like yours with beat detection. Actually I have experimented a lot with BD because I want to get as sharp with the technology as possible, but in real world situations I have only seldom used it. I will avoid to work with time compressed/expanded audio at all times, unless it is unavoidable. I am a producer, and always records several takes for every instrument, until the musical material is completely error safe covered, then I would prepare a comp. take for each instrument, including multitrack drums, editing the best pieces for each part.
If you look a few years back you will probably remember that technologies like pitch shifting were almost unusable because of the amount of artifacts that it produced, it was not considered a bug back then. Technology has advanced very quickly is the last few years, but still it is not the answer to every problem. With your example you are pushing BD possibilities to the limit. Once again, and I think that this is the main point to the thread, I am trying to convince you that this is not a bug.
I know that you are very concern about time integrity, that is why I don't think BD will totally solve your problem. timing and phasing problems are there almost all the time. Think about it for a minute. You do a multitrack drum recording, probably on the overhead and HH channels you will have some likage from the snare drum, right there from the beginnign that represents a phase incoherence determined by de distance from the snare mic to the snare, from the HH mic to the snare and from the OH mics to the snare, all this translates to phase and timing incoherence, but guess what, the ear and human brain are able to tolerate them, a similar kind of phase cancelation will be experienced if you go to a concert hall and hear an orchestra, you will hear the direct sound comming from the stage mixed will the sound bouncing off the roof, the back wall and the side walls. Try to record an orchestra with two mics on the stage and to ambient mics at the back of the hall, then zoom the audio signals in your computer, you probably won't like what you see, but like what you hear. When in doubt close your eyes and listen, try to zoom in with your ears, let them be the final judge. Don't worry to much on what you see on the screen if you are satisfied with what you hear.
Imagine that you are recording an acoustic group that have to play close togheter at the same time. What do you do if one musician makes a mistake? You have to repeat the complete segment, you can't punch in the individual instrument if there is bleed between the mics. I think that the same situation applies to you multitrack drum example.
Going back to your example. Adjusting a multitrack drum is a very complex problem spcecially because of the leakage between the microphones. Supose an ideal situation where you had the snare and kick so well isolated (or gated), that you could splice them and quantize them into the beat, then you will have a serious problem with the overheadas where the sanre likage will be disaligned from the quantized one. One way to cope with this is to find the beats on the quantize snare and copy them to the overhead tracks and then quantize the OH. (it will only quantize the snare hits on the OH). You will have to repeat the same thing with the HH, toms or anything else that has bleed from other tracks. I will never attempt to do this but only experimenting for fun to see how far technoly gets. The only way that you can have efective multitrack quantization is if you use a drumer playing MIDI pads, where every part is separated.
If you have a great drumer, then it would be better to construct a tempo map based on his perfromance and try to adjust everyone else to it. Have you tried creating a groove from the drums performance and groove quantize the bass for example?
By the way you referenced PT and beat detective, if you think there are no timing issues there, I suggest that you take a closser look.
Re: Timing audio - where am I going wrong?
That's a good question, but I'd like to take it up in another thread. The reason is that I'd like to keep this thread focused on the time scaling methods, since that's where the bugs I'm trying to identify live. The method you're talking about slices up the audio into many different little soundbites, and then quantizes the leading edge of each new sound bite. Then you are expected to get them all to cross fade back together. For some material, this method is quite excellent, but, again, which is better, and why, is not the focus of tihs tread.
I'm determined to get these time scaling bugs exposed and exterminated. After you detect beats in the files you mention, but before you have DP slice on beats, open up the quantize window, and scroll down to Beats within Soundbites, choose a resolution, and go for it.
If you look closely at the results, you will witness the bugs I'm talking about. It will be especially easy to see them on recordings that have multiple sources of the same performace (such as two mics on a gtr, or 10 mics on drums.) This is true only because it's easier to see two tracks that looked virtually identical to each other change, and be changed differently by DP, even if they have the same beats - copied from one to the other.
Try it, and let me know what you think,
-James
I'm determined to get these time scaling bugs exposed and exterminated. After you detect beats in the files you mention, but before you have DP slice on beats, open up the quantize window, and scroll down to Beats within Soundbites, choose a resolution, and go for it.
If you look closely at the results, you will witness the bugs I'm talking about. It will be especially easy to see them on recordings that have multiple sources of the same performace (such as two mics on a gtr, or 10 mics on drums.) This is true only because it's easier to see two tracks that looked virtually identical to each other change, and be changed differently by DP, even if they have the same beats - copied from one to the other.
Try it, and let me know what you think,
-James
Re: Timing audio - where am I going wrong?
Alex - sorry, we were posting at the same time, and I was responding to the post above yours.
I totally agree with you again -this is not a technique that should in any way replace good tracking technique that leads you to a comp which doesn't need any of this stuff.
The reason I am insistant about this bug, is that it does exist, and getting time scaling to a degree of evolution similar to pitch scaling will not be possible if these early bugs are not ironed out. All examples are extreme to attemp to illustrat a point. I have been experimenting like mad trying to find out exactly where this bug my be, and I think it is at least associated with time scaling in general.
Here's a very simple test that anyone can do to illustrate something that is very wrong with DP - I'm positive you'll agree. It may also be related to the problems that I was encountering when I started looking for a workaround to this bug, but I can't find one. Here it is:
Record a short click track onto two separate mono audio tracks, as well as one stereo audio track. Since the click is being generated by DP, and recorded back into DP, yo should hav no trouble seeing that these tracks are very similar. Now zoom up and cut out one click across all three tracks. Next select the two mono and one stereo regions that you have just cut out, and copy them. Run out to bar 50 or something, and change the tempo to something slightly lower than that which the click was recorded. Say you recorded it at 120 BPM, change the tempo to 115. Paste the two mono and one stereo click into the three tracks out here at the new tempo.
Set the grid to quarter notes. Now, holding the Command key, stretch the little soundbites so that they end at bar 50 beat 2 (exactly one beat in length as they were when you recorded them, only at the new tempo.) Command ensures that you can only stretch to the grid.
You will notice something wrong. First, the click in track one will be a different length to that in track two, even though they started at the same time, and were stretched to what should have beenthe same length. Next, you might see abunch of loud errant data at the end of each new fil, that is not time-stretching-artifact-like, but rather crazy - you'll see it. Also, you could notice that the left and right sides of the stereo track are the same as each other in length, although still not exactly one beat long.
This last bit is of interest. Why is there consistency to the stretch of two separate audio files when they are in a stereo track, but when they are recorded to two separate mono tracks they lose that consistency??
No matter what you think, ethically, of any of these abilities, it isn;'t the ethics, or the sonic quality that has me picqued here in this thread. It is a eral problem with how DP is scaling time. My feeling is that perhaps if this is pointed out and remedied, it could possibly make any portion of DP that requires time scaling within it to behave more appropriately.
By the way, you get the same results with the click stretching mentioned above if you use the sideways hand, or if you use th scale time command.
See what I mean?
-James
I totally agree with you again -this is not a technique that should in any way replace good tracking technique that leads you to a comp which doesn't need any of this stuff.
The reason I am insistant about this bug, is that it does exist, and getting time scaling to a degree of evolution similar to pitch scaling will not be possible if these early bugs are not ironed out. All examples are extreme to attemp to illustrat a point. I have been experimenting like mad trying to find out exactly where this bug my be, and I think it is at least associated with time scaling in general.
Here's a very simple test that anyone can do to illustrate something that is very wrong with DP - I'm positive you'll agree. It may also be related to the problems that I was encountering when I started looking for a workaround to this bug, but I can't find one. Here it is:
Record a short click track onto two separate mono audio tracks, as well as one stereo audio track. Since the click is being generated by DP, and recorded back into DP, yo should hav no trouble seeing that these tracks are very similar. Now zoom up and cut out one click across all three tracks. Next select the two mono and one stereo regions that you have just cut out, and copy them. Run out to bar 50 or something, and change the tempo to something slightly lower than that which the click was recorded. Say you recorded it at 120 BPM, change the tempo to 115. Paste the two mono and one stereo click into the three tracks out here at the new tempo.
Set the grid to quarter notes. Now, holding the Command key, stretch the little soundbites so that they end at bar 50 beat 2 (exactly one beat in length as they were when you recorded them, only at the new tempo.) Command ensures that you can only stretch to the grid.
You will notice something wrong. First, the click in track one will be a different length to that in track two, even though they started at the same time, and were stretched to what should have beenthe same length. Next, you might see abunch of loud errant data at the end of each new fil, that is not time-stretching-artifact-like, but rather crazy - you'll see it. Also, you could notice that the left and right sides of the stereo track are the same as each other in length, although still not exactly one beat long.
This last bit is of interest. Why is there consistency to the stretch of two separate audio files when they are in a stereo track, but when they are recorded to two separate mono tracks they lose that consistency??
No matter what you think, ethically, of any of these abilities, it isn;'t the ethics, or the sonic quality that has me picqued here in this thread. It is a eral problem with how DP is scaling time. My feeling is that perhaps if this is pointed out and remedied, it could possibly make any portion of DP that requires time scaling within it to behave more appropriately.
By the way, you get the same results with the click stretching mentioned above if you use the sideways hand, or if you use th scale time command.
See what I mean?
-James
Re: Timing audio - where am I going wrong?
James,
I've been following this thread, and I have to admit, I still don't understand what you are trying to do, or what the "bug" is?
So to try and understand, I tried your little experiment.
I did not get what you described, in fact DP worked as it should. All three tracks were the same length. There was no "crazy" data at the end of the track.
Not sure why it dosen't work on your rig. Are you sure your setting are correct? As stated above (near the begining of this thread) that process you originally described sounds redundent and actually might be causing you to have hte problems.
Just my 2•.
<small>[ April 14, 2005, 10:52 PM: Message edited by: westla ]</small>
I've been following this thread, and I have to admit, I still don't understand what you are trying to do, or what the "bug" is?
So to try and understand, I tried your little experiment.
I did not get what you described, in fact DP worked as it should. All three tracks were the same length. There was no "crazy" data at the end of the track.
Not sure why it dosen't work on your rig. Are you sure your setting are correct? As stated above (near the begining of this thread) that process you originally described sounds redundent and actually might be causing you to have hte problems.
Just my 2•.
<small>[ April 14, 2005, 10:52 PM: Message edited by: westla ]</small>