Stereo Tool

OOPS - FORGOT TO UPLOAD THE NEW VERSION! PLEASE IGNORE THIS POST.....

New BETA2 online: "Allow louder highs" now works negatively (QUICK HACK: Set it to +15000 to actually get -15000! Will make that better later, this is just a quick hack to allow testing).

Soundmanfm: Still, the sound in BETA/BETA2 should NOT have been different than in 5.00. But, now you can reach the same quality as in 4.22 for 'S' sounds.

Suggested value for this slider: 15000 (to get a similar sound as in version 4.22). Of course you can also set it higher. Please let me know if this works / improves things!

See remark in previous post. I forgot to upload the new version Will do so tonight...

Test new version on a transmitter this evening.

While that I hear normally.

Сan it is needed to buy new headphones Beyerdynamic DT 990 PRO???

soundmanfm

I've now uploaded the version that I forgot to upload this morning:

New BETA2 online: "Allow louder highs" now works negatively (QUICK HACK: Set it to +15000 to actually get -15000! Will make that better later, this is just a quick hack to allow testing).

Soundmanfm: Still, the sound in BETA/BETA2 should NOT have been different than in 5.00. But, now you can reach the same quality as in 4.22 for 'S' sounds.

Suggested value for this slider: 15000 (to get a similar sound as in version 4.22). Of course you can also set it higher, to get a LOWER maximum level for high frequencies. Please let me know if this works / improves things!

Hello again Hans ,
You may not gonna believe this but i found a strange difference between DSP version and Standalone version in Pre-emphasis work . In DSP version sound after pre-emphasis is more expressive with high`s .... in standalone version is damped .... Is there any difference about this setting in different versions (standalone , DSP , CMD , VST ) ??? I don`t think so but it`s strange ?!

That's very odd, there should be no difference - I'm using the same processing code in both places, only the front-end is different... Or are you using different sample rates etc.?

About AGC
Focussing on improvements sometimes (well, often actually) takes the focus off other things that are negatively impacted by the improvements.

I've compared the AGC of version 5.00 against the one in the BETA's, and I've noticed the following:
1- INSIDE a song, version 5.00 has a far more constant volume level than BETA. This is especially noticeable when a new track starts after some silence.
2- BETA reaches a far more constant level when comparing different tracks
3- BETA has a much better 'remaining peak' removal method

'1' is pretty important, and it's something that got lost in the BETA improvements.

New plan:
- Switch back to 5.00 type of filtering, but based on ITU 1770 levels instead of RMS levels
- But keep improved peak level removal
- During the processing, with a very slow response speed, measure the average output level, and if it's too low or too high for an extended period of time, slowly raise or lower the volume a bit (at most 2 dB, according to the measurements that Bojcha did before - and maybe less because now I'll be using ITU 1770 levels in the first step as well).

This should give much better results.


About highs
1. The limiting in 'allow louder highs' is less effective in version 5.00, which means that the level must be set a bit lower to get good results. In the currently uploaded version, the functionality of the slider has been reversed, set it between 10000 and 15000 to get similar results as in version 4.22 for loud 's' sounds (higher to get - far - less 's' sounds).

In my latest version I have fixed the slider - it can now be moved between -40000 and +40000 - but that version is not yet available.

2. In some cases, Loudness gives slightly (2% or 0.2 dB) higher levels of high frequencies in the output in version 5.00/BETA. To get similar levels to version 4.22, lower band 9 and 10 of the multiband compressor by 2%. You probably won't hear the difference... To do this anyway: 1. If not already enabled, enable Equalizer (preferable After multiband, if you're not already using it); 2. Multiply the equalizer levels of the last 2 bands by 0.98; 3. If you are using 'Before multiband', also multiply the Soft Limit values of the last 2 bands by 0.98.

This only seems to occur when the lowpass filter is used...

hmm .. I thought that I suggest something like that, but is that we already have when Smoothing is turned off ?
Yes, Lowpass rises that highs near selectend frequency by lowpass and you can easy internaly fix that by EQ!
BUT, Lowpass is before multiband, and you can see when you turn on lowpass at 15kHz, last band is rised! Maybe is better to set soft limit lower insted EQ. But again diferently seted Lowpas need different settings of last two bands (EQ or soft limit)

15kHz lowpass - should be flat


i know some processors have "steps" for lowpass/highpass -15, 15.5, 16, 16.5, 17 kHz
And for every step differently seted EQ or whatever

edit: i just checked...
For 15kHz lowpass you need to set EQ (after multiband) 9th band -0.86 and 10th -2.26 that is for LIMIT mode in Multiband
For Compress mode -0.86 and -1.91

That's what I thought too, but it's not! (I don't know why yet).

No Hans In both versions Stereotool shows 176400 output sample rate ... , in case of DSP plugin input rate is what rate have the song (44100) and with standalone version i put manually 44100 for input rate and here we go the difference is here !? And for the protocol i must say that Standalone version is take a little bit more CPU Check this source , It`s for you