New DSPX Firmware is out!

[Sgeirk]

SOURCE MATERIAL IS PARAMOUNT!

Had to respond to the earlier poster about source material.

Bob Orban, years ago wrote an article about maintaining audio quality in the broadcast plant.

Heck, it may have been written in 1975, but it still prevails.

We just went back to linear audio in our automation system. What an incredible difference. We're linear in our STL now, as well. Areas of multipath sound 100% cleaner and the blending is not nearly as noticeable.

It takes time to do it right, but blow out all your compressed audio with clean, 44.1 linear audio and you can't help but notice a big difference.

The cost of today's portable and fixed HD's is making audio codec's not necessary for storage. Transmission/sending via internet, sure...but when 800 songs fit within 40 gigs, what's the point of bit/audio destruction?

And this particularly applies to anyone that has gone HD or is about to. HD radio with compressed audio fed into it sounds slightly better than satellite, IMHO. But, HD radio with linear audio fed into it sounds pretty breathtaking.

 

[BrokeEngineer]

Again, Scott, is the DSPX-AM coded yet? We spoke awhile back about a demo and a trade type thing but you said you were super busy. Is the AM ready for prime time yet?

 

[gtomas]

Re: SOURCE MATERIAL IS PARAMOUNT!

It takes time to do it right, but blow out all your compressed audio with clean, 44.1 linear audio and you can't help but notice a big difference.

I completely agree with you. I've been preaching linear audio in broadcasting ever since mp3 became popular... But people listen to WAV and then to mp3, they don't notice the difference and think it's OK to use mp3 (or some other lossy format) in broadcast. They often forget that on-air processor has a huge influence, from disrupting frequency balance (and therefore "masks" that would otherwise cover the artifacts) to applying 14dB or 17dB high-frequency boost and then compressing it, sevearly affecting the part of spectrum where codecs exhibit most of their artifacts...

And in this digital age you additionally have a final transmission codec be it for streaming or DAB (HD Radio). So on top of your psycho acoustically reduced audio that has gone through major transformation in the processor making it sound even worse (from codec-artifacts point of view) you are now cascading another codec, and this one usually with a very low target bitrate. Common sense tells you this won't sound good and it's even more apparent in real life. Stations that use quality, linear sources sound better both on-air (FM) and even better in digital. There should be only one lossy codec in your transmission chain and it should be at the very end!

Also, people don't realize that some (actually most) codecs clip the audio internally if driven to 0dBfs, which produces additional distortion. If you're already using techniques like look-ahead limiting to keep audio clean for the codec to sound best, it doesn't make sense to then drive the codec so that it clips your carefully peak limited audio. You'll sacrifice some loudness (from my experiments -1,5dBfs is a good compromise to drive most codecs) but you'll keep the quality of your audio!


Regards,
Goran Tomas

 

[BofH]

Yes some great points there Goran about uncompressed linear audio - wherever possible insist on direct CD rips that have been extracted using some form of error correction enabled ripper, such as Exact Audio Copy in "Secure Mode". This ensures that the rip is as bit accurate as the original CD audio as possible.

Another thing I can't emphasise enough is standardising on a standard audio level for all music playout - a technique that has worked well for us is to use the WaveGain utility. http://www.rarewares.org/files/others/wavegain-1.2.7beta.zip It is fully cartchunk compatible so it won't destroy any infoediting etc (test on a few files for your own peace of mind first - also remember that although it has an undo mode, it will not be able to restore the files to their bit original format - eg it is destructive. With this in mind I recommend keeping your master music library non-wavegained and only wavegain music added to your onair machine - this way at any time you can restore the original version if you need to or if you want to try different options - my preferred one is: wavegain -r -y -w -l -d 1 -t -g +3.5 *.wav

It actually works in the opposite manner to a normalising plugin, by digitally rescaling the audio levels DOWN to a preset level instead of UP to 0 dBFS. The way it should be done considering most modern POP is already sitting on 0 dBFS and you can't get all your music library to this same level without some incredibly destructive limiting/clipping.

After running this over your music folder, you'll find your music library will sound consistent from cut to cut, even without any compression applied. This really eases the burden on your onair processor and allows your sweepers and other program elements to "cut through" over even the most hypercompressed modern pop. It also allows plenty of breathing room for the audio processor to really smooth the audio how you like.

For Internet broadcasting, you MUST connect your encoder digitially to the output of your broadcast processor and ensure its a clean feed (no clipping or pre-emphasis - multiband AGC and peak limiting only!)

I've found it's just too easy to clip codecs using analog sound cards - you can buy a cheap soundcard such as the Aopen Cobra AW-850D for under $50 - it works great and has optical and coaxial SPDIF inputs (it accepts AES fine with an XLR > RCA converter)

 

[Sgeirk]

Awesome, thanks for the tip on wavegain...does it work with audio that's already IN the system with "RIFFPCM" formatted files? (already with the song info, etc?)

Wish I would have known this before I started this process, but I'm about to dump another 350 titles in, so I'll use this on the external HD, before loading.

Thanks!!!

 

[wgliradio]

Agree 100% on non-compressed audio and use it every time we can. I wish we could be uncompressed at the High School, but getting just one PC can be a chore, so we do try for high bitrates (at least 256 kbps)ripped right from disc when using our automation.

One good thing about our automation (OTS) is that it does something very similar to WaveGain. All OTS files are referenced so peaks never surpass -6dB. This is done thru a normalizing tehcnique, so there is no signal processing used to keep the level down. This leaves more than enough headroom for the soundcard and equipment ahead of it. A Henry box gets audio in and out of the console the right way.

But there is nothing like playing uncompressed audio direct from disc.

 

[BofH]

Awesome, thanks for the tip on wavegain...does it work with audio that's already IN the system with "RIFFPCM" formatted files? (already with the song info, etc?)

It sure does, CartChunk uses RIFF PCM (WAV) files to store song information. Give it a try on some of your WAV files and see how you go. Make sure you keep the original WAV's on your external HD and only wavegain your Onair system - that way you can experiment / restore the original WAV files without having to re-rip.