Collaborate has a lot going for it and I was seriously considering it for our new online courses BUT a couple of issues really let it down!
1. Audio quality is really poor with a lot of background noise. I presume it's being compressed heavily for internet streaming but to me it's unusable. I've been trying to run my audio through skype alongside collaborate as it's much better but the downside of this is trying to manage two different systems simultaneously. This can be confusing enough on my end but when you consider all the attendees having the same set up it gets crazy!
Another problem with this of course is that recordings won't work as the audio is separate.
I've been an audio engineer for over 20 years so I know a lot about how to get good audio but if anyone has any ideas or solutions I'd really appreciate it.
2. Secondly, not as big a deal as the audio but it's really annoying when you use web browsing, switch to a white board and then you come back an the web page is gone. Wouldn't it be easy to
incorporate a way of storing/bookmarking the visited web pages?
Sorry to be so negative about what is a very good system but I really need the system to be top notch in all departments.
Replies
Elluminate/Collaborate's concentration from early on has been 'no user left behind' so the compression has been a big part to keep users on lower bandwidth able to keep up. It has trumped the need for high audio purity. As more and more users are on broadband, the goals can edge more toward increasing the sound and video stream quality.
I have been a regular participant in Adobe Connect session for Adobe e-Learning training and that would be the next closest competitor to Collaborate. I have actually left sessions because there was so large a percentage of sound drop outs that comprehension was hampered. The design in Connect seems to swing towrd dropping instead of compressing when a user starts to lag behind on connection speed. It has been a while since I have been in other web conferencing system meetings, so those may have improved, but several of our volunteers came to us (VHSG) because the systems they had tried had such audio issues. I am thinking that if you need visual elements to go with the audio, Collaborate will still best what is out there. If you don't need visuals, Skype may be a good way to go if you are finding the quality to be better.
Now for some suggestions on how to get the audio quality as high as you can. We had a music history course two years ago, so we have passed through some of what you are facing right now. We found a variety of ways to work with the music. Loading up music into the media library was the way we worked with pieces that needed the highest sound refinement. You can get CD quality if you make the file CD quality. The participants are essentially getting the digital file and playing it with a player on their own system. This avoids all compression. There are caveats. Sound files have a per file mb limit of 20 mb and the entire library cannot hold more than 100 mb simultaneously. It isn't too hard to eliminate one that has been played that isn't needed any longer and load in another while in session though.
Another option that works quite well is to pair your live presentation with the webtour to send them to a website with the media embedded there. This is what you see in the Classroom 2.0 pre-show. The participants see a webpage with streaming music in a player. They can listen while they wait for the show to begin. The quality is identical to what you would get when you visit the website outside of the live session.
Even though there is a lot of compression when sounds come through the mic, shorter sounds work very nicely through that method. Sound effects, for instance, add some nice spice in to the live voice presentation and the compression is generally not noticed by participants.Good examples are crickets when you get an awkward stretch where you ask for participant input and no one responds. It gives everyone a chuckle and it works pretty much every time to get people typing and talking. The pulsing, rather than melodic nature of the sound works well even with the compression. Other sound effects that are fun and work well even wit the compression are vintage effects like the old 8mm newsreel openings. Everyone expects those to sound old and noisy and so the effect works. It works well into your presentation when you are talking about how things used to be done. You can event make your visuals all old-school newsreel looking. If you have a hard time finding that sound effect and want it, let me know and I can send you a copy of my sound effect file.
OK, I officially am starting the Tammy Moore fan club.
Hi Steve,
The mic is definitely not the issue as I have access to top quality studio recording mics and external preamps. I even converted the signal to digital and fed it to the system in digital format to see if it was an analog to digital issue but it's not.
This is such a pity. I've tried every system I can get my hands on and definitely collaborate is the winner but because of the type of tuition I offer, I simply can't have audio that isn't of good quality. Sure it's fine for basic communication but not for what I need.
As regards the web tour, Yes, bookmarks would be great!
Hello Will,
it is very interesting for me to read more about what you discovered around Collaborate´s audio artifacts by using studio signal chain setups. Stability of total syllabic perceptability was one of the strongest selling points of Elluminate. The "no one left behind" approach meant "No spoken word is lossed". Situation changed as the Gemini developers did touch the internal audio processing / codec parameters somehow. Dealing with audio drivers, the situation may be different for winXP, winV7, MacOS or Ubuntu. Some drivers to provide filter and echo cancellation on low level (to be switched off by Java commands IMHO). What systems did you use in the studio?
Here are my investigational quests I want to share with you and other AVengineers reading this (sorry to everyone reading questionmarks only, it´s AVnglish ;-):
Hypothesis 1: Fishbowling artifacts are caused by higher frequencies than speech. I assume that there is no steeply sloping low-pass filter in the digital chain, or they do not match it to the sample rate they made adjustable. Therefore, the suppression needs to be sloped by preprocessor.
Your result: As you fed in digital format, did you try to limit the analog input at e.g. ~ 7 kHz for a 22kHz sample?
Hypothesis 2: Fishbowling comes from non-constant background, say reverb and echos. I assume that the codec uses compression or bad shaped AEC ) with pumping attach/decay/sustain parameters. Codec simply identifies background as a wanted voice signal to be emphasised. In other words: They prolonged some critical time constants in dynamical amplification as they did step from v10 to v11.
Your result:
Did you switch off the internal dynamic of the codec as I wrote in the end user thread on external audio ?
Did you try to noisegate ? Keying-out by background-only microphone ?
Did you hear any pumping effects in the Collaborate records you´ve done with your external signal chain?
Not shure what you mean here. In any "simulcast" gateway setup (Collaborate+Skype, Collaborate+Adobe,Collaborate+Videoconference), I have to use a PC per channel, which is able to record that channel.How do you share audio between two conference applications on the same machine using external hardware signal processing? Might be too tricky to hand over to the average participant.
Did you try to insert external audio through a telephone hybrid into a conference provider being bridged to Collaborate by telefone bridge ? This might be more straightforward than Skype (for most of my German clients, Skype is evil, but SIP is acceptable as this is standardized VoiceOverIP crossing firewall borders in a controlled environment)
Will
unfortunately, most of my clients here in Germany decided to step back to Elluminate v10 because they did buy a multilingual (say German) solution, which was switched to English-only all of a sudden. So my soundchecks happen at own events only. In my lab, I´m working on a mixed V10, V11, v12ß workbench to overcome migration issues in the next months. My goal is to contribute to a best-of v12 recovering v10´s USP strengths because I really love the total feature set of the engine (like you do I guess).
Any help is very appreciated.
Mero
Just to make sure--you have tried a standard headset mic, and so it's not some fluke of your more expert equipment actually giving a worse result...? I'm guessing your standards must be really high, as while I would have said Skype is superior to Collaborate, I would not have said it was out-of-the-world better. What kind of sessions are you holding?
Hi, Will! My two cents:
1. My experiences with Audio on Collaborate (previously Elluminate) have been that the audio is not at Skype level, but has always been adequate to very good... I'm wondering if there is another issue. The headset or microphone that I use does seems to make a big difference. I use a $25 Logitech USB headset on my PC, and my sound is usually quite clear.
2. The switch between whiteboard and web tour used to work the way you have described, and sometimes I miss that, but other times I like the simplicity of the new model. I'm with you on the bookmarking idea! I'd love a dropdown to show the sites/pages I've recently visited so I don't have to type it in again.
You don't need to apologize for asking good questions! Let's see if we can help with #1 and I'll try to make sure #2's suggestion gets to the right people. :)