[solved] Rendered File Does Not Sound The Same As In Renoise

[TomPrice]

I made a piece of music and rendered it. When I play back the wav file, the output of a send track sounds significantly quieter than it does when I play the piece of music in Renoise. What is going on here? THank you to whoever provides assistance.

[00.1]

i believe its a -6db cut when you render selection to sample or render to disk
when i render selection to sample, i auto proceed with “+ 6.041 w/ option to cut” in the instrument settings
so when i render to disk, i don’t hear the cut

from what i’ve read, the reason for this auto cut is to allow headroom

edits:
i also auto proceed with “+ 6.041 w/ option to cut” in the VST instrument Properties before loading a VST
and proceed again after i render selection to sample
so when i render to disk, i don’t hear the cut

[taktik]

At which interpolationmode and samplerate do you render the song? If you use the same settings as you do for playback, does it then sound the same?
Some VSTs are not “samplerate safe”, some behave strange when rendering them not in “realtime speed”. Using The “sinc” interpolation might as sound different (thats the point of it).

Would be great to have at least a small snipplet of the song so that we could take a look at this problem. Could you prepare and upload one?

If you are rendering a selection, not the whole song, then 00.1 is right: There is a headroom that avoid clipping, so you have to push the output by 6 db to get it as loud as before.

[TomPrice]

I am rendering with Arguru’s sinc, at 44100 samplerate. That is the same samplerate as in my device settings.

Also, I am rendering the whole song, not a selection.

I did some experimenting to try and understand the problem and this is what happened:

First of all I turned off all effects on the send track. (there were no effects on the tracks being sent to the send track). I rendered again and the send track was still quieter than it should be. No VST instrument sounds were being sent to the send track, so this eliminates any sort of VST weirdness as the source of the problem.

Then I changed the tracks that send their output to the send track so that they just play their output instead of sending it to the send track. I rendered again and the output of all of these tracks was still quieter than it should be compared to the output of the rest of the tracks. This would seem like a very odd coincidence except there is only one sample used among all the tracks that formerly sent their output to the send track. So I figured it was something weird with the sample.

I confirmed this by making a new track and playing the sample in that track. I rendered the song again and got the same problem with the sample in the new track.

I then did some further experimenting by using the “Render Selection to Sample” feature to copy the problematic sample instrument. I adjusted the volume of the copied instrument so that it sounded identical to the original, problematic sample instrument. Then I tested what would happen to the new instrument when I rendered. The new instrument would stay at the same volume like it should, and the original instrument still got quieter in the rendered wav.

So I still haven’t figured out how to fix this, but at least I’ve figured out how to work around it…

Let me know if you still want an audio sample after all this descriptive stuff.

[taktik]

@Tom Price

I am rendering with Arguru’s sinc, at 44100 samplerate. That is the same samplerate as in my device settings.

Also, I am rendering the whole song, not a selection.

I did some experimenting to try and understand the problem and this is what happened:

First of all I turned off all effects on the send track. (there were no effects on the tracks being sent to the send track). I rendered again and the send track was still quieter than it should be. No VST instrument sounds were being sent to the send track, so this eliminates any sort of VST weirdness as the source of the problem.

Then I changed the tracks that send their output to the send track so that they just play their output instead of sending it to the send track. I rendered again and the output of all of these tracks was still quieter than it should be compared to the output of the rest of the tracks. This would seem like a very odd coincidence except there is only one sample used among all the tracks that formerly sent their output to the send track. So I figured it was something weird with the sample.

I confirmed this by making a new track and playing the sample in that track. I rendered the song again and got the same problem with the sample in the new track.

I then did some further experimenting by using the “Render Selection to Sample” feature to copy the problematic sample instrument. I adjusted the volume of the copied instrument so that it sounded identical to the original, problematic sample instrument. Then I tested what would happen to the new instrument when I rendered. The new instrument would stay at the same volume like it should, and the original instrument still got quieter in the rendered wav.

So I still haven’t figured out how to fix this, but at least I’ve figured out how to work around it…

Let me know if you still want an audio sample after all this descriptive stuff.

Thanks for testing this in detail.

Please try to render the same thing not in Arguru Sinc mode but with “Cubic” interpolation. The Sinc mode will definitely sound different on some samples. With some samples its not that noticeable, with others it can make a big difference.

If Arguru Sinc is really the cause of this, could you then please upload a small example song so that we could take a look at this here?

[TomPrice]

Rendered with Arguru’s sinc:
http://www.mediafire.com/?zaaaiaawaaa
Rendered with cubic interpolation:
http://www.mediafire.com/?aawoonaaana

In each of these files you hear sample A alternating with sample B, starting with A.
Sample A is the original sample that was having this problem.
Sample B is a sample that I got from using “Render selection to sample” on a selection containing sample A and then adjusting the volume of the resulting sample so that it sounded identical to sample A.

With cubic interpolation it sounds exactly like what I hear in renoise, with Arguru’s sinc you can hear sample A is quieter than sample B.

[vvoois (vV)]

@Tom Price

Rendered with Arguru’s sinc:
http://www.mediafire.com/?zaaaiaawaaa
Rendered with cubic interpolation:
http://www.mediafire.com/?aawoonaaana

In each of these files you hear sample A alternating with sample B, starting with A.
Sample A is the original sample that was having this problem.
Sample B is a sample that I got from using “Render selection to sample” on a selection containing sample A and then adjusting the volume of the resulting sample so that it sounded identical to sample A.

With cubic interpolation it sounds exactly like what I hear in renoise, with Arguru’s sinc you can hear sample A is quieter than sample B.

http://tutorials.renoise.com/Renoise/RenderToFile:

bBeware when using organic or bad shaped samples. They may have a perfect effect during[/b] play in the editor, when you use Arguru’s sinc interpolation, the sample may sound very different from the generated wave-file than it does in Renoise! Listen to the two examples.Cubic rendered < vs. > Arguru rendered
Notice the different sounds between the two mp3-files. Though they use the exact same sample and the exact same bitrate and frequency, the interpolation changes the sample-structure drasticly enough to create these differences. So try cubic interpolation first as well as the current frequency rate your sound-card is currently set to, to play your samples in Renoise before you do any bug-report about this. Also, some settings may affect proper play of VST instruments, e.g. instruments are sampled at 44Khz, but you render in 96Khz and those samples suddenly play at 1.5 times it’s original pitch.
Things to bare in mind

[TomPrice]

Alright, TYVM for the assistance.

[taktik]

Can we do anything to avoid such confusion in future? Not everybody will read the full tutorials before rendering stuff.

As we can not avoid that arguru sinc sounds different (thats why we actually have it), should we maybe show a warning message each time you render with sinc interpolation?

[00.1]

@taktik

Can we do anything to avoid such confusion in future? Not everybody will read the full tutorials before rendering stuff.

As we can not avoid that arguru sinc sounds different (thats why we actually have it), should we maybe show a warning message each time you render with sinc interpolation?

That’s actually pretty cool that arguru sinc sounds different, yet another subtle tool to use

After reading this, I’m more aware of its purpose
I may forget its purpose since I don’t render to disk that often
I think it would be good for new users who don’t read full tutorials

[TomPrice]

@taktik

Can we do anything to avoid such confusion in future? Not everybody will read the full tutorials before rendering stuff.

As we can not avoid that arguru sinc sounds different (thats why we actually have it), should we maybe show a warning message each time you render with sinc interpolation?

I think that would be a good addition. Maybe with one of those little checkboxes that says “do not show this message in the future” or something like that.

[taktik]

@crytek

I’ve always rendered in Argru sync and never even bothered to check on on the quality. Would rendering to Cubic be the best choice for rendering?

Depends on what you want: If you want that the rendered file sounds exactly as you have heard it before, then choose cubic. If you want this resampling quality http://www.simonv.com/tutorials/quality.php (see Renoise), then sinc is your choice.