not correct on mixdown
since N has had ‘on the fly’ sample rate conversion, you dont get that box telling you that the sample rates do not match, in build 2312 (if you have good eyesight) at the begining of the track it now says ‘re-sample on’ - however there is something wrong with this which can be seen on mixdown -
in the screenshot track 1 is 44100hz - track 2 is 48000 - track 3 is the mixdown track -
the 44100hz part of track 1 is rendered correctly in the mixdown, the 48000hz is approx 1/64th of a measure forward (fast) of where it should be -
http://i235.photobucket.com/albums/ee177/eurodog6/myscreenshot-31.jpg
Dr J
Good catch Doc. I am wondering if there is a problem with the summing maths in the audio engine as this seems related to the other problems we have seen on mixdown. When you apply dither you are rolling some of the LSB digits around by adding a very low level “noise” so I am thinking something is amiss in the summing maths. Does that make sense?
I have another experiment in mind if I have the time a little later. I have a dither plugin I would like to slap on the master channel and see if it affects the mixdown in the same way as using the built-in dither.
D
Well I tried my experiment… with Master FX (dither) or with NO FX at all, no n-Track built-in dither, the mixdown does not match the project. Upon inserting the rendered file into the project and making SURE the start time is zero, the newly rendered file sounds like it is playing back through a slap back delay guitar pedal. It should align perfectly with the project and play sample accurate with it. It does not.
One more thing to try, then I give up until Flavio posts a fix…
D
EDIT: The project files are 24 bit and I thought loading the rendered 16 bit file might be tossing a monkey wrench in the works but no… rendering to 24 bit yields the same WRONG results. sigh
Best not to use this feature, if you care about fidelity. Use a program like r8brain to convert all tracks to the same rate.
I did a test with no dither, all audio no midi, 24bt/48k mixdown stereo track, mixed it down to a 16bt/44100 no dither or noise shaping, no plug ins, straight mixdown. Song was 3min 42secs long.
SonarLE, Reaper, nTrack. the mixdown waves from Sonar and Reaper Identical in time, nTrack .042 longer.
Yaz
Are you sure you aren’t seeing an offset caused by compensation? There’s been a LONG standing issue (bug) where there will be a bit of dead air added to the beginning of almost any mixdown that is about the same as the plug-in compensation. It’s quite noticeable if after doing a mixdown the resulting file is inserted back into the song. This has been an issue since at least V3.
Good point phoo. Although I did do tests with NO plugins and still got erroneous mix downs. They’re not consistent either. It wouldn’t be that big of a deal if the user knew to always offset the track by XXX milliseconds every time.
I don’t see this behavior in other DAW apps I have. I think bug reports are going in and Flav is on the case… we’ll just have to wait and see.
D
I didn’t say you had to be using plug-ins. You don’t. There are other built in things that can cause compensation to be non-zero.
yes but only part of the mixdown in the image is offset.
That’s a GREAT clue that it’s not the old long standing compensation problem. Thanks!