One-Synth-Challenge 52: Podolski - Voting and Results

[Mike777]

... I honestly didn't expect my first-ever OSC submission to garner even a single point. To walk away with two points, I'm ecstatic -- thanks again you two for making my day!

Because of the voting system only voting for five entries, that means a vote of even zero points (like one of mine) is that out of 65 entries that entry did not make anyone's 1-5 place choice, that's all it means. But it could have been someone's 6th, 7th, 8th, etc place choice if voting system required voting for top 10 or top 20 or top 40.

So even zero points with current voting system doesn't indicate where that song would be if everyone had to vote on all songs (or say the top 20 or 40). So I support the idea of voting all entries.

Congrats to all for great music made!

[tomtoo]

... .. So I support the idea of voting all entries...

I dont think so. Iam shure it would not change much. It would only make voting more complex, what would in the end lead to less voters.

There are so many different reasons why people give there points, and the more people, the more interesting it gets, imho.

[Mike777]

... .. So I support the idea of voting all entries...

I dont think so. Iam shure it would not change much. It would only make voting more complex, what would in the end lead to less voters.

There are so many different reasons why people give there points, and the more people, the more interesting it gets, imho.

It would be very easy to give every song a score of 1 through 5. Super easy. Nothing complex about it. AND it helps prove everyone listened to every song, which is another very good reason to score all songs. imo

[tomtoo]

... .. So I support the idea of voting all entries...

I dont think so. Iam shure it would not change much. It would only make voting more complex, what would in the end lead to less voters.

There are so many different reasons why people give there points, and the more people, the more interesting it gets, imho.

It would be very easy to give every song a score of 1 through 5. Super easy. Nothing complex about it. AND it helps prove everyone listened to every song, which is another very good reason to score all songs. imo

I not see where it should help to prove everyone listened to every song ?
To be honest i dont think there is ANY way to proof this.
Everybody has his own way to listen thru to songs. I could say if the start is not catchy i skip.
Or if i dont like the melody or the style of music.

If many people vote there will do it on there way and i am not the person that likes to tell whats the right way.

But iam shure that a easy voting system brings more voters.

[V'ger]

Agree 0 points sounds worse than it actually is, but not sure a system to root out any songs getting 0 points is any better for the songs that will eventually come last anyway? We will anyway have a vote on this shortly and let the majority settle it.

I just noticed that the track time being displayed in the Archive.Com player for my track "The Night Is Dark..." is a whopping 33 minutes and 29 seconds. Even the twelve-inch extended dance version, "The Night Is Young And Full Of Teenyboppers", is not quite that long!!!

Indeed, very bizarre that, don't know what's going on there.. I've tried a few things without success, so now I'm de-tagging it to see if that will help. If not maybe it needs to be recoded which might be best from the DAW instead of mp3 to mp3, so can you upload please?

I have uploaded my own encoded MP3 file of the song to this location. Please let me know how this file works out.

Thanks in advance.
Tony

Cheers, but no matter if I use the new file, rename it, de-tag it or whatever, it always ends up with the faulty length at Archive.org, it's very strange.. What mp3 app are you using? Do you mind trying another one? Here is the latest Lame: http://www.rarewares.org/mp3-lame-bundle.php

[tonedef71]

Cheers, but no matter if I use the new file, rename it, de-tag it or whatever, it always ends up with the faulty length at Archive.org, it's very strange.. What mp3 app are you using? Do you mind trying another one? Here is the latest Lame: http://www.rarewares.org/mp3-lame-bundle.php

Initially, I exported a WAV file rendering of the track using my Sonar X2 DAW. Then I used Adobe Audition 5 to encode it to an MP3 file.

This time, I downloaded the latest lame-enc.dll from the URL you provided and used it with MediaMonkey 4 to re-encode the previously rendered WAV file to a CBR 320 bps MP3 file. This new MP3 has been uploaded to the following location. Crossing my fingers that this re-encoding does the trick.

BTW: Does it matter that the bitrate of the audio is 48KHz?

Thanks in advance.
--Tony

[tonedef71]

... I honestly didn't expect my first-ever OSC submission to garner even a single point. To walk away with two points, I'm ecstatic -- thanks again you two for making my day!

Because of the voting system only voting for five entries, that means a vote of even zero points (like one of mine) is that out of 65 entries that entry did not make anyone's 1-5 place choice, that's all it means. But it could have been someone's 6th, 7th, 8th, etc place choice if voting system required voting for top 10 or top 20 or top 40.

True dat... I'll never know if my track could have placed in anybody's personal top ten or such, but with the large number of excellent-sounding entries, I had low expectations right from the outset. My track being voted the fifth place choice of two of the voters (an outcome that far exceeded my expectations) was such a pleasant surprise, that it really made my day.

Whoever said there is no upside to lowered expectations? This rhetorical question is akin to one of my favorite quotations, which I'll paraphrase: "It is better to under-promise and over-deliver, than to over-promise and under-deliver."

[ghettosynth]

... .. So I support the idea of voting all entries...

I dont think so. Iam shure it would not change much. It would only make voting more complex, what would in the end lead to less voters.

There are so many different reasons why people give there points, and the more people, the more interesting it gets, imho.

It would be very easy to give every song a score of 1 through 5. Super easy. Nothing complex about it. AND it helps prove everyone listened to every song, which is another very good reason to score all songs. imo

First, it won't ensure that every song gets listened to, people would just give a 1 to songs that they didn't want to really evaluate. Rankings are useful, they don't require any subjective evaluation of the value of the score. By that I mean, what's a "2", what's a "1"? Ranking just says, "I like this song better than that song."

This is why I suggested providing an easily sorted list of songs so that people can just sort the list. If you rank the songs so that highest rank = liked best and lowest rank = liked least then the average rank (or median rank) gives a nice meaningful reflection of what the community thinks of your music.

Think about this a bit. Giving 1's to everyone won't change the perception of not being able to distinguish the list of equals at the bottom. But, ranking the tracks gives a meaningful perception. I'm not saying, "oh, these are all equally uninteresting to me, I'll give them all a one", I'm forced to say, "well, these are the least interesting, but, I like this one better than that one, and this other one better than both of them."

Forget the numbers, just have voters order the songs from most favorite to least favorite. In this case, even with few voters, you will still not have similarity among the lists because every voter gives an ordering and unless the number of songs are very small the probability that all users like all songs in the same order is very unlikely.

Now, with the sorted lists, you can use the scoring method that you currently have, or you can give every song a score i.e. instead of 5 down to 1, the number of entries down to 1. Or you could assign some scaled scoring all the way down, e.g. log of the position. There's probably some good information on what makes an effective scoring system in these types of things, I'll see what I can find.

[bjporter]

... .. So I support the idea of voting all entries...

I dont think so. Iam shure it would not change much. It would only make voting more complex, what would in the end lead to less voters.

There are so many different reasons why people give there points, and the more people, the more interesting it gets, imho.

It would be very easy to give every song a score of 1 through 5. Super easy. Nothing complex about it. AND it helps prove everyone listened to every song, which is another very good reason to score all songs. imo

First, it won't ensure that every song gets listened to, people would just give a 1 to songs that they didn't want to really evaluate. Rankings are useful, they don't require any subjective evaluation of the value of the score. By that I mean, what's a "2", what's a "1"? Ranking just says, "I like this song better than that song."

This is why I suggested providing an easily sorted list of songs so that people can just sort the list. If you rank the songs so that highest rank = liked best and lowest rank = liked least then the average rank gives a nice meaningful reflection of what the community thinks of your music.

Think about this a bit. Giving 1's to everyone won't change the perception of not being able to distinguish the list of equals at the bottom. But, ranking the tracks gives a meaningful perception. I'm not saying, "oh, these are all equally uninteresting to me, I'll give them all a one", I'm forced to say, "well, these are the least interesting, but, I like this one better than that one, and this other one better than both of them."

Forget the numbers, just have voters order the songs from most favorite to least favorite. In this case, even with few voters, you will still not have similarity among the lists because every song gets an ordering.

Now, with the sorted lists, you can use the scoring method that you currently have, or you can give every song a score i.e. instead of 5 down to 1, the number of entries down to 1. Or you could assign some scaled scoring all the way down, e.g. log of the position. There's probably some good information on what makes an effective scoring system in these types of things, I'll see what I can find.

Agreed...

Also people are getting a ton of plays and comments, which is very tough to do without this contest... that's quite a bit of goodness there... I'd rather have 200 plays and lots of really helpful comments, than a bunch of points that still puts me at the bottom.

[V'ger]

Cheers, but no matter if I use the new file, rename it, de-tag it or whatever, it always ends up with the faulty length at Archive.org, it's very strange.. What mp3 app are you using? Do you mind trying another one? Here is the latest Lame: http://www.rarewares.org/mp3-lame-bundle.php

Initially, I exported a WAV file rendering of the track using my Sonar X2 DAW. Then I used Adobe Audition 5 to encode it to an MP3 file.

This time, I downloaded the latest lame-enc.dll from the URL you provided and used it with MediaMonkey 4 to re-encode the previously rendered WAV file to a CBR 320 bps MP3 file. This new MP3 has been uploaded to the following location. Crossing my fingers that this re-encoding does the trick.

BTW: Does it matter that the bitrate of the audio is 48KHz?

Let's see if that does the trick. Will have to wait between tries as Archive.org puts all jobs in a queue.

48khz should have nothing to do with it as there have been many mp3s with that sample rate before. In fact I believe this is the first time we've seen this time issue.

But curious why you don't export to, or convert as, mp3 in Sonar directly? Result not good enough?

[V'ger]

Well that fixed it. Seems it was Adobe Audition 5 as encodes to two different files produced the same error, but as soon as another encoder was used there was no error.

[tonedef71]

But curious why you don't export to, or convert as, mp3 in Sonar directly? Result not good enough?

For the longest time, SONAR only permitted MP3 export if you purchased a special license, which was extra. Apparently they added free support for MP3 export in the Sonar X series, and I just never knew about it. I've always used Adobe Audition to master/finalize my exported WAV audio anyway, so it just made sense to do the MP3 encoding from in there. When ripping audio CDs, I use Media Monkey 4, which uses the LAME encoder.

[tonedef71]

Well that fixed it. Seems it was Adobe Audition 5 as encodes to two different files produced the same error, but as soon as another encoder was used there was no error.

Thank you so much for taking care of this for me!

I remember that Adobe Audition used to utilize an MP3 encoder called MP3 Pro. I'm not sure which encoder they use now in Audition 5, but I did read somewhere that they completely rewrote the audio plugin architecture from scratch -- many of the audio formats previously supported by Audition are no longer supported in version 5 because the old plugins are incompatible with the new audio plugin architecture. The Adobe bakers have built-in support for all of the audio formats that they believe still matter anymore in the industry -- they have no plans to support old audio formats and the new plugin architecture is not open source. I suppose I no longer require support for old audio file formats like Creative Labs .VOC and Sun Microsystems .AU anyway.

[V'ger]

I see Adobe Audition 5 uses the payware Fraunhofer encoder which I guess was in the old Sonar too. I suppose Sonar now uses Lame or similar since it's free but anyway better at calculating length it seems.

In my DAW you can render the song directly to mp3, but never did tests to see if it was better than rendering to a HQ wav and converting after. Probably about the same, anybody know?