Freeware Downmixing program?

[DragonQ]

I need a program to downmix 6 WAV files (24-Bit, 96kHz) to a stereo WAV file. I'm using Audacity to play about with the files, but it has no proper downmixing module. You can tell it which channels go where, and the volume of each channel in terms of dB (although not exactly, only to the nearest mouse movement), whereas I need a program that will allow me to input downmixing values (such as the normalised values on this website http://ac3filter.net/guides/mixing_matrix).

I tried WaveWizard, but it doesn't work properly. I give it all the normalised values and then turn the program's normalise function off, and it results in a stereo WAV that is way too quiet O_o. Anyone have any suggestions?

[SolarWind]

Cool Edit's (Audobe Audition) "Mix Down to File" function works quite well for this purpose. (The Cool Edit 2.1 demo that will let you use all features can still be found in the shareware archives on the net):

Multitrack View: Edit > Mix Down to File

This function creates a two-track stereo mix of all tracks and places the resulting waveform into Cool Edit Pro’s Edit View. The contents of all or selected enabled tracks are combined, with track and waveform properties (such as volume and pan) affecting the way the final mix sounds.

After you've loaded 6 individual mono channels in the multitrack view, you need to adjust "Volume" and "Pan" factor for every input channel according to your mix preferences (that would be the "mixing matrix") before applying "Mix Down to File". A typical 5.1 to streo downmix stategy (values to start with) would be something like this:

Volume / Pan
Track1_LF -6.0dB / Left 100%
Track2_RF -6.0dB / Right 100%
Track3_C -9.0dB / 0
Track4_LFE -15.0dB / 0
Track5_LS -12.0dB / Left 100%
Track6_RS -12.0dB / Right 100%

Hope this helps. See also this thread (http://www.hydrogenaudio.org/forums/index.php?showtopic=55442) for more info on 5.1 -> stereo downmix coefficients.

[DragonQ]

Is there any way to find out the downmix coefficients that are stored on the DVD-V or DVD-A (that standalone players should use)?

Also, thanks for those values. I already have that program but I didn't know what the coefficients were in dB rather than percentages.

[SolarWind]

Sorry, don't know how to view the emedded downmix coeficients.
I only know that the "recommended" downmix coefficients are indeed part of DVD-Audio (.mlp) and DVD-V Dolby Digital (.ac3) multichannel streams, so there's got to be a way to see them, too. You could try the DVD-Audio Explorer / SurCode MLP encoder/decoder or similar tools. Other forum members may be able to provide more help.

DVD-Audio includes specialized downmixing features for PCM channels. Unlike DVD-Video, where the decoder determines how to mix from 6 channels down to 2, DVD-Audio includes coefficient tables to control mixdown and avoid volume buildup from channel aggregation. Up to 16 tables can be defined by each Audio Title Set (album), and each track can be identified with a table. Coefficients range from 0dB to 60dB. This feature goes by the horribly contrived name of SMART (system-managed audio resource technique). Dolby Digital, supported in both DVD-Audio and DVD-Video, also includes downmixing information that can be set at encode time.

[Stefan]

If you turn the normalize function "off" of course it's not going to work! Besides, normalization is very dependent on the source material. All it does is adjust the file so that the waudio waveform peaks hit the target level. If the waveforms are basically at a low level but with a few big peaks, it's not going to sound very loud. A better approach would be Replaygain, which measure the RMS or average level and applies scientifically designed loudness curves. If you have Adobe Audition (or Cool Edit Pro), you could also try the Group Waveform Normalize function using the Equal Loudness values. It's similar to Replaygain.

This will give you a much more effective management of audio levels. Then try AC3filter or other programs.

[DragonQ]

But surely it doesn't need normalising if I use the correct coefficients? And when I say it's quiet I don't mean there are a few peaks but the rest is quiet, I mean the whole track could be increased by about 6-9dB without clipping. I don't understand why this would happen if I'm giving it sensible coefficients in the first place. Here is a screenshot:

http://dragonq.pwp.blueyonder.co.uk/Quiet Waveform.PNG


I don't want to lose resolution or anything by taking the low level signal and boosting it, I just want to get it so that when I downmix, all channels are at levels that make it peak at or near 100% (0dB) and sounds like the stereo mix (in terms of loudness of different parts). I do have Adobe Audition 2.0, but it's just so complicated to use (or it just seems so daunting I can't figure out what things do - simply mixing two tracks together is harder and more complicated than Audacity) - it seems to have its own downmixing thing, but it doesn't have any levels to choose from. It looks like I can do it manually if I put all the channels into the multitrack, set the panning and volume manually, but I didn't know what volumes to use in dB (only %), which was why I couldn't do it. I can try it now when I have some time thanks to SolarWind's suggestion.

[DragonQ]

OK, I just did that...in Audacity the wave looks fine, pretty loud, but no clipping. Sounds pretty good too. I save it as a 24-Bit, 96kHz WAV file and open it in Audacity...and it appears quiet again! What could be going on here??

[DragonQ]

Crap, I meant in Adobe Audition (the program I used to make the downmix) the wave looks fine (volume-wise), but when I then open it in Audacity it looks quiet like the screenshot.

[dmckean]

I think foobar 2000 has a built-in plugin that downmixes 6 channel audio.

[SolarWind]

OK, I just did that...in Audacity the wave looks fine, pretty loud, but no clipping. Sounds pretty good too. I save it as a 24-Bit, 96kHz WAV file and open it in Audacity...and it appears quiet again! What could be going on here??
You're likely saving 24-bit audio to a WAV sub-type Audacity cannot handle properly.
Audacity is likely fooled into beleiving that your file has 32-bit dynamic range, wheras your real dynamic range is still 24-bit of course. Since your zero dB mark is then 32-bit in Audacity (which is incorrect) the WAV is being displayed scaled after it loads. Note: The saved WAV material is still correct, it is just a "feature" of Audacity to scale the vertical axis range to 32-bit. This is my take on what may be going on.

Try this: When you finished your 5.1->stereo mixdown in Adobe Audition and you click on "File"->"Save As...". The "Save Waveform As" dialog pops up. Now, before you push "Save", click on "Options..." button in it.

Save to "24-bit packed int (type 1, 24-bit)" rather than "32-bit Normalized Float (type 3) - Default".
After this even Audacity should be able to recognize that the file's dynamic range is 24-bit.

32-bit int (type 1 - 32-bit)
This format saves 32-bit audio as 32-bit integers.

32-bit 24.0 float (type 1 - 24-bit)
Full 32-bit floats are actually saved (and in the range of +/-8million), but the .wav BitsPerSample field is set to 24 while BlockAlign is still set to 4 bytes per channel.

32-bit 16.8 float (type 1 - 32-bit)
Cool Edit Pro’s internal format (floating point values in the range of +/-32768.0 but larger and smaller values are valid and not clipped since the floating point exponent is saved as well). The .wav BitsPerSample field is set to 32 and BlockAlign to 4 bytes per channel.

32-bit 0.24 normalized float (type 3 - 32-bit)
Standard floating point format for type 3 .wav files. Values are normalized to the range +/-1.0 and although values above and below this range are saved, some programs may clip when reading them back in (Cool Edit will not clip, but read the save value back if it is beyond this range).

24-bit packed int (type 1 - 24-bit)
Straight 24-bit integers are saved (so any data beyond the bounds are clipped). The .wav BitsPerSample field is set to 24 and BlockAlign to 3 bytes per channel.

24-bit packed int (type 1 - 20-bit)
Straight 24-bit integers are saved (so any data beyond the bounds are clipped). The .wav BitsPerSample field is set to 20 and BlockAlign to 3 bytes per channel. The extra 4 bits that are saved are actually the remaining valid bits when saving, and are used when reading (thus giving 24-bit accuracy still if those bits were actually when writing). Applications should either fill those last 4 bits with zeros, or with actual data, but generally A/D converters that generate 20 bits of valid data automatically set the remaining 4 bits to zero. Any type 1 format with BlockAlign set to 3 bytes per channel is assumed to be packed integers, and a BitsPerSample field between 17 and 24 inclusively will read in all 24 bits fine and assume the remaining bits are either accurate or set to zero.