Old 01-27-2008, 12:55 AM   #1
Billoon
Human being with feelings
 
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
Default Record output bug?

edit: There is a timing variation when using Record output(stereo, latency compensated) as compared to the midi recording, sometimes its a bit early, sometimes a bit late.

[IMG]http://img156.**************/img156/3307/roslcie3.png[/IMG]


I setup a track with Reasamp. and duplicated it...recorded the midi on the first and the output on the second. Then rendered the midi to stem and dropped the original sample on a track.

The midi/stem tracks are accurate but the recorded output is early in this case.



I tried record output(not latency compensated) but that was way later than the other two methods.


edit: This probably should have gone in the nitpick section as its only a few samples.....doh.

Curious though that if i route output of the VSTi track to a new track and record its output as i play, it comes out even earlier.

Last edited by Billoon; 01-27-2008 at 08:21 AM. Reason: edited for accuracy...
Billoon is offline   Reply With Quote
Old 01-27-2008, 06:50 AM   #2
Dstruct
Human being with feelings
 
Join Date: Jul 2006
Posts: 12,480
Default

Quote:
Originally Posted by Billoon View Post
Record output(stereo, latency compensated)
that's the WRONG option in this case! "latency compensated" means the latency reported by the asio-driver for external inputs or the manual offset.

don't mix it up with pdc. pdc is always enabled. so you just have to use "Record output (stereo)" in this case (when recording stuff internally in reaper)!



@justin: you really should give those options a different name. too many people get confused by it!
Dstruct is offline   Reply With Quote
Old 01-27-2008, 07:16 AM   #3
Billoon
Human being with feelings
 
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
Default

Isnt PDC disabled on record armed tracks and i did mention in my OP that i tried the non-LC option and there was a greater discrepancy between the recorded midi and the recorded audio in that case...so what is actually confusing me is why the LC option is closer to the midi output than the non-LC option.

Heres a pic of the same situation but using the non-LC record output option...as you can see the difference is 1048 samples late as oppased to 8 samples early when using the LC option.

[IMG]http://img168.**************/img168/3496/ronlcnk2.png[/IMG]

So if the LC option is the wrong one, why is it apparently more accurate?
Billoon is offline   Reply With Quote
Old 01-27-2008, 07:32 AM   #4
Dstruct
Human being with feelings
 
Join Date: Jul 2006
Posts: 12,480
Default

mmh, you also can confuse people by not explaining the pictures properly/completely.

WHAT recorded output is late? If you record midi, you should select "Record Output (Midi)" ...


Seems to work fine here.
Dstruct is offline   Reply With Quote
Old 01-27-2008, 07:58 AM   #5
Billoon
Human being with feelings
 
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
Default

Quote:
Originally Posted by Dandruff View Post
mmh, you also can confuse people by not explaining the pictures properly/completely.
OK. Sorry for not being clear. Its the recorded audio output that is late in post two(when using the non-LC option).

In post one...

tk1 is the recorded midi input, this track has ReaSamp loaded.

tk2 has the sample that was loaded in ReaSamp lined up with the first midi note.

tk3 is rendered stem of tk1.

tk4 is the same as tk1 except i recorded the audio output of the track using record output(stereo, latency compensated).

In post two...i did the same thing except tk1 has the recorded output using the non-LC record output(stereo) option.


Ive a attached a project file if you want to try it. Hit record and play a few midi notes, then render the midi track to stem and see if it lines up with the recorded audio.

Ive done it a few times and the LC one is always closest to the midi stem...but each time i do it there is a variation (+ and -) in timing.


What i would like to know is which one is the most accurate...the midi recording or the audio recording(LC)?
Attached Files
File Type: zip rob.zip (104.6 KB, 185 views)
Billoon is offline   Reply With Quote
Old 01-27-2008, 08:36 AM   #6
Dstruct
Human being with feelings
 
Join Date: Jul 2006
Posts: 12,480
Default

Ok, now one can understand it

I don't have this Reasamp plugin installed, so I've tested it with my Surge VSTi (with the OSC set to retrigger to get exact phase start). See the attached screesnhot. It's sampleaccurate here. Looks like this must be a problem of Reasamp?

Last edited by Dstruct; 02-25-2008 at 03:22 PM.
Dstruct is offline   Reply With Quote
Old 01-27-2008, 08:38 AM   #7
Billoon
Human being with feelings
 
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
Default

OK. Thanks for checking.

Im not using any midi offset and ill try it with another plug.
Billoon is offline   Reply With Quote
Old 01-27-2008, 08:48 AM   #8
Dstruct
Human being with feelings
 
Join Date: Jul 2006
Posts: 12,480
Default

yeah, i think the midi offset (for midi outputs) doesn't has to do anything with it.
Dstruct is offline   Reply With Quote
Old 01-27-2008, 09:19 AM   #9
Billoon
Human being with feelings
 
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
Default

OK. Tried a few other VSTis(including Shortcircuit). In every case the record output(stereo, latency compensated) option is closer(but not sample accurate) to the stem.

The non-LC option is always noticably later(double notes heard).


Im downloading the Surge demo now to see what happens there.
Billoon is offline   Reply With Quote
Old 01-27-2008, 09:48 AM   #10
Dstruct
Human being with feelings
 
Join Date: Jul 2006
Posts: 12,480
Default

Quote:
Originally Posted by Billoon View Post
OK. Tried a few other VSTis(including Shortcircuit). In every case the record output(stereo, latency compensated) option is closer(but not sample accurate) to the stem.

The non-LC option is always noticably later(double notes heard).
I don't know what you're doing. Just tested with Shortcircuit (1.1.2) -> fine. Recorded (non LC) output is sampleaccurate to the rendered stem.

Really not a single problem here.
Dstruct is offline   Reply With Quote
Old 01-27-2008, 10:04 AM   #11
Billoon
Human being with feelings
 
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
Default

Hmm, i wonder if this is system specific. Im using a M-Audio keyboard connected with a USB cable and an M-Audio card.

Hopefully a few others will try the project i posted.


Tried Surge demo...same thing. LC is either a little early or a little late and non-lc is always very late.


If this is just a setting or something, id love to work out what it is.


Dandruff...if you use the LC option is, is the recorded audio a lot earlier than the stem?
Billoon is offline   Reply With Quote
Old 01-27-2008, 10:18 AM   #12
Dstruct
Human being with feelings
 
Join Date: Jul 2006
Posts: 12,480
Default

Quote:
Originally Posted by Billoon View Post
Dandruff...if you use the LC option is, is the recorded audio a lot earlier than the stem?
no, not a lot. 591 samples earlier to be exact (with "use audio driver reported latency" enabled + manual input offset of "-90 samples")


Quote:
Originally Posted by Billoon View Post
Hmm, i wonder if this is system specific. Im using a M-Audio keyboard connected with a USB cable and an M-Audio card.
i doubt that this is caused by certain hardware, because you record the stuff internally in reaper with the "record output" modes.

Last edited by Dstruct; 01-27-2008 at 10:22 AM.
Dstruct is offline   Reply With Quote
Old 01-27-2008, 10:29 AM   #13
Billoon
Human being with feelings
 
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
Default

Thanks for testing.

Tried using an ESI Julia(that has midi input),...same result.

Hopefully someone can suggest a way to fix this. Id love to get this working properly.


Would be great if a few other people tried this.
Billoon is offline   Reply With Quote
Old 01-27-2008, 10:40 AM   #14
Dstruct
Human being with feelings
 
Join Date: Jul 2006
Posts: 12,480
Default

what has the midi-input to to with it?


in my tests i didn't record any midi-notes from my keyboard. i've just drawed in the midi-notes with my mouse into an empty item ...

Last edited by Dstruct; 01-27-2008 at 10:43 AM.
Dstruct is offline   Reply With Quote
Old 01-27-2008, 11:25 AM   #15
Billoon
Human being with feelings
 
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
Default

Quote:
Originally Posted by Dandruff View Post
what has the midi-input to to with it?


in my tests i didn't record any midi-notes from my keyboard. i've just drawed in the midi-notes with my mouse into an empty item ...
Ahh, well try record from the keyboard and see if they line up then. Thats what im doing.


Personally, id rather just capture the original performance as audio (to save the time rendering to stem and cause some synths sound slightly different when you render them).

But if recording the midi then rendering to stem is going to be a more accurate representation of the original performance(timing wise)...ill do that.


The biggest variation ive got so far is about 60 samples, so its not that big a deal...but id still rather know which one is most accurate.
Billoon is offline   Reply With Quote
Old 01-27-2008, 11:33 AM   #16
Dstruct
Human being with feelings
 
Join Date: Jul 2006
Posts: 12,480
Default

Quote:
Originally Posted by Billoon View Post
Ahh, well try record from the keyboard and see if they line up then. Thats what im doing.

yeah, know i got it. you've recorded the midi and the audio from the vsti both at the same time.

yes, in this case the recorded (non-lc) audio is late here (576 samples) too.

might be a bug!? justin?


Quote:
Originally Posted by Billoon View Post
But if recording the midi then rendering to stem is going to be a more accurate representation of the original performance(timing wise)...ill do that.

yeah, that's how you should do it (at least for the moment).
Dstruct is offline   Reply With Quote
Old 01-28-2008, 07:05 AM   #17
Billoon
Human being with feelings
 
Join Date: Sep 2006
Location: Arse end of the earth.
Posts: 2,988
Default

Whew, we got there in the end.


Thanks for confirming, Dandruff. Hopefully Justin will take a look sometime.
Billoon is offline   Reply With Quote
Old 01-30-2008, 01:44 PM   #18
Dstruct
Human being with feelings
 
Join Date: Jul 2006
Posts: 12,480
Default

justin? is this a limitation of the engine or a bug?
Dstruct is offline   Reply With Quote
Old 08-07-2008, 12:54 PM   #19
Oxytoxine
Human being with feelings
 
Oxytoxine's Avatar
 
Join Date: Aug 2007
Posts: 374
Default

are there any news on this?

not to be too expectant/barefaced, but this really should get some attention (or clarification - maybe there is a fundamental misunderstanding and there's an easy solution)

thanks!
Oxytoxine is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 12:50 PM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.