Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Backup of SQAM samples (Read 3412 times) previous topic - next topic
0 Members and 2 Guests are viewing this topic.

Backup of SQAM samples

Hi,

everyone is referencing these samples on HA, but the downloads are down. (https://sound.media.mit.edu/resources/mpeg4/audio/sqam/) Since I agree with the HA idea of ABXing, I want to try the most terrible encoder killers myself.

Does anyone still have them and could upload them here?

Thank you!


Re: Backup of SQAM samples

Reply #2
Wow I've been googling and opening dozens of HA threads in hopes of running across harp4 and did not see this.

Thank you very much Chris!

Re: Backup of SQAM samples

Reply #3
Wow I've been googling and opening dozens of HA threads in hopes of running across harp4 and did not see this.
Is this the one you're looking for?

Edit: the thread where this sample is in is one of the few where most links are still fully operational, as the samples seem to have been attached to it.
Listen to the music, not the media it's on.
União e reconstrução

Re: Backup of SQAM samples

Reply #4
@includemeout no, it was the one people refer to as harp40_1 which is listed here https://sound.media.mit.edu/resources/mpeg4/audio/sqam/ and subsequently https://tech.ebu.ch/publications/sqamcd as track 40. But thank you very much for checking!

I think I got all the most brutal samples now and my results so far are quite brutal as well. I believed I would never be able to ABX them with OPUS, but I can get at least 20% (in foo abx) already on one of them on my crappy speakers with 320 VBR.. I know 5% or less are more bulletproof, but I was expecting total randomness as OPUS has shown to be fully transparent in all the high bitrate tests on HA.
It seems 'killer' samples indeed trigger any encoder. And then I saw this one where other people confirm even the highest settings don't work: https://hydrogenaud.io/index.php?topic=120193.25

Chiptunes are also terribly easy to ABX at times. I thought I was hallucinating, but then tried and actually got it right. I was firmly in the camp that 'it's all transparent period'. Learned my lesson. I found one that I can ABX almost perfectly because the 'smooth yet distinct' effect on a 'chirp' (I forgot what that's called in tracking) turns to mush. I never really sat down to ABX previously, but I heard that one right away.

This really made my decision hard to convert a bit of my collection due to lack of space. Yes, these situations are probably very rare, but if even someone like me, avid music fan, but not trained ABXer, can ABX them when they do occur, that's not very good for a permanent archive.

Re: Backup of SQAM samples

Reply #5
Agreed. I would not replace a lossless audio collection by a lossy coded one. Storage has gotten cheap enough to have e.g. a FLAC archive. I transcoded to xHE-AAC only for portable, casual listening, but kept lossless rips of my favorite audio CDs due to CD(-ROM) readers nearing extinction in modern devices.

But also keep in mind that test signals, chiptunes, and isolated musical instruments as on the SQAM CD, are very different from typical music. On most material, modern lossy audio codecs do very well.

Chris
If I don't reply to your reply, it means I agree with you.

Re: Backup of SQAM samples

Reply #6
Both good points. I may have never noticed if I had just 'casually' listened. But I do listen 'consciously' probably more often than average. Storage is a problem though in my case, my collection has grown too large for affordable retail storage if I plan ahead. (with a backup) I have a lot of experimental audio that only fits the mood once a year or less, but I don't want to get rid of it. Not sure if that's more a stupid collecting addiction or 'sophistication'. I gravitate towards the former, but I can't hit the delete button :).

Re-encoding is not possible in many instances here as well. Friends digitized some tapes on their gear for me, I don't have a CD reader anymore, and who knows how long the modern web release platforms will exist for possible re-download. (watching EAC was satisfying in the CD era)

My thesis for conversion was that (at least until I hit these specific samples) it was all fully transparent and there's no real reason to store data that just sits there. I felt I was being too 'audiophile'. Even with these it seems to be highly dependent on time of day and mood if I can recognize it.
On the other hand, when we still had CDs we had full 16/44 no questions asked. But I don't want to go offtopic in the listening test area. Thanks for helping me find the samples and the food for thought.