Topic: Outputting from JRiver to the Spectrum Lab program (Read 19202 times)

mojave · « **on:** November 29, 2012, 10:24:42 am »

There are threads at AVS and elsewhere in which users are outputting the audio from their soundcard, routing it back to the input, and measuring the maximum and average output levels of the bass in a movie using the Spectrum Lab program. This seems a rather archaic method and going directly from the media program seems like a better method.

Spectrum Lab can accept input via WM_COPYDATA messages or TCP. It would be a neat feature if JRiver could output one of these ways and could get more people interested in JRiver. There was a WinAmp plugin that sent the data to Spectrum Lab, but I can't get it to work with JRiver.

There is a thread at the Data-Bass.com forum that is creating a nice collection of movies based on bass output. Here is an example:

The Incredible Hulk
Level - 5 Stars (269.4dBHz)
Extension - 5 Stars (2Hz)
Dynamics - 4 Stars (25.6dB)
Execution - 5 Stars - Still the track I hold all others to. It has everything.

Overall Rating: 4.75 Stars

HuskerOmaha · « **Reply #1 on:** November 30, 2012, 11:12:44 am »

I am also interested in having this functionality.

Matt · « **Reply #2 on:** November 30, 2012, 11:38:46 am »

Is the Spectrum Lab author interested at all?

Maybe they'd be willing to help write the output plugin code?

mojave · « **Reply #3 on:** November 30, 2012, 12:50:57 pm »

I'll see if I can get in contact.

The WinAmp plugin I linked to earlier explains a little how it is done.

mojave · « **Reply #4 on:** November 30, 2012, 12:54:53 pm »

Here is the info on Spectrum Lab's website on how to communicate with it:

http://www.qsl.net/dl4yhf/yhf_comm/yhf_comm_info.htm

mojave · « **Reply #5 on:** June 12, 2013, 03:38:26 pm »

Now that JRiver can convert video to audio I can just convert Particles and analyze the wav file with Spectrum Lab! Spectrum Lab only accepts stereo or mono wav files. Using the DSP to provide mono output works fine and combines all channels properly. However, if I want to analyze just the LFE channel or other channels there is no way to do so.

I guess I could make a multi-channel wav with all channels muted except the one I want. Then I could re-convert that to a mono wav. I want to see the original levels, though. If I wanted to see the front right channel, for example, and the others were muted then the front right would still be lowered by 10 dB when mixed with the muted LFE channel. Would I just add 10 dB back to that channel?

mojave · « **Reply #6 on:** June 19, 2013, 02:55:45 pm »

JRiver is clipping when converting to a mono wav file probably due to the levels of all combined channels. Is there a way to prevent the clipping since the conversion to mono takes place before any other DSP?

Matt · « **Reply #7 on:** June 19, 2013, 03:15:08 pm »

It's not a clip until it's output as something that can't hold it (normally an integer format at the end of the chain).

You could apply a volume reduction later in the chain.

JRSS mixing doesn't avoid all possibilities of clips. There was a lot of discussion about this a couple years ago, and I think you helped us derive the current solution that compromises volume vs clip-resistance.

6233638 · « **Reply #8 on:** June 19, 2013, 03:28:08 pm »

Quote from: Matt on June 19, 2013, 03:15:08 pm

JRSS mixing doesn't avoid all possibilities of clips. There was a lot of discussion about this a couple years ago, and I think you helped us derive the current solution that compromises volume vs clip-resistance.

I wish it was an option.

mojave · « **Reply #9 on:** June 19, 2013, 04:02:05 pm »

Thanks, Matt. I didn't realize the clipping only happened at the end of the chain. I reduced the level by 3 dB in Equalizer and that took care of the issue.

mojave · « **Reply #10 on:** June 20, 2013, 07:39:27 am »

Which DSP volume adjustments are relative and which are absolute? What I need is for the LFE to be attenuated by 10.2 dB and the other channels by 20.2 dB before being mixed. I noticed that when using the Parametric Equalizer that volume adjustments are just relative and I end up with 10 dB of attenuation. If I use Mix Channels and attenuate a channel is it relative to Adjust the Volume?

Does mixing to mono always behave the same way or is it dynamic based on content? I'm trying to establish a repeatable process which exactly matches what has been done by others using other software.

If I use a combination of Equalizer and Mix Channels I get what I want on the subwoofer channel. If I set Output Format to mono with JRSS off, what channel does the subwoofer need to be copied to? I can't seem to get it to work. I just get an empty mono channel.

mojave · « **Reply #11 on:** June 21, 2013, 09:31:14 am »

Quote

I'm trying to establish a repeatable process which exactly matches what has been done by others using other software.

If mixing to mono in JRiver always does the exact same thing regarding levels then I think I'm good to go. However, it probably differs for 5.1 vs 7.1. I found that using -7.1 dB in the Equalizer preamp (the dB level isn't very granular) combined with JRSS is close.

What is currently being done with other software is that all the channels are reduced by 20.2 dB except the LFE which is reduced by 10.2 dB. The channels are then all mixed to a mono wav file. This is done with both 5.1 and 7.1 audio. This prevents any chance of clipping with a 7.1 source.

I've been asked to write up a tutorial on AVS on how to use JRiver to create wav files from movies for analysis in Spectrum Lab for The New Master List of BASS in Movies with Frequency Charts thread. The thread already has 578,604 views. However, I need to be able to match what has already been done and make sure it is consistent.

Matt · « **Reply #12 on:** June 21, 2013, 09:56:05 am »

Do you have to have a mono file? Or can Spectrum Lab use just a specific channel of a multi-channel WAV file?

mojave · « **Reply #13 on:** June 21, 2013, 10:13:59 am »

Spectrum lab only accepts mono or stereo wav files. It doesn't accept any multi-channel files.

This is why I was trying to use mono output without JRSS as mentioned in the latest build thread. I thought if I could use Mix Channels to combine everything to the front left it would come out when switching to mono.

mojave · « **Reply #14 on:** June 21, 2013, 10:32:34 am »

I just tried something else that I thought for sure would work . . . but it doesn't.

Equalizer preamp set to -10.1 dB
Room correction subwoofer set to +10.1 dB. This should push all other channels down by another 10.1 dB, right?
Parametric EQ set to add all channels to the subwoofer.

I left as multi-channel and then opened in Audacity. Audacity can save just a single channel of a multi-channel file as 24-bit wav.

There is still clipping when opening in Audacity. JRiver is too smart and doesn't honor the double volume change.

I tried using some VST plugins to lower levels, but convert format doesn't seem to honor any VST settings in the DSP.

mojave · « **Reply #15 on:** June 21, 2013, 10:41:12 am »

I put the link to the thread that charts the bass at data-bass.com in the first post, but it may have been overlooked by those following this thread. Here it is again:

The Low Frequency Content Thread (films, games, music, etc)

6233638 · « **Reply #16 on:** June 21, 2013, 10:56:24 am »

Do you actually need downmixed audio for this, or just the LFE channel? It seems like it should be easy to output LFE unmodified?

(I still wish we had an option to use a downmix that won't allow clipping to happen though)

Matt · « **Reply #17 on:** June 21, 2013, 11:07:45 am »

Looking at this uncovered a problem. We were adding mixing matrix components with addition instead of doing the more complex (but correct) incoherent volume addition when converting down to mono.

Next build:
Fixed: JRSS downmixing was not always energy neutral when downmixing to mono.

mojave · « **Reply #18 on:** June 21, 2013, 11:17:48 am »

All channels need to be included since there is a lot of bass in the other channels. For example, the subway crash in Skyfall has incredible bass output from 80 Hz down to 15 Hz. All channels, including surrounds, contain info in this region that combine to create the effect.

Matt · « **Reply #19 on:** June 21, 2013, 11:26:41 am »

Once you have the updated build, when you use JRSS to play 5.1 as 1.0, it will use -11.76dB for the main channels and -1.76 for the LFE.

So add a Parametric Equalizer that applies volume to get the remaing ~9dB and you're all done.

Does that work?

mojave · « **Reply #20 on:** June 21, 2013, 11:51:24 am »

Quote from: Matt on June 21, 2013, 11:26:41 am

Once you have the updated build, when you use JRSS to play 5.1 as 1.0, it will use -11.76dB for the main channels and -1.76 for the LFE.

So add a Parametric Equalizer that applies volume to get the remaing ~9dB and you're all done.

Does that work?

That should work perfectly! Thanks for the exact numbers. What about 7.1 as 1.0? Sorry for the trouble.

Matt · « **Reply #21 on:** June 21, 2013, 12:00:11 pm »

7.1 to 1.0 will use:
-2.3045 dB for the LFE
-12.3045 dB for the other speakers

mojave · « **Reply #22 on:** June 24, 2013, 04:22:53 pm »

I found it ironic on Friday evening that I had spent part of the day trying to make downmixed mono file and ended up watching a movie with the family that has a mono soundtrack (The Wizard of Oz).

Matt · « **Reply #23 on:** June 24, 2013, 05:54:06 pm »

Quote from: mojave on June 24, 2013, 04:22:53 pm

I found it ironic on Friday evening that I had spent part of the day trying to make downmixed mono file and ended up watching a movie with the family that has a mono soundtrack (The Wizard of Oz).

Or is that the opposite of ironic?

We could ask Alanis Morissette, who defines irony as dying in a plane crash.

mojave · « **Reply #24 on:** June 25, 2013, 08:48:52 am »

It is Ironic, Isn't It?

According to the above link, Alanis had the following in her song

:

Attempts at describing irony: 11

Successful attempts: 2.5

Confusion of irony with "poor sense of timing": 6

Completely missing the term: 3

mojave · « **Reply #25 on:** June 26, 2013, 11:01:12 am »

I have created a guide at AVSForum for creating mono files with JRiver that can be analyzed with Spectrum Lab. I will also post the guide here.

INTERACT FORUM

Author Topic: Outputting from JRiver to the Spectrum Lab program (Read 19202 times)