Vinyl Gain #602
Vinyl Gain #602
Conversation
|
The test are failing as desired. This way we can see the difference to the golden sample. |
daschuer
force-pushed
the
scaler3
branch
2 times, most recently
from
06c806a
to
305ac9f
Compare
|
Hi! I've already tested in the official release, and I see that the autogain control for scratching is too much agressive, because when I spin it back at a low speed, it get too much loud and it's a little bit disgusting. Here you have a comparison with Traktor (that I think it has only a limiter or something, because cares for not being so loud), that respect the low volumes when scratching (and that's nice!) In both I spin the vinyl control back a little (more or less at a half speed), but gain mantain always the same volume (and in the Mixxx recording you can't say where I do that) |
|
Cool, I did not know that Traktor has this feature. Is there a way to configure it somehow? |
Here you have the waves with 1 KHz with start and stop only. My turntables have adjustable start and stop speeds: they are set to the longest when I tested start and stop.
Well, as I said, I feel it too much aggressive. I prefer that gain was lower to lower speeds, like Traktor does. Here, the same scratch sequence performed with both programs (without crossfader cuts)
I feel more natural some gain changing. It's ok to avoid spikes (louder than original at higher speeds) By the way, I tested Traktor with and without autogain and autolimiter, but autogain I think it's the classic option that just set a constant gain after finishing analyzing the track, and autolimiter may avoid spikes when mixing two or more tracks... but in my tests the autogain only set gain lower in all of them and I don't think the autolimiter did something.
|
|
Just a thought, the low gain at lower speeds issue could be and adjustable option as well in Mixxx preferences, because I guess some people will prefer the actual autogain set for scratching (and maybe not other DJ software have the option of change that) And Be's right: if you want me to test alphas, let me know! :) |
|
Here you have the start-stop tests of the first screen shot (Mixxx and Traktor with the 1 KHz tone). If you prefer better quality, just let me know. (Autogain and limiter were activated in Traktor, but I modified gain after recording in both tracks) |
|
It looks like Traktor limits the gain change by speed to lower speeds. Faster speeds are not effected. I hope we have windows builds soon for testing. Due to waveform scratching, it is possible to get clipping think I will limit the gain changing to just 5.5 dB to avoid that. This should be no problems because we are above the physical possible speed. @Bronxio: @Be-ing: |
|
I don't have the hardware to test this myself, but I doubt that a preference option would be appropriate. I presume there's one way to make it sound best. |
|
@Be-ing: You could use jog-wheel scratching like me. Please try to listen to the scratching sound on one hand and try to find a beat on the other hand. It is a good sight for no preference option, if you do not find any regressions. |
|
I don't have any controllers with jog wheels. I didn't realize this PR affected jog wheels. I thought it only affected vinyl control. In that case there are many other contributors who could test this. @ronso0, @radusuciu want to test? |
|
Here is the improved formula. This should be natural for normal scratching, but no clipping at insane waveform scratching speeds at default gain leveling. |
|
I agree with Be, I actually think it's not really necessary neither. If the best setting is found and there's no advantage to use it in another way, better not to give any option. |
|
Since we have actually an over proportional gain boost on high speeds, my last formular was just rubbish. I think we schould go with the original formular, but apply a hard limit at x 5 where we do not increase the gain any further. In addition, my new formular is also not 0 dB at normal speed, which is also wrong. |
…mulator speed to 5 (5.5 dB)
|
@daschuer Do you want to change the formula or is this now working as intended? Just tell me when to test and I'll give it a try |
|
Now I am satisfied. Windows builds should be here soon: |
|
Oh this feels much better! |
|
Wooops, I thought that alpha wasn't ready yet because the checks messages! I tested that alpha as well, and it's very nice! I've testing it about one hour and sounds more natural for me! I recorded a comparison with Traktor and it's very similar (with no crossfader cuts, and please don't mind the end, I stopped the tone with Mixxx and not with Traktor)
If I notice something, I'll let you know; but I think it's great like this for the official 2.1 release. Thanks, daschuer! |
|
Yeaiy! |
|
So it seems we have 3 people saying this sounds better. I cannot test it. Ready for merge? |
|
Yes, it looks like. If we get complains later, we can consider to bring the old behaviour back as preferences option. |
|
Until today, I never associated a known problem (at least by myself) with the magic formula (I thought it was a MixVibes control signal native problem or something, because I remember failures like this in MixVibes Cross); but just to be sure... Could this bug be related with it? Because gain doesn't clip, I don't think so... It only happens with fast spins: the wave position changes in the control vinyl. I recorded a video to show how it happens. |
|
This bug effects only the gain of the Mixxx output, not the transport. Please file a new bug, link the very good video and add also a vinyl pass through recording (the time code sound) from a scratching similar to that in the video. |
|
A new bug at https://bugs.launchpad.net/mixxx/+filebug |
|
Relative. Great, I'll record the scratch in pass through and open this bug then in a few moments. |
|
Ready! bug reported. |
|
I tested clicking and dragging the waveform and it sounds better with this. |
This introduces the magic vinyl gain formular, discussed
https://bugs.launchpad.net/mixxx/+bug/1450665
Currently the vinyl gain it is applied all the time.
In my test it provides actually a more natural scratching sound especial notable on Vocal samples.
I have originally planed to conditional enable the gain. But I am unsure what the conditions are since also pitch changes with a scale of 90 % are benefit from it and sounding more natural now. At a 10% pitch range there is a nearly no notable gain difference.
Any ideas, for the conditions or improvements for the magic formular?