• Get involved.
    We want your input!
    Apply for Membership and join the conversations about everything related to broadcasting.

    After we receive your registration, a moderator will review it. After your registration is approved, you will be permitted to post.
    If you use a disposable or false email address, your registration will be rejected.

    After your membership is approved, please take a minute to tell us a little bit about yourself.
    https://www.radiodiscussions.com/forums/introduce-yourself.1088/

    Thanks in advance and have fun!
    RadioDiscussions Administrators

soundprocessing on livestreams in the US (NYC) some good, some very bad...

E

erwin33

I wonder why stations don't sound so good on there livestreams, for example in NYC. WPLJ's livestream sounded very, very bad soft, less or no soundprocessing on there stream. Other stations in NYC sounding great for example all CBS NYC stations and also the Clear Channel stations sounding good since a couple of months because they struggle a long time with bad processing on there streams (Z100, Lite and KTU).
Why take some stations it not seriously as they do with there off air processing? In these days where internet radio becomes much important, they must take it seriously. Some do, some don't.
 
I wish that streams were 100% simulcasts of what goes over the air, including commercials and EAS tests.
 
Too much to do and not enough time to do it would probably at least one pat of it.

Also, real processing people are few and far between.

Also remember that the while web sites may bring some revenue streaming itself is considerably harder to profit from.
 
Perceptually coded audio needs to be processed very gently and carefully. The reason is that you don't want level build up at a frequency that the encoder deems should go away.

Remember-the perceptual encoding process used by the coder is designed to work with UNPROCESSED audio!

I recommend a VERY SLOW gentle AGC (basically a 'jock watcher') , coupled to a SLOW multiband processor (six bands maximum), coupled to a look ahead peak limiter that's HARDLY EVER USED-it's only purpose is to prevent digital clipping.

People have also had very good results with a compellor or CRL SPP-800 as the only processing (with the output turned down a few db to accomodate peaks).
 
Just a testimonial...Being in portuguese and living in Portugal, with free time to listen to streams on the station where I work, after listening dozens of US stations, I got stucked at KVYN, for the music and for the processing (almost none). Having nice speakers and reasonable acoustics at he studios... sounds quite nice...
 
Again with KVYN, the bitrate is low and so is the audio quality... I would rate the audio quality as "I'll put up with it if the content really, really interest me". But otherwise I don't see lots of people listening for long. It seems the majority of US broadcasters are looking at web streaming as a side-thing to their on-air broadcast, almost like a feature to have. Not as a service that should capture _and keep_ the listener. More like, look we have a web stream, check it out but we don't expect you to actually listen to it...

On the topic of stream processing, I'll agree that it's better to have light processing than heavy processing on the web stream. I actually did a blind listening test with a couple of dozens university students, where I increased the aggressiveness of audio processing progressively (using Orban PC-1100), encoded it to aacPlus and then people rated the audio quality of each sample. The results were that with increased aggressiveness of dynamics processing, the artifacts get progressively worse. IOW, light processing produces less artifacts so the audio qulity is better. I don't have that paper at hand, but I'll try to post it later.

Apart from the overall density of processing and in particular the use of final look-ahead limiters, I would say the second most important thing to carefully tailor in a web stream processing (and any digital broadcast for that matter) is the high end. The codec has the most trouble with the high-end and that's where the majority of codec artifacts are produced. You want a very controlled, if not a bit tamed down high-end to minimize the coding artifacts. Putting lots of high end to a codec spells disaster in the artifacts department!

There are other things to take care of, the output level (codecs produce overshoots, in my tests up to 3dB), controlling the width of the stereo image, source audio quality, etc. If you want your web stream or digital broadcast to sound good, you need to take it seriously and pay attention to the details.


Regards,
Goran Tomas
 
It's VERY easy for someone to criticize a station's audio stream when you're not the one writing the checks to keep it running.

In case you've been in a cave for the last eighteen months or so - ad revenues for all media are WAY down. Stations are having a hard enough time selling out avails on the terrestrial side, let alone selling internet-only ads for the webstream that for the most part is listened to by people who will never have a PPM meter for the originating market.

In my opinion many US broadcasters are doing an OUTSTANDING job with their streams considering the shoestring budget they're working under. Sorry they're not up to the standards demanded by the world - but for the most part they're not targeted to please people outside of their home markets.
 
Goran Tomas said:
On the topic of stream processing, I'll agree that it's better to have light processing than heavy processing on the web stream. I actually did a blind listening test with a couple of dozens university students, where I increased the aggressiveness of audio processing progressively (using Orban PC-1100), encoded it to aacPlus and then people rated the audio quality of each sample. The results were that with increased aggressiveness of dynamics processing, the artifacts get progressively worse. IOW, light processing produces less artifacts so the audio qulity is better. I don't have that paper at hand, but I'll try to post it later.

Apart from the overall density of processing and in particular the use of final look-ahead limiters, I would say the second most important thing to carefully tailor in a web stream processing (and any digital broadcast for that matter) is the high end. The codec has the most trouble with the high-end and that's where the majority of codec artifacts are produced. You want a very controlled, if not a bit tamed down high-end to minimize the coding artifacts. Putting lots of high end to a codec spells disaster in the artifacts department!

Regards,
Goran Tomas
Click to expand...

Perhaps we should use a peak limiter with a 25 us curve in it's control loop.
 
My reply on the audio quality of web streams was not directed to US stations in particular (it was the original poster that put US stations in the name of thread) but to any station broadcasting from anywhere in the world which is web streaming with low bitrate and low audio quality. No, I haven't been in the cave. Apart from the stations that I worked for and set-up web streaming (one of them is non-comm and non-profit and still has a 128 kbps AAC stream) there are many stations on the web (US, European and others) that are streaming high bitrate web streams.

It all boils down to what the particular station wants to do with its web streams, how does it perceive it and how much does it want to invest in grabbing the on-line audience and whether it sees a future there. Regardless of where it's placed on the globe. I'm not saying that all those web streams are profitable - most are not. But some stations are accepting that they may not initially make money on their web streams, but still consider them as an investment in the future and as such they want to maintain the audio quality on their web broadcast as they would on the air. Others don't and have lower audio quality. Forgive us for noticing that. There will always be pioneers and followers as well as better and worse sounding radio stations...


Regards,
Goran Tomas
 
LA_Guy said:
Perhaps we should use a peak limiter with a 25 us curve in it's control loop.
Click to expand...

Personally, I'm more inclined to comprehensive audio processing that provides all the benefits of on-air broadcast (signature sound, level and loudness consistency, etc) with the control and shaping of audio for perceptual coding (spectral balance and its consistency, stereo width control, peak level control with low distortion, bandwidth control, audio density control, etc).

There are many processing manufacturers that produce dynamics processors for web streaming and digital broadcast. From what is available on the market today, I personally prefer the Orban PC-1101 as the best choice, processing quality wise, feature wise and price wise.


Regards,
Goran Tomas
 
Goran Tomas said:
My reply on the audio quality of web streams was not directed to US stations in particular (it was the original poster that put US stations in the name of thread) but to any station broadcasting from anywhere in the world which is web streaming with low bitrate and low audio quality. No, I haven't been in the cave. Apart from the stations that I worked for and set-up web streaming (one of them is non-comm and non-profit and still has a 128 kbps AAC stream) there are many stations on the web (US, European and others) that are streaming high bitrate web streams.

It all boils down to what the particular station wants to do with its web streams, how does it perceive it and how much does it want to invest in grabbing the on-line audience and whether it sees a future there. Regardless of where it's placed on the globe. I'm not saying that all those web streams are profitable - most are not. But some stations are accepting that they may not initially make money on their web streams, but still consider them as an investment in the future and as such they want to maintain the audio quality on their web broadcast as they would on the air. Others don't and have lower audio quality. Forgive us for noticing that. There will always be pioneers and followers as well as better and worse sounding radio stations...


Regards,
Goran Tomas
Click to expand...

Goran,

Your reply just confirmed my first sentence in the previous post. Doesn't matter if it's a US broadcast station or an internet only outlet - you can say whatever you want about what an enterprise does with its stream when you aren't the one paying the bills.

The perspective changes dramatically when it's your money on the line. And this notion comes from personal experience BTW.
 
All of the stations that I know that use the Orban card just duck way too much.

For the net, I recommend either Breakaway Broadcast or Live...I like using one of their aggressive presets...but back down the setings. Phenomenal sound.
 
Are you folks suggesting that the original stream to the re-broadcasting server farm be uncompressed WAV files? (Assuming one does not have a server farm & big pipe in house.) I've always worried about an MP3 stream being further tweaked by encoding, but that's how many folks do it. At Cox we played MP2s--and that music was then encoded. It was processed "going in," by standard radio rackgear.
 
Rob Stutson said:
The perspective changes dramatically when it's your money on the line. And this notion comes from personal experience BTW.
Click to expand...

I guess you missed the part where I said I worked for radio stations and set up their web streams. As an engineer I had to elaborate to the management why the costs involved and they had to approve it. It wasn't directly my money, but most of it did come from an engineering budget.


Regards,
Goran Tomas
 
CaptainShowbiz said:
Are you folks suggesting that the original stream to the re-broadcasting server farm be uncompressed WAV files? (Assuming one does not have a server farm & big pipe in house.) I've always worried about an MP3 stream being further tweaked by encoding, but that's how many folks do it. At Cox we played MP2s--and that music was then encoded. It was processed "going in," by standard radio rackgear.
Click to expand...

The best audio quality will be achieved if the web stream (or any other digital broadcast) is encoded only once, at the very end. So - uncompressed from source, over any links all the way to the transmission encoder.

If you are coding a stream that your CDN/ISP is only distributing, IOW you encoded a stream that you are sending directly to the streaming server at your CDN/ISP, then you are fine. Provided everything before your encoder is uncompressed.

If however, you were encoding audio that was decoded at the CDN/ISP and re-encoded to another format and/or bitrate, than your audio quality will go down. How much depends on types of codecs and bitrates used. This is called codec cascading and to preserve audio quality must be avoided.

If your source audio files (your music library) are not linear, IOW it's already uncompressed in MP3 or MP2 format, that adds yet another layer of encoding/decoding which further degrades your audio quality.

Again, audio for digital broadcast should be encoded only once, at the very end of the chain (by the transmission/stream encoder).


Regards,
Goran Tomas
 
Muchly agreed! The final streaming product will sound best when the original source material originates as linear, uncompressed audio. It makes a big difference with aacplus streams.
 
Kirk said:
Check out the streams from www.goom.com. In the US, this URL is redirected to www.goomradio.us.

These streams sound very good.

Kirk Harnack
Click to expand...


I dunno, Kirk- I built a AAA station and wife came into the room to listen to it (she likes AAA). After about 30 seconds she asked me to turn the sound off. I asked her why and she said the sound was 'pushing ' itself at her. Further listening shows that the audio IS quite aggressive-it does try to 'push' itself at the listener. Maybe that's what they want-but it could negatively affect women a lot.

Oh and their site caused my browser (IE8) to hang/crash while setting up the station-twice. The third time worked fine however.

I present an alternative form of processing for you all to listen to. It's loud but not agressive at all. It follows my beliefs stated above.

You need to paste these into Winamp or Itunes. They will also play with Windows Media Player if you've installed the Orban AAC+ plug in.

http://75.125.94.68:8000 is a 128 kbps MP3 stream

http://75.125.94.68:8010 is a 160 kbps MP3 stream

http://75.125.94.68:8020 is a 64 kbps AAC+ stream

http://75.125.94.68:8030 is a 32 kbps AAC+ joint stereo stream

http://75.125.94.68:8040 is a 16 kbps 22 kHz AAC+ MONO stream
 
Status
This thread has been closed due to inactivity. You can create a new thread to discuss this topic.


Back
Top Bottom