TIFF--> JXL -Comparing Affinty "100% Quality" to XnConvert Lossless. No comparison?
Posted: Fri Jul 04, 2025 7:17 am
A friend was convering a 545 MB TIFF to JXl in Affinty and got a 81 MB. We both assumed that when Affinity said "100% Quality" export it meant Lossless. Now I think it means "visually lossless", i.s.lossy.
Lossless with XnConvert was 368 MB. I tried a variety of combinations for 99% to 95% quality and compression from 1-10 in XnConvert and found that XnConvert was giving simliar sizes to Affinity when 95% to 96% quality was used, specifically 96% was 77MB at 1 and 86MB at 10. He said on his Apple Max Studio M4 Max with 128GB it took him about 20 seconds to export the TIFF. I get not disimilar results with Quality 90% and compresion 7 on a 12th Gen Intel(R) Core(TM) i7-1260P 2.10 GHz with 64GB (Intel NUC from 2022).
When I research this on the web, I get this answer: "Affinity's 100% Quality setting is not technically lossless. While it aims to preserve as much image quality as possible, it typically uses lossy compression methods, which can result in some loss of data and quality."That seems correct to me.
My friend understandably equates "Quality = 100%" with lossless.
While every file conversion is unique and no general rule can confidently be applied using just one example. So got the Affinity 7 day trial and test the 100% quality on 30 other TIFF that I had already converted with XnConvert. In each case the Affinty 100% Quality was yielding compressions similar to XnConvert 97% compression 2.
So I am thinking to tell my friend that as visually satisfying as the Affinity Quality=100% maybe he shouldn't be lulled into the sense it is actually lossless.
I have thought to convert back and forth between TIFF and JXL to empirically verify that Afffinity is not truly lossless. I think that would be the acid test.
I just need a batch converter. XnConvert is ideal for me. He works 100 hours on one image. So, it should matter to him if his format is lossless or not. Though I think wisely he probably works in TIFF and only makes the final conversion to JXL at the end. At least I hope so.
Lossless with XnConvert was 368 MB. I tried a variety of combinations for 99% to 95% quality and compression from 1-10 in XnConvert and found that XnConvert was giving simliar sizes to Affinity when 95% to 96% quality was used, specifically 96% was 77MB at 1 and 86MB at 10. He said on his Apple Max Studio M4 Max with 128GB it took him about 20 seconds to export the TIFF. I get not disimilar results with Quality 90% and compresion 7 on a 12th Gen Intel(R) Core(TM) i7-1260P 2.10 GHz with 64GB (Intel NUC from 2022).
When I research this on the web, I get this answer: "Affinity's 100% Quality setting is not technically lossless. While it aims to preserve as much image quality as possible, it typically uses lossy compression methods, which can result in some loss of data and quality."That seems correct to me.
My friend understandably equates "Quality = 100%" with lossless.
While every file conversion is unique and no general rule can confidently be applied using just one example. So got the Affinity 7 day trial and test the 100% quality on 30 other TIFF that I had already converted with XnConvert. In each case the Affinty 100% Quality was yielding compressions similar to XnConvert 97% compression 2.
So I am thinking to tell my friend that as visually satisfying as the Affinity Quality=100% maybe he shouldn't be lulled into the sense it is actually lossless.
I have thought to convert back and forth between TIFF and JXL to empirically verify that Afffinity is not truly lossless. I think that would be the acid test.
I just need a batch converter. XnConvert is ideal for me. He works 100 hours on one image. So, it should matter to him if his format is lossless or not. Though I think wisely he probably works in TIFF and only makes the final conversion to JXL at the end. At least I hope so.