| |||||||||||
|
These are the old trial logs for principally OAS Playout V1.x - for the latest see the main Trials page. These are the results of my efforts to put together a 1 hour pseudo-broadcast using the kit I have here and OAS Playout - principally to see how it performs in general use and to identify areas for improvement. EnvironmentSystem is a P266/128MB running Windows 98SE. All the audio is located on a seperate machine linked via 10MB ethernet. The machine has 2 sound cards, a Sound Blaster Live and an old ISA ESS Audiodrive module.
The whole lot is in a spare bedroom. Outputs from each soundcard are fed into a mixer along with Mic, CD and tape. Mixed output is fed to amp and minidisc for recording. The mixer is a home made job (circa '93) - amazingly enough still going strong - 3 stereo channels, 2 mono and 2 mic channels.
Trial logsMost recent first.... 4. OAS Playout v2.0, Alpha 1 (delta), 24/07/02First trial using the new (mine) audio core to replace the xaudio DLL, this was a pre-beta release. 4. ~72 minutes The software in use is essentially 'v2.0, Alpha 1' but with a couple of minor bug fixes applied. proved as reliable as with the xaudio variant, still some minor audio 'glitches' to iron out but no crashes. Some cross fade occuring between channels on the mixer but that's neither here nor there. comments..... went very well considering this version has a new audio core which tries as much as possible to mimic the xaudio functionality. The codec is audiably different but much closer to the Winamp sound which is pretty much deemed the de-facto standard now. Many of the minor problems occurring seem markedly similar to one's I'd seen with the xaudio dll, notably: - incurring small bursts of audio from a previous jingle/snippet when starting a new one - random delays of .5-1s when hitting 'start' to the jingle playing Both these I'm hopeful will disappear when the Player starts making use of the 'buffer preload' capability of the new audio core, something the xaudio DLL couldn't do. I don't plan to put this in just yet, for now it looks like the 'audio bursts' are to do with the buffer not being cleared down sufficiently, a quick mod to clear the whole lot seems to have definately cleaned up the odd snippet being heard so at least it should be in the same state as the v1 package. Only other thing of note is in the volume reduction level applied when the jingle player fades the main one - still not overly happy with my % vol level to db vol level logic but this is very clear cut to fix. 3. OAS Playout v1.0, Beta 3, 27/01/02~70 minutes fairly smooth, a few level problems, couple of vocal crashes - again operator incompetence. comments.... no functional problems, s/w behaved fine surviving even a database import mid way through (1 track added to music store). found i tended to have 1 or 2 tracks cued at any one time on both players. generally doing back2back on 1 player but for jingle inserts switching to the other player for the jingle then playing the track on that player till the next jingle. for the sequences where i played snippets over music, made use of the auto-fade facility of the jingle player - it's a lot smoother now particularly the volume restore of main player - this is where u notice the 'jingle start' delay tho - main track fades there is a pause before the clip starts. now all the audio is held as mp3/44hz/stereo we appear to have abolished all the 'odd' effects seen earlier - "wrap round" and clipping etc. the "preloading" of jingle still clips the first word of the occassional sound bite - noticeable with the "clousea" clips. jrb, 27/01/02 2. OAS Playout v1.0, Beta 2, 30/07/01~60 minutes better than first attempt, couple of technical fudge ups later on but down to operator incompetence & broken cd player.. comments.... as before no functional problems/crashes etc. still dubious about the new xaudio dll, i'm sure it was this causing audio lockups early on..... this time round, seemed to have even more probs with .wav files, stalled a lot getting them from the network which didn't happen before - again think this must be the new xaudio dll. either way, wav's are still buggered. found i tended to have 1 or 2 tracks cued at any one time on both players. generally doing back2back on 1 player but for jingle inserts switching to the other player for the jingle then playing the track on that player till the next jingle. occassionally used the other player for playing jingle/snippet over a track. 2 cards are a must. the show was alot tighter than last... the "preload" mechanism of jingles seems to be working tho there were a couple of instances where delays did crop up - mostly on wavs but i thought some mp3s did it as well. there's at least 1 incident on the mdisc where this delay causes a clip to crash the vocals. there's also (at least one) where you can just about hear the start has been clipped. finally - fudge ups on the recording down to batteries running out on the mdisc.... me failing to switch over i/p channels on the mixer - no probs with the main app. jb, 31/07/01 OAS Playout v1.0, Beta 2, 23/07/01~60 minutes, 23/07/01, generally naff - i'm crap coz i aint done this for nigh on 8 years... OAS Playout, pre beta2 release although no functions over beta1 used in trial. comments.... it stayed up for the entire duration, there were no lockups or crashes. ignoring my inability to speak properly or occassionally hit the button at the right time it all hung together. had a couple of pauses during playback of 2 tracks, notably when i was cueing up wav files on the other player. found i tended to have 1 or 2 tracks cued at any one time on both players. generally doing back2back on 1 player but for jingle inserts switching to the other player for the jingle then playing the track on that player till the next jingle. occassionally used the other player for playing jingle/snippet over a track. 2 cards are a must. this bug is the biggest git and makes the auto-fade facility damn near useless: "Notable pause of aprox 1s occuring randomly on audio clips through the Jingle Console" i could almost live with it if the music didn't duck early then get this zonking pause of nothing. turned it off for most of the show. seriously need to come up with a hack for this at least on the jingle side. found wav files aren't handled at all well or consistently in auto-seq mode, the ends frequently get chopped - this is now logged as a bug. the problems with short clips didn't really come up - I avoided them for the most part. summary ------- avoiding the bits we know are troublesome - auto-fade & really short wavs i think it held together okay. jb, 23/07/01 | ||||||||||